diff --git "a/checkpoint-30000/trainer_state.json" "b/checkpoint-30000/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-30000/trainer_state.json" @@ -0,0 +1,216652 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.5957818644000477, + "global_step": 30000, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-07, + "loss": 10.5441, + "theoretical_loss": 20.841284642178216, + "tokens_seen": 65536 + }, + { + "epoch": 0.0, + "learning_rate": 1.984126984126984e-06, + "loss": 10.5415, + "theoretical_loss": 17.59466794495971, + "tokens_seen": 131072 + }, + { + "epoch": 0.0, + "learning_rate": 2.9761904761904763e-06, + "loss": 10.4877, + "theoretical_loss": 15.967943933467634, + "tokens_seen": 196608 + }, + { + "epoch": 0.0, + "learning_rate": 3.968253968253968e-06, + "loss": 10.3405, + "theoretical_loss": 14.920783596619636, + "tokens_seen": 262144 + }, + { + "epoch": 0.0, + "learning_rate": 4.96031746031746e-06, + "loss": 10.1428, + "theoretical_loss": 14.164683778394041, + "tokens_seen": 327680 + }, + { + "epoch": 0.0, + "learning_rate": 5.9523809523809525e-06, + "loss": 9.9242, + "theoretical_loss": 13.581028313181289, + "tokens_seen": 393216 + }, + { + "epoch": 0.0, + "learning_rate": 6.944444444444444e-06, + "loss": 9.7074, + "theoretical_loss": 13.110275842032255, + "tokens_seen": 458752 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507936e-06, + "loss": 9.5301, + "theoretical_loss": 12.71859646611439, + "tokens_seen": 524288 + }, + { + "epoch": 0.0, + "learning_rate": 8.928571428571428e-06, + "loss": 9.3999, + "theoretical_loss": 12.38505930450469, + "tokens_seen": 589824 + }, + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-06, + "loss": 9.253, + "theoretical_loss": 12.095879447666144, + "tokens_seen": 655360 + }, + { + "epoch": 0.0, + "learning_rate": 1.0912698412698412e-05, + "loss": 9.1543, + "theoretical_loss": 11.841533071732742, + "tokens_seen": 720896 + }, + { + "epoch": 0.0, + "learning_rate": 1.1904761904761905e-05, + "loss": 9.013, + "theoretical_loss": 11.615186049337796, + "tokens_seen": 786432 + }, + { + "epoch": 0.0, + "learning_rate": 1.2896825396825396e-05, + "loss": 8.9241, + "theoretical_loss": 11.411780980817854, + "tokens_seen": 851968 + }, + { + "epoch": 0.0, + "learning_rate": 1.3888888888888888e-05, + "loss": 8.8292, + "theoretical_loss": 11.227478542742938, + "tokens_seen": 917504 + }, + { + "epoch": 0.0, + "learning_rate": 1.4880952380952381e-05, + "loss": 8.7366, + "theoretical_loss": 11.059300758608979, + "tokens_seen": 983040 + }, + { + "epoch": 0.0, + "learning_rate": 1.5873015873015872e-05, + "loss": 8.6875, + "theoretical_loss": 10.904894927088016, + "tokens_seen": 1048576 + }, + { + "epoch": 0.0, + "learning_rate": 1.6865079365079364e-05, + "loss": 8.786, + "theoretical_loss": 10.76237258069379, + "tokens_seen": 1114112 + }, + { + "epoch": 0.0, + "learning_rate": 1.7857142857142855e-05, + "loss": 8.6125, + "theoretical_loss": 10.630196716861345, + "tokens_seen": 1179648 + }, + { + "epoch": 0.0, + "learning_rate": 1.884920634920635e-05, + "loss": 8.6043, + "theoretical_loss": 10.507101012987707, + "tokens_seen": 1245184 + }, + { + "epoch": 0.0, + "learning_rate": 1.984126984126984e-05, + "loss": 8.557, + "theoretical_loss": 10.392030784394397, + "tokens_seen": 1310720 + }, + { + "epoch": 0.0, + "learning_rate": 2.0833333333333333e-05, + "loss": 8.3781, + "theoretical_loss": 10.284099061527812, + "tokens_seen": 1376256 + }, + { + "epoch": 0.0, + "learning_rate": 2.1825396825396824e-05, + "loss": 8.452, + "theoretical_loss": 10.182553393901085, + "tokens_seen": 1441792 + }, + { + "epoch": 0.0, + "learning_rate": 2.2817460317460315e-05, + "loss": 8.3968, + "theoretical_loss": 10.086750402363664, + "tokens_seen": 1507328 + }, + { + "epoch": 0.0, + "learning_rate": 2.380952380952381e-05, + "loss": 8.3586, + "theoretical_loss": 9.996136019471344, + "tokens_seen": 1572864 + }, + { + "epoch": 0.0, + "learning_rate": 2.48015873015873e-05, + "loss": 8.2227, + "theoretical_loss": 9.910229967024176, + "tokens_seen": 1638400 + }, + { + "epoch": 0.0, + "learning_rate": 2.5793650793650793e-05, + "loss": 8.3317, + "theoretical_loss": 9.828613432171625, + "tokens_seen": 1703936 + }, + { + "epoch": 0.0, + "learning_rate": 2.6785714285714284e-05, + "loss": 8.2379, + "theoretical_loss": 9.75091918753851, + "tokens_seen": 1769472 + }, + { + "epoch": 0.0, + "learning_rate": 2.7777777777777776e-05, + "loss": 8.2491, + "theoretical_loss": 9.676823599712613, + "tokens_seen": 1835008 + }, + { + "epoch": 0.0, + "learning_rate": 2.876984126984127e-05, + "loss": 8.1307, + "theoretical_loss": 9.606040111776705, + "tokens_seen": 1900544 + }, + { + "epoch": 0.0, + "learning_rate": 2.9761904761904762e-05, + "loss": 8.1334, + "theoretical_loss": 9.538313887395919, + "tokens_seen": 1966080 + }, + { + "epoch": 0.0, + "learning_rate": 3.075396825396825e-05, + "loss": 8.1187, + "theoretical_loss": 9.473417378263001, + "tokens_seen": 2031616 + }, + { + "epoch": 0.0, + "learning_rate": 3.1746031746031745e-05, + "loss": 8.0782, + "theoretical_loss": 9.411146631541524, + "tokens_seen": 2097152 + }, + { + "epoch": 0.0, + "learning_rate": 3.273809523809524e-05, + "loss": 8.0006, + "theoretical_loss": 9.351318194865048, + "tokens_seen": 2162688 + }, + { + "epoch": 0.0, + "learning_rate": 3.373015873015873e-05, + "loss": 7.9401, + "theoretical_loss": 9.293766507291341, + "tokens_seen": 2228224 + }, + { + "epoch": 0.0, + "learning_rate": 3.472222222222222e-05, + "loss": 8.107, + "theoretical_loss": 9.238341688075822, + "tokens_seen": 2293760 + }, + { + "epoch": 0.0, + "learning_rate": 3.571428571428571e-05, + "loss": 7.8785, + "theoretical_loss": 9.184907653139359, + "tokens_seen": 2359296 + }, + { + "epoch": 0.0, + "learning_rate": 3.670634920634921e-05, + "loss": 7.8914, + "theoretical_loss": 9.133340503044415, + "tokens_seen": 2424832 + }, + { + "epoch": 0.0, + "learning_rate": 3.76984126984127e-05, + "loss": 7.7732, + "theoretical_loss": 9.0835271371648, + "tokens_seen": 2490368 + }, + { + "epoch": 0.0, + "learning_rate": 3.8690476190476195e-05, + "loss": 7.7819, + "theoretical_loss": 9.035364057274894, + "tokens_seen": 2555904 + }, + { + "epoch": 0.0, + "learning_rate": 3.968253968253968e-05, + "loss": 7.8541, + "theoretical_loss": 8.988756330540422, + "tokens_seen": 2621440 + }, + { + "epoch": 0.0, + "learning_rate": 4.067460317460318e-05, + "loss": 7.6904, + "theoretical_loss": 8.943616687271678, + "tokens_seen": 2686976 + }, + { + "epoch": 0.0, + "learning_rate": 4.1666666666666665e-05, + "loss": 7.6758, + "theoretical_loss": 8.89986473310929, + "tokens_seen": 2752512 + }, + { + "epoch": 0.0, + "learning_rate": 4.265873015873016e-05, + "loss": 7.4939, + "theoretical_loss": 8.857426258784564, + "tokens_seen": 2818048 + }, + { + "epoch": 0.0, + "learning_rate": 4.365079365079365e-05, + "loss": 7.4488, + "theoretical_loss": 8.816232633409479, + "tokens_seen": 2883584 + }, + { + "epoch": 0.0, + "learning_rate": 4.464285714285714e-05, + "loss": 7.6424, + "theoretical_loss": 8.776220269542522, + "tokens_seen": 2949120 + }, + { + "epoch": 0.0, + "learning_rate": 4.563492063492063e-05, + "loss": 7.4913, + "theoretical_loss": 8.737330150151898, + "tokens_seen": 3014656 + }, + { + "epoch": 0.0, + "learning_rate": 4.6626984126984126e-05, + "loss": 7.459, + "theoretical_loss": 8.699507409139912, + "tokens_seen": 3080192 + }, + { + "epoch": 0.0, + "learning_rate": 4.761904761904762e-05, + "loss": 7.4548, + "theoretical_loss": 8.662700958366539, + "tokens_seen": 3145728 + }, + { + "epoch": 0.0, + "learning_rate": 4.8611111111111115e-05, + "loss": 7.3781, + "theoretical_loss": 8.62686315516739, + "tokens_seen": 3211264 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 8371, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 7.391097068786621, + "objective/train/theoretical_loss": 8.591949505242134, + "objective/train/tokens_used": 23736800, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 4.96031746031746e-05, + "loss": 7.3088, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 5.05952380952381e-05, + "loss": 7.2003, + "theoretical_loss": 8.557918396526182, + "tokens_seen": 3342336 + }, + { + "epoch": 0.0, + "learning_rate": 5.1587301587301586e-05, + "loss": 7.1391, + "theoretical_loss": 8.524730860277067, + "tokens_seen": 3407872 + }, + { + "epoch": 0.0, + "learning_rate": 5.257936507936508e-05, + "loss": 7.1691, + "theoretical_loss": 8.492350356128252, + "tokens_seen": 3473408 + }, + { + "epoch": 0.0, + "learning_rate": 5.357142857142857e-05, + "loss": 7.0778, + "theoretical_loss": 8.460742578303845, + "tokens_seen": 3538944 + }, + { + "epoch": 0.0, + "learning_rate": 5.4563492063492063e-05, + "loss": 7.0848, + "theoretical_loss": 8.429875280561783, + "tokens_seen": 3604480 + }, + { + "epoch": 0.0, + "learning_rate": 5.555555555555555e-05, + "loss": 7.0849, + "theoretical_loss": 8.399718117751275, + "tokens_seen": 3670016 + }, + { + "epoch": 0.0, + "learning_rate": 5.6547619047619046e-05, + "loss": 7.0073, + "theoretical_loss": 8.370242502142142, + "tokens_seen": 3735552 + }, + { + "epoch": 0.0, + "learning_rate": 5.753968253968254e-05, + "loss": 7.0037, + "theoretical_loss": 8.341421472916394, + "tokens_seen": 3801088 + }, + { + "epoch": 0.0, + "learning_rate": 5.8531746031746036e-05, + "loss": 6.8098, + "theoretical_loss": 8.313229577412248, + "tokens_seen": 3866624 + }, + { + "epoch": 0.0, + "learning_rate": 5.9523809523809524e-05, + "loss": 6.8654, + "theoretical_loss": 8.28564276288293, + "tokens_seen": 3932160 + }, + { + "epoch": 0.0, + "learning_rate": 6.051587301587302e-05, + "loss": 6.8585, + "theoretical_loss": 8.258638277681179, + "tokens_seen": 3997696 + }, + { + "epoch": 0.0, + "learning_rate": 6.15079365079365e-05, + "loss": 6.7138, + "theoretical_loss": 8.232194580909036, + "tokens_seen": 4063232 + }, + { + "epoch": 0.0, + "learning_rate": 6.25e-05, + "loss": 6.7182, + "theoretical_loss": 8.20629125968411, + "tokens_seen": 4128768 + }, + { + "epoch": 0.0, + "learning_rate": 6.349206349206349e-05, + "loss": 6.7703, + "theoretical_loss": 8.180908953270682, + "tokens_seen": 4194304 + }, + { + "epoch": 0.0, + "learning_rate": 6.448412698412699e-05, + "loss": 6.5948, + "theoretical_loss": 8.156029283408564, + "tokens_seen": 4259840 + }, + { + "epoch": 0.0, + "learning_rate": 6.547619047619048e-05, + "loss": 6.6068, + "theoretical_loss": 8.131634790246775, + "tokens_seen": 4325376 + }, + { + "epoch": 0.0, + "learning_rate": 6.646825396825397e-05, + "loss": 6.609, + "theoretical_loss": 8.107708873353706, + "tokens_seen": 4390912 + }, + { + "epoch": 0.0, + "learning_rate": 6.746031746031745e-05, + "loss": 6.498, + "theoretical_loss": 8.084235737332481, + "tokens_seen": 4456448 + }, + { + "epoch": 0.0, + "learning_rate": 6.845238095238096e-05, + "loss": 6.462, + "theoretical_loss": 8.061200341620074, + "tokens_seen": 4521984 + }, + { + "epoch": 0.0, + "learning_rate": 6.944444444444444e-05, + "loss": 6.4268, + "theoretical_loss": 8.038588354092902, + "tokens_seen": 4587520 + }, + { + "epoch": 0.0, + "learning_rate": 7.043650793650793e-05, + "loss": 6.2784, + "theoretical_loss": 8.016386108140445, + "tokens_seen": 4653056 + }, + { + "epoch": 0.0, + "learning_rate": 7.142857142857142e-05, + "loss": 6.3876, + "theoretical_loss": 7.994580562902867, + "tokens_seen": 4718592 + }, + { + "epoch": 0.0, + "learning_rate": 7.242063492063492e-05, + "loss": 6.4554, + "theoretical_loss": 7.973159266399054, + "tokens_seen": 4784128 + }, + { + "epoch": 0.0, + "learning_rate": 7.341269841269842e-05, + "loss": 6.2915, + "theoretical_loss": 7.952110321298584, + "tokens_seen": 4849664 + }, + { + "epoch": 0.0, + "learning_rate": 7.440476190476191e-05, + "loss": 6.3854, + "theoretical_loss": 7.931422353115133, + "tokens_seen": 4915200 + }, + { + "epoch": 0.0, + "learning_rate": 7.53968253968254e-05, + "loss": 6.2229, + "theoretical_loss": 7.911084480620269, + "tokens_seen": 4980736 + }, + { + "epoch": 0.0, + "learning_rate": 7.63888888888889e-05, + "loss": 6.1383, + "theoretical_loss": 7.891086288295694, + "tokens_seen": 5046272 + }, + { + "epoch": 0.0, + "learning_rate": 7.738095238095239e-05, + "loss": 6.2541, + "theoretical_loss": 7.871417800659003, + "tokens_seen": 5111808 + }, + { + "epoch": 0.0, + "learning_rate": 7.837301587301588e-05, + "loss": 6.2144, + "theoretical_loss": 7.8520694583133785, + "tokens_seen": 5177344 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507937e-05, + "loss": 6.0549, + "theoretical_loss": 7.833032095585231, + "tokens_seen": 5242880 + }, + { + "epoch": 0.0, + "learning_rate": 8.035714285714287e-05, + "loss": 6.0873, + "theoretical_loss": 7.8142969196261385, + "tokens_seen": 5308416 + }, + { + "epoch": 0.0, + "learning_rate": 8.134920634920635e-05, + "loss": 6.0059, + "theoretical_loss": 7.79585549086643, + "tokens_seen": 5373952 + }, + { + "epoch": 0.0, + "learning_rate": 8.234126984126984e-05, + "loss": 6.1625, + "theoretical_loss": 7.777699704717685, + "tokens_seen": 5439488 + }, + { + "epoch": 0.0, + "learning_rate": 8.333333333333333e-05, + "loss": 6.1117, + "theoretical_loss": 7.759821774430384, + "tokens_seen": 5505024 + }, + { + "epoch": 0.0, + "learning_rate": 8.432539682539683e-05, + "loss": 6.0106, + "theoretical_loss": 7.742214215020986, + "tokens_seen": 5570560 + }, + { + "epoch": 0.0, + "learning_rate": 8.531746031746032e-05, + "loss": 5.952, + "theoretical_loss": 7.72486982819, + "tokens_seen": 5636096 + }, + { + "epoch": 0.0, + "learning_rate": 8.630952380952381e-05, + "loss": 5.931, + "theoretical_loss": 7.707781688159247, + "tokens_seen": 5701632 + }, + { + "epoch": 0.0, + "learning_rate": 8.73015873015873e-05, + "loss": 5.9401, + "theoretical_loss": 7.690943128362394, + "tokens_seen": 5767168 + }, + { + "epoch": 0.0, + "learning_rate": 8.82936507936508e-05, + "loss": 5.9093, + "theoretical_loss": 7.674347728928345, + "tokens_seen": 5832704 + }, + { + "epoch": 0.0, + "learning_rate": 8.928571428571429e-05, + "loss": 5.8234, + "theoretical_loss": 7.6579893049019265, + "tokens_seen": 5898240 + }, + { + "epoch": 0.0, + "learning_rate": 9.027777777777777e-05, + "loss": 5.9277, + "theoretical_loss": 7.641861895150762, + "tokens_seen": 5963776 + }, + { + "epoch": 0.0, + "learning_rate": 9.126984126984126e-05, + "loss": 5.8891, + "theoretical_loss": 7.625959751911346, + "tokens_seen": 6029312 + }, + { + "epoch": 0.0, + "learning_rate": 9.226190476190476e-05, + "loss": 5.8799, + "theoretical_loss": 7.610277330930954, + "tokens_seen": 6094848 + }, + { + "epoch": 0.0, + "learning_rate": 9.325396825396825e-05, + "loss": 5.803, + "theoretical_loss": 7.5948092821654365, + "tokens_seen": 6160384 + }, + { + "epoch": 0.0, + "learning_rate": 9.424603174603175e-05, + "loss": 5.7666, + "theoretical_loss": 7.579550440996015, + "tokens_seen": 6225920 + }, + { + "epoch": 0.0, + "learning_rate": 9.523809523809524e-05, + "loss": 5.8446, + "theoretical_loss": 7.564495819930985, + "tokens_seen": 6291456 + }, + { + "epoch": 0.0, + "learning_rate": 9.623015873015874e-05, + "loss": 5.8103, + "theoretical_loss": 7.549640600760828, + "tokens_seen": 6356992 + }, + { + "epoch": 0.0, + "learning_rate": 9.722222222222223e-05, + "loss": 5.7041, + "theoretical_loss": 7.534980127137563, + "tokens_seen": 6422528 + }, + { + "epoch": 0.0, + "learning_rate": 9.821428571428572e-05, + "loss": 5.7111, + "theoretical_loss": 7.520509897551315, + "tokens_seen": 6488064 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 9704, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.717738151550293, + "objective/train/theoretical_loss": 7.5062255586790965, + "objective/train/tokens_used": 27013600, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 9.92063492063492e-05, + "loss": 5.7683, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010019841269841271, + "loss": 5.8045, + "theoretical_loss": 7.492122899082528, + "tokens_seen": 6619136 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001011904761904762, + "loss": 5.5694, + "theoretical_loss": 7.478197843232982, + "tokens_seen": 6684672 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010218253968253968, + "loss": 5.7093, + "theoretical_loss": 7.464446445844085, + "tokens_seen": 6750208 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010317460317460317, + "loss": 5.7607, + "theoretical_loss": 7.450864886492971, + "tokens_seen": 6815744 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010416666666666667, + "loss": 5.5396, + "theoretical_loss": 7.437449464512941, + "tokens_seen": 6881280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010515873015873016, + "loss": 5.5782, + "theoretical_loss": 7.424196594141411, + "tokens_seen": 6946816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010615079365079365, + "loss": 5.6965, + "theoretical_loss": 7.4111027999080985, + "tokens_seen": 7012352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010714285714285714, + "loss": 5.7661, + "theoretical_loss": 7.398164712249439, + "tokens_seen": 7077888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010813492063492064, + "loss": 5.5926, + "theoretical_loss": 7.3853790633361545, + "tokens_seen": 7143424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010912698412698413, + "loss": 5.6062, + "theoretical_loss": 7.3727426831017535, + "tokens_seen": 7208960 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011011904761904761, + "loss": 5.6133, + "theoretical_loss": 7.360252495460585, + "tokens_seen": 7274496 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001111111111111111, + "loss": 5.5821, + "theoretical_loss": 7.347905514704743, + "tokens_seen": 7340032 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001121031746031746, + "loss": 5.5828, + "theoretical_loss": 7.33569884206989, + "tokens_seen": 7405568 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011309523809523809, + "loss": 5.6814, + "theoretical_loss": 7.32362966246063, + "tokens_seen": 7471104 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011408730158730158, + "loss": 5.4025, + "theoretical_loss": 7.311695241326718, + "tokens_seen": 7536640 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011507936507936508, + "loss": 5.4938, + "theoretical_loss": 7.299892921681898, + "tokens_seen": 7602176 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011607142857142858, + "loss": 5.6077, + "theoretical_loss": 7.288220121257689, + "tokens_seen": 7667712 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011706349206349207, + "loss": 5.5523, + "theoretical_loss": 7.276674329784936, + "tokens_seen": 7733248 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011805555555555556, + "loss": 5.5232, + "theoretical_loss": 7.265253106396328, + "tokens_seen": 7798784 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011904761904761905, + "loss": 5.6514, + "theoretical_loss": 7.253954077143563, + "tokens_seen": 7864320 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012003968253968255, + "loss": 5.5026, + "theoretical_loss": 7.242774932623186, + "tokens_seen": 7929856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012103174603174604, + "loss": 5.6399, + "theoretical_loss": 7.2317134257054665, + "tokens_seen": 7995392 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012202380952380953, + "loss": 5.3427, + "theoretical_loss": 7.220767369361081, + "tokens_seen": 8060928 + }, + { + "epoch": 0.0, + "learning_rate": 0.000123015873015873, + "loss": 5.4781, + "theoretical_loss": 7.2099346345805815, + "tokens_seen": 8126464 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001240079365079365, + "loss": 5.4975, + "theoretical_loss": 7.1992131483820305, + "tokens_seen": 8192000 + }, + { + "epoch": 0.0, + "learning_rate": 0.000125, + "loss": 5.3893, + "theoretical_loss": 7.188600891902338, + "tokens_seen": 8257536 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001259920634920635, + "loss": 5.4292, + "theoretical_loss": 7.178095898568204, + "tokens_seen": 8323072 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012698412698412698, + "loss": 5.3613, + "theoretical_loss": 7.167696252342704, + "tokens_seen": 8388608 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012797619047619048, + "loss": 5.373, + "theoretical_loss": 7.15740008604386, + "tokens_seen": 8454144 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012896825396825398, + "loss": 5.403, + "theoretical_loss": 7.147205579731682, + "tokens_seen": 8519680 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012996031746031748, + "loss": 5.3827, + "theoretical_loss": 7.1371109591604105, + "tokens_seen": 8585216 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013095238095238096, + "loss": 5.3453, + "theoretical_loss": 7.127114494292836, + "tokens_seen": 8650752 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013194444444444446, + "loss": 5.2623, + "theoretical_loss": 7.117214497873759, + "tokens_seen": 8716288 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013293650793650793, + "loss": 5.211, + "theoretical_loss": 7.107409324059815, + "tokens_seen": 8781824 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013392857142857144, + "loss": 5.4685, + "theoretical_loss": 7.097697367103032, + "tokens_seen": 8847360 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001349206349206349, + "loss": 5.2639, + "theoretical_loss": 7.0880770600856335, + "tokens_seen": 8912896 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001359126984126984, + "loss": 5.1417, + "theoretical_loss": 7.0785468737037345, + "tokens_seen": 8978432 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001369047619047619, + "loss": 5.3841, + "theoretical_loss": 7.069105315097691, + "tokens_seen": 9043968 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013789682539682541, + "loss": 5.2652, + "theoretical_loss": 7.059750926726999, + "tokens_seen": 9109504 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001388888888888889, + "loss": 5.3788, + "theoretical_loss": 7.050482285287744, + "tokens_seen": 9175040 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001398809523809524, + "loss": 5.3131, + "theoretical_loss": 7.041298000670675, + "tokens_seen": 9240576 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014087301587301586, + "loss": 5.3443, + "theoretical_loss": 7.032196714958137, + "tokens_seen": 9306112 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014186507936507937, + "loss": 5.2737, + "theoretical_loss": 7.023177101458123, + "tokens_seen": 9371648 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014285714285714284, + "loss": 5.2961, + "theoretical_loss": 7.014237863773846, + "tokens_seen": 9437184 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014384920634920634, + "loss": 5.2381, + "theoretical_loss": 7.005377734907255, + "tokens_seen": 9502720 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014484126984126984, + "loss": 5.2879, + "theoretical_loss": 6.996595476395081, + "tokens_seen": 9568256 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014583333333333335, + "loss": 5.3349, + "theoretical_loss": 6.98788987747596, + "tokens_seen": 9633792 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014682539682539685, + "loss": 5.2528, + "theoretical_loss": 6.979259754287375, + "tokens_seen": 9699328 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014781746031746032, + "loss": 5.2108, + "theoretical_loss": 6.970703949091095, + "tokens_seen": 9764864 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 11015, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.321908950805664, + "objective/train/theoretical_loss": 6.962221329525966, + "objective/train/tokens_used": 30290400, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014880952380952382, + "loss": 5.3033, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001498015873015873, + "loss": 5.2703, + "theoretical_loss": 6.953810787886883, + "tokens_seen": 9895936 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001507936507936508, + "loss": 5.2393, + "theoretical_loss": 6.945471240428864, + "tokens_seen": 9961472 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015178571428571427, + "loss": 5.262, + "theoretical_loss": 6.937201626695213, + "tokens_seen": 10027008 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001527777777777778, + "loss": 5.2831, + "theoretical_loss": 6.929000908868754, + "tokens_seen": 10092544 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015376984126984128, + "loss": 5.1947, + "theoretical_loss": 6.920868071145226, + "tokens_seen": 10158080 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015476190476190478, + "loss": 5.2418, + "theoretical_loss": 6.912802119127937, + "tokens_seen": 10223616 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015575396825396825, + "loss": 5.3916, + "theoretical_loss": 6.904802079242816, + "tokens_seen": 10289152 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015674603174603175, + "loss": 5.1326, + "theoretical_loss": 6.896866998173063, + "tokens_seen": 10354688 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015773809523809523, + "loss": 5.2225, + "theoretical_loss": 6.888995942312617, + "tokens_seen": 10420224 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015873015873015873, + "loss": 5.2378, + "theoretical_loss": 6.881187997237696, + "tokens_seen": 10485760 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001597222222222222, + "loss": 5.1909, + "theoretical_loss": 6.873442267195708, + "tokens_seen": 10551296 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016071428571428573, + "loss": 5.3429, + "theoretical_loss": 6.8657578746108605, + "tokens_seen": 10616832 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001617063492063492, + "loss": 5.0062, + "theoretical_loss": 6.858133959605815, + "tokens_seen": 10682368 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001626984126984127, + "loss": 5.1487, + "theoretical_loss": 6.8505696795387845, + "tokens_seen": 10747904 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016369047619047618, + "loss": 5.1328, + "theoretical_loss": 6.843064208555461, + "tokens_seen": 10813440 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016468253968253969, + "loss": 5.1871, + "theoretical_loss": 6.835616737155247, + "tokens_seen": 10878976 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016567460317460316, + "loss": 4.999, + "theoretical_loss": 6.828226471771211, + "tokens_seen": 10944512 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016666666666666666, + "loss": 5.0724, + "theoretical_loss": 6.8208926343632905, + "tokens_seen": 11010048 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016765873015873016, + "loss": 4.8736, + "theoretical_loss": 6.813614462024214, + "tokens_seen": 11075584 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016865079365079366, + "loss": 5.1555, + "theoretical_loss": 6.806391206597704, + "tokens_seen": 11141120 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016964285714285717, + "loss": 5.0609, + "theoretical_loss": 6.799222134308489, + "tokens_seen": 11206656 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017063492063492064, + "loss": 5.1907, + "theoretical_loss": 6.792106525403691, + "tokens_seen": 11272192 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017162698412698414, + "loss": 5.012, + "theoretical_loss": 6.785043673805189, + "tokens_seen": 11337728 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017261904761904762, + "loss": 5.0829, + "theoretical_loss": 6.778032886772552, + "tokens_seen": 11403264 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017361111111111112, + "loss": 5.1912, + "theoretical_loss": 6.771073484576156, + "tokens_seen": 11468800 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001746031746031746, + "loss": 5.0558, + "theoretical_loss": 6.764164800180141, + "tokens_seen": 11534336 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001755952380952381, + "loss": 4.9835, + "theoretical_loss": 6.757306178934827, + "tokens_seen": 11599872 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001765873015873016, + "loss": 5.0662, + "theoretical_loss": 6.750496978278292, + "tokens_seen": 11665408 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001775793650793651, + "loss": 5.0258, + "theoretical_loss": 6.743736567446755, + "tokens_seen": 11730944 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017857142857142857, + "loss": 4.8796, + "theoretical_loss": 6.737024327193478, + "tokens_seen": 11796480 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017956349206349207, + "loss": 4.9038, + "theoretical_loss": 6.73035964951589, + "tokens_seen": 11862016 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018055555555555555, + "loss": 5.0887, + "theoretical_loss": 6.723741937390625, + "tokens_seen": 11927552 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018154761904761905, + "loss": 5.1635, + "theoretical_loss": 6.717170604516246, + "tokens_seen": 11993088 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018253968253968252, + "loss": 5.0405, + "theoretical_loss": 6.71064507506334, + "tokens_seen": 12058624 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018353174603174602, + "loss": 5.0501, + "theoretical_loss": 6.704164783431781, + "tokens_seen": 12124160 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018452380952380953, + "loss": 5.012, + "theoretical_loss": 6.6977291740148805, + "tokens_seen": 12189696 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018551587301587303, + "loss": 4.9063, + "theoretical_loss": 6.691337700970224, + "tokens_seen": 12255232 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001865079365079365, + "loss": 5.1759, + "theoretical_loss": 6.684989827996945, + "tokens_seen": 12320768 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001875, + "loss": 5.1608, + "theoretical_loss": 6.6786850281192365, + "tokens_seen": 12386304 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001884920634920635, + "loss": 5.0333, + "theoretical_loss": 6.672422783475897, + "tokens_seen": 12451840 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018948412698412698, + "loss": 5.2021, + "theoretical_loss": 6.666202585115692, + "tokens_seen": 12517376 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019047619047619048, + "loss": 5.0713, + "theoretical_loss": 6.66002393279837, + "tokens_seen": 12582912 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019146825396825398, + "loss": 5.0407, + "theoretical_loss": 6.653886334801115, + "tokens_seen": 12648448 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019246031746031748, + "loss": 4.9208, + "theoretical_loss": 6.647789307730289, + "tokens_seen": 12713984 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019345238095238096, + "loss": 4.9735, + "theoretical_loss": 6.641732376338277, + "tokens_seen": 12779520 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019444444444444446, + "loss": 4.9885, + "theoretical_loss": 6.6357150733452785, + "tokens_seen": 12845056 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019543650793650793, + "loss": 4.963, + "theoretical_loss": 6.629736939265876, + "tokens_seen": 12910592 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019642857142857144, + "loss": 4.9707, + "theoretical_loss": 6.623797522240246, + "tokens_seen": 12976128 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001974206349206349, + "loss": 4.9616, + "theoretical_loss": 6.617896377869851, + "tokens_seen": 13041664 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 12170, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.082708358764648, + "objective/train/theoretical_loss": 6.61203306905748, + "objective/train/tokens_used": 33567200, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001984126984126984, + "loss": 5.0375, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019940476190476191, + "loss": 4.8738, + "theoretical_loss": 6.6062071658514965, + "tokens_seen": 13172736 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020039682539682542, + "loss": 5.0729, + "theoretical_loss": 6.600418245294164, + "tokens_seen": 13238272 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002013888888888889, + "loss": 4.9334, + "theoretical_loss": 6.594665891273929, + "tokens_seen": 13303808 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002023809523809524, + "loss": 5.0793, + "theoretical_loss": 6.588949694381529, + "tokens_seen": 13369344 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020337301587301587, + "loss": 4.9229, + "theoretical_loss": 6.583269251769811, + "tokens_seen": 13434880 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020436507936507937, + "loss": 4.9859, + "theoretical_loss": 6.577624167017157, + "tokens_seen": 13500416 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020535714285714284, + "loss": 4.8612, + "theoretical_loss": 6.5720140499943955, + "tokens_seen": 13565952 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020634920634920634, + "loss": 4.8802, + "theoretical_loss": 6.566438516735092, + "tokens_seen": 13631488 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020734126984126985, + "loss": 4.8757, + "theoretical_loss": 6.560897189309134, + "tokens_seen": 13697024 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020833333333333335, + "loss": 4.835, + "theoretical_loss": 6.5553896956994855, + "tokens_seen": 13762560 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020932539682539685, + "loss": 4.889, + "theoretical_loss": 6.5499156696820435, + "tokens_seen": 13828096 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021031746031746032, + "loss": 4.8896, + "theoretical_loss": 6.544474750708482, + "tokens_seen": 13893632 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021130952380952382, + "loss": 4.874, + "theoretical_loss": 6.539066583792016, + "tokens_seen": 13959168 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002123015873015873, + "loss": 4.9977, + "theoretical_loss": 6.533690819395976, + "tokens_seen": 14024704 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002132936507936508, + "loss": 4.8998, + "theoretical_loss": 6.528347113325134, + "tokens_seen": 14090240 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021428571428571427, + "loss": 4.8949, + "theoretical_loss": 6.523035126619684, + "tokens_seen": 14155776 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002152777777777778, + "loss": 4.9361, + "theoretical_loss": 6.517754525451811, + "tokens_seen": 14221312 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021626984126984128, + "loss": 4.9161, + "theoretical_loss": 6.512504981024765, + "tokens_seen": 14286848 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021726190476190478, + "loss": 4.8396, + "theoretical_loss": 6.507286169474376, + "tokens_seen": 14352384 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021825396825396825, + "loss": 4.9154, + "theoretical_loss": 6.502097771772935, + "tokens_seen": 14417920 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021924603174603176, + "loss": 4.8739, + "theoretical_loss": 6.496939473635374, + "tokens_seen": 14483456 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022023809523809523, + "loss": 4.7639, + "theoretical_loss": 6.491810965427682, + "tokens_seen": 14548992 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022123015873015873, + "loss": 4.8504, + "theoretical_loss": 6.486711942077496, + "tokens_seen": 14614528 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002222222222222222, + "loss": 4.846, + "theoretical_loss": 6.481642102986796, + "tokens_seen": 14680064 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022321428571428573, + "loss": 4.9861, + "theoretical_loss": 6.476601151946659, + "tokens_seen": 14745600 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002242063492063492, + "loss": 4.8742, + "theoretical_loss": 6.471588797054005, + "tokens_seen": 14811136 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002251984126984127, + "loss": 4.8472, + "theoretical_loss": 6.46660475063028, + "tokens_seen": 14876672 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022619047619047618, + "loss": 4.9104, + "theoretical_loss": 6.46164872914203, + "tokens_seen": 14942208 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022718253968253969, + "loss": 4.97, + "theoretical_loss": 6.456720453123305, + "tokens_seen": 15007744 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022817460317460316, + "loss": 4.8924, + "theoretical_loss": 6.4518196470998515, + "tokens_seen": 15073280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022916666666666666, + "loss": 4.7734, + "theoretical_loss": 6.446946039515035, + "tokens_seen": 15138816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023015873015873016, + "loss": 4.7345, + "theoretical_loss": 6.442099362657457, + "tokens_seen": 15204352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023115079365079367, + "loss": 4.5916, + "theoretical_loss": 6.437279352590206, + "tokens_seen": 15269888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023214285714285717, + "loss": 4.7577, + "theoretical_loss": 6.432485749081721, + "tokens_seen": 15335424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023313492063492064, + "loss": 4.7562, + "theoretical_loss": 6.427718295538199, + "tokens_seen": 15400960 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023412698412698414, + "loss": 4.7501, + "theoretical_loss": 6.422976738937518, + "tokens_seen": 15466496 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023511904761904762, + "loss": 4.8521, + "theoretical_loss": 6.418260829764641, + "tokens_seen": 15532032 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023611111111111112, + "loss": 4.9394, + "theoretical_loss": 6.413570321948453, + "tokens_seen": 15597568 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002371031746031746, + "loss": 4.7242, + "theoretical_loss": 6.40890497279999, + "tokens_seen": 15663104 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002380952380952381, + "loss": 4.87, + "theoretical_loss": 6.404264542952033, + "tokens_seen": 15728640 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002390873015873016, + "loss": 4.7524, + "theoretical_loss": 6.399648796300032, + "tokens_seen": 15794176 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002400793650793651, + "loss": 4.7057, + "theoretical_loss": 6.395057499944313, + "tokens_seen": 15859712 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024107142857142857, + "loss": 4.8799, + "theoretical_loss": 6.390490424133546, + "tokens_seen": 15925248 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024206349206349207, + "loss": 4.894, + "theoretical_loss": 6.3859473422094375, + "tokens_seen": 15990784 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024305555555555555, + "loss": 4.925, + "theoretical_loss": 6.381428030552619, + "tokens_seen": 16056320 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024404761904761905, + "loss": 4.886, + "theoretical_loss": 6.376932268529695, + "tokens_seen": 16121856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024503968253968255, + "loss": 4.8673, + "theoretical_loss": 6.372459838441432, + "tokens_seen": 16187392 + }, + { + "epoch": 0.0, + "learning_rate": 0.000246031746031746, + "loss": 4.7909, + "theoretical_loss": 6.368010525472036, + "tokens_seen": 16252928 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024702380952380955, + "loss": 4.843, + "theoretical_loss": 6.363584117639526, + "tokens_seen": 16318464 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 13382, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.550128936767578, + "objective/train/theoretical_loss": 6.35918040574715, + "objective/train/tokens_used": 36844000, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.000248015873015873, + "loss": 4.7236, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002490079365079365, + "loss": 4.7381, + "theoretical_loss": 6.354799183335816, + "tokens_seen": 16449536 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025, + "loss": 4.6256, + "theoretical_loss": 6.350440246637536, + "tokens_seen": 16515072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002509920634920635, + "loss": 4.6346, + "theoretical_loss": 6.3461033945298295, + "tokens_seen": 16580608 + }, + { + "epoch": 0.01, + "learning_rate": 0.000251984126984127, + "loss": 4.7132, + "theoretical_loss": 6.341788428491087, + "tokens_seen": 16646144 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025297619047619046, + "loss": 4.6882, + "theoretical_loss": 6.337495152556851, + "tokens_seen": 16711680 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025396825396825396, + "loss": 4.7217, + "theoretical_loss": 6.333223373277006, + "tokens_seen": 16777216 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025496031746031746, + "loss": 4.903, + "theoretical_loss": 6.32897289967384, + "tokens_seen": 16842752 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025595238095238096, + "loss": 4.7354, + "theoretical_loss": 6.324743543200984, + "tokens_seen": 16908288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002569444444444444, + "loss": 4.7234, + "theoretical_loss": 6.320535117703164, + "tokens_seen": 16973824 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025793650793650796, + "loss": 4.7379, + "theoretical_loss": 6.316347439376797, + "tokens_seen": 17039360 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025892857142857146, + "loss": 4.7087, + "theoretical_loss": 6.312180326731362, + "tokens_seen": 17104896 + }, + { + "epoch": 0.01, + "learning_rate": 0.00025992063492063497, + "loss": 4.7339, + "theoretical_loss": 6.308033600551573, + "tokens_seen": 17170432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002609126984126984, + "loss": 4.4952, + "theoretical_loss": 6.303907083860294, + "tokens_seen": 17235968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002619047619047619, + "loss": 4.8616, + "theoretical_loss": 6.2998006018822075, + "tokens_seen": 17301504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002628968253968254, + "loss": 4.6561, + "theoretical_loss": 6.295713982008203, + "tokens_seen": 17367040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002638888888888889, + "loss": 4.7343, + "theoretical_loss": 6.2916470537604745, + "tokens_seen": 17432576 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026488095238095237, + "loss": 4.6208, + "theoretical_loss": 6.287599648758312, + "tokens_seen": 17498112 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026587301587301587, + "loss": 4.499, + "theoretical_loss": 6.283571600684564, + "tokens_seen": 17563648 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026686507936507937, + "loss": 4.4384, + "theoretical_loss": 6.279562745252756, + "tokens_seen": 17629184 + }, + { + "epoch": 0.01, + "learning_rate": 0.00026785714285714287, + "loss": 4.6564, + "theoretical_loss": 6.2755729201748665, + "tokens_seen": 17694720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002688492063492063, + "loss": 4.7756, + "theoretical_loss": 6.271601965129717, + "tokens_seen": 17760256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002698412698412698, + "loss": 4.7313, + "theoretical_loss": 6.267649721731981, + "tokens_seen": 17825792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002708333333333333, + "loss": 4.6158, + "theoretical_loss": 6.263716033501797, + "tokens_seen": 17891328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002718253968253968, + "loss": 4.5259, + "theoretical_loss": 6.259800745834964, + "tokens_seen": 17956864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002728174603174603, + "loss": 4.6126, + "theoretical_loss": 6.255903705973709, + "tokens_seen": 18022400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002738095238095238, + "loss": 4.6509, + "theoretical_loss": 6.252024762978021, + "tokens_seen": 18087936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002748015873015873, + "loss": 4.5702, + "theoretical_loss": 6.248163767697523, + "tokens_seen": 18153472 + }, + { + "epoch": 0.01, + "learning_rate": 0.00027579365079365083, + "loss": 4.7434, + "theoretical_loss": 6.244320572743888, + "tokens_seen": 18219008 + }, + { + "epoch": 0.01, + "learning_rate": 0.00027678571428571433, + "loss": 4.633, + "theoretical_loss": 6.240495032463768, + "tokens_seen": 18284544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002777777777777778, + "loss": 4.5732, + "theoretical_loss": 6.236687002912243, + "tokens_seen": 18350080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002787698412698413, + "loss": 4.5841, + "theoretical_loss": 6.232896341826766, + "tokens_seen": 18415616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002797619047619048, + "loss": 4.4406, + "theoretical_loss": 6.229122908601596, + "tokens_seen": 18481152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002807539682539683, + "loss": 4.7306, + "theoretical_loss": 6.225366564262705, + "tokens_seen": 18546688 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028174603174603173, + "loss": 4.6097, + "theoretical_loss": 6.2216271714431635, + "tokens_seen": 18612224 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028273809523809523, + "loss": 4.6049, + "theoretical_loss": 6.217904594358968, + "tokens_seen": 18677760 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028373015873015873, + "loss": 4.6094, + "theoretical_loss": 6.214198698785328, + "tokens_seen": 18743296 + }, + { + "epoch": 0.01, + "learning_rate": 0.00028472222222222223, + "loss": 4.5876, + "theoretical_loss": 6.210509352033379, + "tokens_seen": 18808832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002857142857142857, + "loss": 4.7526, + "theoretical_loss": 6.206836422927337, + "tokens_seen": 18874368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002867063492063492, + "loss": 4.7591, + "theoretical_loss": 6.203179781782049, + "tokens_seen": 18939904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002876984126984127, + "loss": 4.6716, + "theoretical_loss": 6.199539300380979, + "tokens_seen": 19005440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002886904761904762, + "loss": 4.67, + "theoretical_loss": 6.195914851954575, + "tokens_seen": 19070976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002896825396825397, + "loss": 4.6089, + "theoretical_loss": 6.19230631115903, + "tokens_seen": 19136512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002906746031746032, + "loss": 4.557, + "theoretical_loss": 6.188713554055438, + "tokens_seen": 19202048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002916666666666667, + "loss": 4.4268, + "theoretical_loss": 6.185136458089309, + "tokens_seen": 19267584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002926587301587302, + "loss": 4.6528, + "theoretical_loss": 6.181574902070463, + "tokens_seen": 19333120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002936507936507937, + "loss": 4.6325, + "theoretical_loss": 6.178028766153277, + "tokens_seen": 19398656 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029464285714285714, + "loss": 4.6518, + "theoretical_loss": 6.174497931817287, + "tokens_seen": 19464192 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029563492063492064, + "loss": 4.6178, + "theoretical_loss": 6.17098228184813, + "tokens_seen": 19529728 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029662698412698414, + "loss": 4.5005, + "theoretical_loss": 6.167481700318827, + "tokens_seen": 19595264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 14656, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.503334045410156, + "objective/train/theoretical_loss": 6.163996072571392, + "objective/train/tokens_used": 40120800, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.00029761904761904765, + "loss": 4.5771, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002986111111111111, + "loss": 4.5788, + "theoretical_loss": 6.160525285198771, + "tokens_seen": 19726336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0002996031746031746, + "loss": 4.5308, + "theoretical_loss": 6.157069226027087, + "tokens_seen": 19791872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003005952380952381, + "loss": 4.4183, + "theoretical_loss": 6.153627784098204, + "tokens_seen": 19857408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003015873015873016, + "loss": 4.5105, + "theoretical_loss": 6.150200849652586, + "tokens_seen": 19922944 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030257936507936505, + "loss": 4.5325, + "theoretical_loss": 6.146788314112464, + "tokens_seen": 19988480 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030357142857142855, + "loss": 4.4606, + "theoretical_loss": 6.143390070065279, + "tokens_seen": 20054016 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030456349206349205, + "loss": 4.5863, + "theoretical_loss": 6.140006011247426, + "tokens_seen": 20119552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003055555555555556, + "loss": 4.6739, + "theoretical_loss": 6.136636032528262, + "tokens_seen": 20185088 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030654761904761905, + "loss": 4.5882, + "theoretical_loss": 6.133280029894401, + "tokens_seen": 20250624 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030753968253968255, + "loss": 4.589, + "theoretical_loss": 6.129937900434269, + "tokens_seen": 20316160 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030853174603174605, + "loss": 4.4872, + "theoretical_loss": 6.126609542322921, + "tokens_seen": 20381696 + }, + { + "epoch": 0.01, + "learning_rate": 0.00030952380952380956, + "loss": 4.493, + "theoretical_loss": 6.123294854807119, + "tokens_seen": 20447232 + }, + { + "epoch": 0.01, + "learning_rate": 0.000310515873015873, + "loss": 4.6428, + "theoretical_loss": 6.11999373819066, + "tokens_seen": 20512768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003115079365079365, + "loss": 4.328, + "theoretical_loss": 6.116706093819951, + "tokens_seen": 20578304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003125, + "loss": 4.5559, + "theoretical_loss": 6.113431824069829, + "tokens_seen": 20643840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003134920634920635, + "loss": 4.5108, + "theoretical_loss": 6.110170832329612, + "tokens_seen": 20709376 + }, + { + "epoch": 0.01, + "learning_rate": 0.000314484126984127, + "loss": 4.5167, + "theoretical_loss": 6.10692302298939, + "tokens_seen": 20774912 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031547619047619046, + "loss": 4.4875, + "theoretical_loss": 6.103688301426538, + "tokens_seen": 20840448 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031646825396825396, + "loss": 4.6132, + "theoretical_loss": 6.100466573992453, + "tokens_seen": 20905984 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031746031746031746, + "loss": 4.5981, + "theoretical_loss": 6.097257747999514, + "tokens_seen": 20971520 + }, + { + "epoch": 0.01, + "learning_rate": 0.00031845238095238096, + "loss": 4.5378, + "theoretical_loss": 6.094061731708257, + "tokens_seen": 21037056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003194444444444444, + "loss": 4.4524, + "theoretical_loss": 6.090878434314753, + "tokens_seen": 21102592 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032043650793650796, + "loss": 4.4526, + "theoretical_loss": 6.0877077659382, + "tokens_seen": 21168128 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032142857142857147, + "loss": 4.5552, + "theoretical_loss": 6.084549637608715, + "tokens_seen": 21233664 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032242063492063497, + "loss": 4.5234, + "theoretical_loss": 6.081403961255324, + "tokens_seen": 21299200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003234126984126984, + "loss": 4.5241, + "theoretical_loss": 6.078270649694149, + "tokens_seen": 21364736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003244047619047619, + "loss": 4.4889, + "theoretical_loss": 6.07514961661678, + "tokens_seen": 21430272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003253968253968254, + "loss": 4.3799, + "theoretical_loss": 6.072040776578845, + "tokens_seen": 21495808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003263888888888889, + "loss": 4.5607, + "theoretical_loss": 6.068944044988752, + "tokens_seen": 21561344 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032738095238095237, + "loss": 4.4685, + "theoretical_loss": 6.065859338096619, + "tokens_seen": 21626880 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032837301587301587, + "loss": 4.3811, + "theoretical_loss": 6.0627865729833745, + "tokens_seen": 21692416 + }, + { + "epoch": 0.01, + "learning_rate": 0.00032936507936507937, + "loss": 4.5747, + "theoretical_loss": 6.059725667550044, + "tokens_seen": 21757952 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033035714285714287, + "loss": 4.3786, + "theoretical_loss": 6.0566765405071905, + "tokens_seen": 21823488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003313492063492063, + "loss": 4.4018, + "theoretical_loss": 6.0536391113645305, + "tokens_seen": 21889024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003323412698412698, + "loss": 4.4855, + "theoretical_loss": 6.050613300420713, + "tokens_seen": 21954560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003333333333333333, + "loss": 4.3616, + "theoretical_loss": 6.047599028753265, + "tokens_seen": 22020096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003343253968253968, + "loss": 4.5438, + "theoretical_loss": 6.04459621820868, + "tokens_seen": 22085632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003353174603174603, + "loss": 4.4438, + "theoretical_loss": 6.041604791392675, + "tokens_seen": 22151168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003363095238095238, + "loss": 4.4097, + "theoretical_loss": 6.038624671660598, + "tokens_seen": 22216704 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033730158730158733, + "loss": 4.3068, + "theoretical_loss": 6.035655783107973, + "tokens_seen": 22282240 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033829365079365083, + "loss": 4.3482, + "theoretical_loss": 6.0326980505612084, + "tokens_seen": 22347776 + }, + { + "epoch": 0.01, + "learning_rate": 0.00033928571428571433, + "loss": 4.3094, + "theoretical_loss": 6.029751399568434, + "tokens_seen": 22413312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003402777777777778, + "loss": 4.3689, + "theoretical_loss": 6.026815756390491, + "tokens_seen": 22478848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003412698412698413, + "loss": 4.4719, + "theoretical_loss": 6.023891047992053, + "tokens_seen": 22544384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003422619047619048, + "loss": 4.4678, + "theoretical_loss": 6.020977202032884, + "tokens_seen": 22609920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003432539682539683, + "loss": 4.3266, + "theoretical_loss": 6.018074146859234, + "tokens_seen": 22675456 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034424603174603173, + "loss": 4.4691, + "theoretical_loss": 6.015181811495358, + "tokens_seen": 22740992 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034523809523809523, + "loss": 4.3061, + "theoretical_loss": 6.012300125635177, + "tokens_seen": 22806528 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034623015873015873, + "loss": 4.3645, + "theoretical_loss": 6.009429019634049, + "tokens_seen": 22872064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 15907, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.475631237030029, + "objective/train/theoretical_loss": 6.006568424500674, + "objective/train/tokens_used": 43397600, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.00034722222222222224, + "loss": 4.436, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003482142857142857, + "loss": 4.1448, + "theoretical_loss": 6.003718271889123, + "tokens_seen": 23003136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003492063492063492, + "loss": 4.3228, + "theoretical_loss": 6.000878494090979, + "tokens_seen": 23068672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003501984126984127, + "loss": 4.4723, + "theoretical_loss": 5.998049024027599, + "tokens_seen": 23134208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003511904761904762, + "loss": 4.4769, + "theoretical_loss": 5.995229795242498, + "tokens_seen": 23199744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003521825396825397, + "loss": 4.3468, + "theoretical_loss": 5.99242074189383, + "tokens_seen": 23265280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003531746031746032, + "loss": 4.2293, + "theoretical_loss": 5.989621798747002, + "tokens_seen": 23330816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003541666666666667, + "loss": 4.274, + "theoretical_loss": 5.98683290116738, + "tokens_seen": 23396352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003551587301587302, + "loss": 4.2337, + "theoretical_loss": 5.9840539851131105, + "tokens_seen": 23461888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003561507936507937, + "loss": 4.3068, + "theoretical_loss": 5.981284987128044, + "tokens_seen": 23527424 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035714285714285714, + "loss": 4.2633, + "theoretical_loss": 5.97852584433477, + "tokens_seen": 23592960 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035813492063492064, + "loss": 4.2789, + "theoretical_loss": 5.975776494427746, + "tokens_seen": 23658496 + }, + { + "epoch": 0.01, + "learning_rate": 0.00035912698412698415, + "loss": 4.3126, + "theoretical_loss": 5.973036875666525, + "tokens_seen": 23724032 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036011904761904765, + "loss": 4.3333, + "theoretical_loss": 5.970306926869093, + "tokens_seen": 23789568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003611111111111111, + "loss": 4.4774, + "theoretical_loss": 5.967586587405295, + "tokens_seen": 23855104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003621031746031746, + "loss": 4.2515, + "theoretical_loss": 5.96487579719035, + "tokens_seen": 23920640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003630952380952381, + "loss": 4.2325, + "theoretical_loss": 5.962174496678481, + "tokens_seen": 23986176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003640873015873016, + "loss": 4.1292, + "theoretical_loss": 5.959482626856606, + "tokens_seen": 24051712 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036507936507936505, + "loss": 4.4078, + "theoretical_loss": 5.956800129238152, + "tokens_seen": 24117248 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036607142857142855, + "loss": 4.2105, + "theoretical_loss": 5.954126945856933, + "tokens_seen": 24182784 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036706349206349205, + "loss": 4.2613, + "theoretical_loss": 5.951463019261125, + "tokens_seen": 24248320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003680555555555556, + "loss": 4.1763, + "theoretical_loss": 5.948808292507333, + "tokens_seen": 24313856 + }, + { + "epoch": 0.01, + "learning_rate": 0.00036904761904761905, + "loss": 4.1927, + "theoretical_loss": 5.946162709154724, + "tokens_seen": 24379392 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037003968253968255, + "loss": 4.2673, + "theoretical_loss": 5.943526213259268, + "tokens_seen": 24444928 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037103174603174606, + "loss": 4.2252, + "theoretical_loss": 5.940898749368037, + "tokens_seen": 24510464 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037202380952380956, + "loss": 4.401, + "theoretical_loss": 5.9382802625136, + "tokens_seen": 24576000 + }, + { + "epoch": 0.01, + "learning_rate": 0.000373015873015873, + "loss": 4.229, + "theoretical_loss": 5.935670698208488, + "tokens_seen": 24641536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003740079365079365, + "loss": 4.2107, + "theoretical_loss": 5.933070002439743, + "tokens_seen": 24707072 + }, + { + "epoch": 0.01, + "learning_rate": 0.000375, + "loss": 4.2292, + "theoretical_loss": 5.930478121663538, + "tokens_seen": 24772608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003759920634920635, + "loss": 4.2642, + "theoretical_loss": 5.927895002799874, + "tokens_seen": 24838144 + }, + { + "epoch": 0.01, + "learning_rate": 0.000376984126984127, + "loss": 4.3279, + "theoretical_loss": 5.925320593227351, + "tokens_seen": 24903680 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037797619047619046, + "loss": 4.1869, + "theoretical_loss": 5.922754840778015, + "tokens_seen": 24969216 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037896825396825396, + "loss": 4.1617, + "theoretical_loss": 5.9201976937322645, + "tokens_seen": 25034752 + }, + { + "epoch": 0.01, + "learning_rate": 0.00037996031746031746, + "loss": 4.3255, + "theoretical_loss": 5.917649100813846, + "tokens_seen": 25100288 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038095238095238096, + "loss": 4.1971, + "theoretical_loss": 5.915109011184899, + "tokens_seen": 25165824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003819444444444444, + "loss": 4.2252, + "theoretical_loss": 5.9125773744410814, + "tokens_seen": 25231360 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038293650793650797, + "loss": 4.2593, + "theoretical_loss": 5.91005414060676, + "tokens_seen": 25296896 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038392857142857147, + "loss": 4.3073, + "theoretical_loss": 5.90753926013026, + "tokens_seen": 25362432 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038492063492063497, + "loss": 4.2613, + "theoretical_loss": 5.90503268387919, + "tokens_seen": 25427968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003859126984126984, + "loss": 4.1771, + "theoretical_loss": 5.9025343631358185, + "tokens_seen": 25493504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003869047619047619, + "loss": 4.3719, + "theoretical_loss": 5.900044249592523, + "tokens_seen": 25559040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003878968253968254, + "loss": 4.33, + "theoretical_loss": 5.897562295347297, + "tokens_seen": 25624576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003888888888888889, + "loss": 4.1681, + "theoretical_loss": 5.895088452899312, + "tokens_seen": 25690112 + }, + { + "epoch": 0.01, + "learning_rate": 0.00038988095238095237, + "loss": 4.2915, + "theoretical_loss": 5.892622675144553, + "tokens_seen": 25755648 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039087301587301587, + "loss": 4.1884, + "theoretical_loss": 5.890164915371495, + "tokens_seen": 25821184 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039186507936507937, + "loss": 4.2524, + "theoretical_loss": 5.887715127256854, + "tokens_seen": 25886720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003928571428571429, + "loss": 4.2715, + "theoretical_loss": 5.885273264861381, + "tokens_seen": 25952256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003938492063492063, + "loss": 4.3223, + "theoretical_loss": 5.882839282625721, + "tokens_seen": 26017792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003948412698412698, + "loss": 4.1229, + "theoretical_loss": 5.880413135366323, + "tokens_seen": 26083328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003958333333333333, + "loss": 4.068, + "theoretical_loss": 5.877994778271411, + "tokens_seen": 26148864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 17033, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.266489028930664, + "objective/train/theoretical_loss": 5.8755841668969895, + "objective/train/tokens_used": 46674400, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003968253968253968, + "loss": 4.3211, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0003978174603174603, + "loss": 4.0408, + "theoretical_loss": 5.873181257162926, + "tokens_seen": 26279936 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039880952380952383, + "loss": 4.1185, + "theoretical_loss": 5.870786005349071, + "tokens_seen": 26345472 + }, + { + "epoch": 0.01, + "learning_rate": 0.00039980158730158733, + "loss": 4.1377, + "theoretical_loss": 5.868398368091427, + "tokens_seen": 26411008 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040079365079365083, + "loss": 4.1952, + "theoretical_loss": 5.866018302378377, + "tokens_seen": 26476544 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040178571428571433, + "loss": 4.0419, + "theoretical_loss": 5.863645765546956, + "tokens_seen": 26542080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004027777777777778, + "loss": 4.1441, + "theoretical_loss": 5.861280715279168, + "tokens_seen": 26607616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004037698412698413, + "loss": 4.0659, + "theoretical_loss": 5.858923109598366, + "tokens_seen": 26673152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004047619047619048, + "loss": 4.0085, + "theoretical_loss": 5.8565729068656545, + "tokens_seen": 26738688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004057539682539683, + "loss": 4.1483, + "theoretical_loss": 5.85423006577636, + "tokens_seen": 26804224 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040674603174603173, + "loss": 4.2296, + "theoretical_loss": 5.851894545356539, + "tokens_seen": 26869760 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040773809523809523, + "loss": 4.0618, + "theoretical_loss": 5.849566304959527, + "tokens_seen": 26935296 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040873015873015874, + "loss": 3.8736, + "theoretical_loss": 5.84724530426254, + "tokens_seen": 27000832 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040972222222222224, + "loss": 4.0506, + "theoretical_loss": 5.844931503263309, + "tokens_seen": 27066368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004107142857142857, + "loss": 4.1426, + "theoretical_loss": 5.842624862276773, + "tokens_seen": 27131904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004117063492063492, + "loss": 4.0093, + "theoretical_loss": 5.840325341931795, + "tokens_seen": 27197440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004126984126984127, + "loss": 3.8588, + "theoretical_loss": 5.838032903167934, + "tokens_seen": 27262976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004136904761904762, + "loss": 4.1939, + "theoretical_loss": 5.835747507232257, + "tokens_seen": 27328512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004146825396825397, + "loss": 4.1288, + "theoretical_loss": 5.833469115676177, + "tokens_seen": 27394048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004156746031746032, + "loss": 4.1349, + "theoretical_loss": 5.8311976903523535, + "tokens_seen": 27459584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004166666666666667, + "loss": 4.1272, + "theoretical_loss": 5.828933193411611, + "tokens_seen": 27525120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004176587301587302, + "loss": 4.0323, + "theoretical_loss": 5.826675587299913, + "tokens_seen": 27590656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004186507936507937, + "loss": 4.1329, + "theoretical_loss": 5.824424834755357, + "tokens_seen": 27656192 + }, + { + "epoch": 0.01, + "learning_rate": 0.00041964285714285714, + "loss": 3.9974, + "theoretical_loss": 5.8221808988052235, + "tokens_seen": 27721728 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042063492063492065, + "loss": 4.0567, + "theoretical_loss": 5.819943742763051, + "tokens_seen": 27787264 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042162698412698415, + "loss": 4.0377, + "theoretical_loss": 5.817713330225754, + "tokens_seen": 27852800 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042261904761904765, + "loss": 4.0387, + "theoretical_loss": 5.815489625070766, + "tokens_seen": 27918336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004236111111111111, + "loss": 3.9902, + "theoretical_loss": 5.813272591453236, + "tokens_seen": 27983872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004246031746031746, + "loss": 3.9238, + "theoretical_loss": 5.811062193803242, + "tokens_seen": 28049408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004255952380952381, + "loss": 4.0464, + "theoretical_loss": 5.808858396823043, + "tokens_seen": 28114944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004265873015873016, + "loss": 3.9472, + "theoretical_loss": 5.806661165484377, + "tokens_seen": 28180480 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042757936507936505, + "loss": 4.1283, + "theoretical_loss": 5.804470465025777, + "tokens_seen": 28246016 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042857142857142855, + "loss": 4.087, + "theoretical_loss": 5.802286260949924, + "tokens_seen": 28311552 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042956349206349205, + "loss": 4.0772, + "theoretical_loss": 5.8001085190210375, + "tokens_seen": 28377088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004305555555555556, + "loss": 4.1279, + "theoretical_loss": 5.797937205262292, + "tokens_seen": 28442624 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043154761904761905, + "loss": 3.9086, + "theoretical_loss": 5.795772285953266, + "tokens_seen": 28508160 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043253968253968256, + "loss": 3.9767, + "theoretical_loss": 5.793613727627429, + "tokens_seen": 28573696 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043353174603174606, + "loss": 3.7997, + "theoretical_loss": 5.791461497069644, + "tokens_seen": 28639232 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043452380952380956, + "loss": 4.0363, + "theoretical_loss": 5.789315561313717, + "tokens_seen": 28704768 + }, + { + "epoch": 0.01, + "learning_rate": 0.000435515873015873, + "loss": 4.0149, + "theoretical_loss": 5.787175887639959, + "tokens_seen": 28770304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004365079365079365, + "loss": 4.068, + "theoretical_loss": 5.785042443572798, + "tokens_seen": 28835840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004375, + "loss": 3.926, + "theoretical_loss": 5.782915196878396, + "tokens_seen": 28901376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004384920634920635, + "loss": 3.9741, + "theoretical_loss": 5.780794115562314, + "tokens_seen": 28966912 + }, + { + "epoch": 0.01, + "learning_rate": 0.000439484126984127, + "loss": 4.1026, + "theoretical_loss": 5.778679167867196, + "tokens_seen": 29032448 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044047619047619046, + "loss": 3.9757, + "theoretical_loss": 5.7765703222704765, + "tokens_seen": 29097984 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044146825396825396, + "loss": 4.0369, + "theoretical_loss": 5.7744675474821285, + "tokens_seen": 29163520 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044246031746031746, + "loss": 4.0837, + "theoretical_loss": 5.772370812442426, + "tokens_seen": 29229056 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044345238095238096, + "loss": 3.947, + "theoretical_loss": 5.770280086319733, + "tokens_seen": 29294592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004444444444444444, + "loss": 4.0545, + "theoretical_loss": 5.768195338508333, + "tokens_seen": 29360128 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044543650793650797, + "loss": 3.894, + "theoretical_loss": 5.766116538626263, + "tokens_seen": 29425664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 18174, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.157195091247559, + "objective/train/theoretical_loss": 5.7640436565131905, + "objective/train/tokens_used": 49951200, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044642857142857147, + "loss": 4.0514, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044742063492063497, + "loss": 3.9977, + "theoretical_loss": 5.761976662228307, + "tokens_seen": 29556736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004484126984126984, + "loss": 3.8908, + "theoretical_loss": 5.759915526048243, + "tokens_seen": 29622272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004494047619047619, + "loss": 3.9623, + "theoretical_loss": 5.757860218465023, + "tokens_seen": 29687808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004503968253968254, + "loss": 4.0131, + "theoretical_loss": 5.755810710184019, + "tokens_seen": 29753344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004513888888888889, + "loss": 3.8964, + "theoretical_loss": 5.753766972121954, + "tokens_seen": 29818880 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045238095238095237, + "loss": 4.0609, + "theoretical_loss": 5.751728975404911, + "tokens_seen": 29884416 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045337301587301587, + "loss": 4.0626, + "theoretical_loss": 5.7496966913663705, + "tokens_seen": 29949952 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045436507936507937, + "loss": 3.9034, + "theoretical_loss": 5.747670091545274, + "tokens_seen": 30015488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004553571428571429, + "loss": 3.9837, + "theoretical_loss": 5.745649147684103, + "tokens_seen": 30081024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004563492063492063, + "loss": 4.0769, + "theoretical_loss": 5.743633831726987, + "tokens_seen": 30146560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004573412698412698, + "loss": 3.9591, + "theoretical_loss": 5.741624115817827, + "tokens_seen": 30212096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004583333333333333, + "loss": 3.9491, + "theoretical_loss": 5.739619972298446, + "tokens_seen": 30277632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004593253968253968, + "loss": 4.0368, + "theoretical_loss": 5.737621373706753, + "tokens_seen": 30343168 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046031746031746033, + "loss": 3.9124, + "theoretical_loss": 5.735628292774945, + "tokens_seen": 30408704 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046130952380952383, + "loss": 3.8811, + "theoretical_loss": 5.7336407024276985, + "tokens_seen": 30474240 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046230158730158733, + "loss": 3.8806, + "theoretical_loss": 5.731658575780418, + "tokens_seen": 30539776 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046329365079365083, + "loss": 3.863, + "theoretical_loss": 5.729681886137476, + "tokens_seen": 30605312 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046428571428571433, + "loss": 3.7816, + "theoretical_loss": 5.727710606990488, + "tokens_seen": 30670848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004652777777777778, + "loss": 3.9672, + "theoretical_loss": 5.725744712016601, + "tokens_seen": 30736384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004662698412698413, + "loss": 3.905, + "theoretical_loss": 5.723784175076802, + "tokens_seen": 30801920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004672619047619048, + "loss": 3.8629, + "theoretical_loss": 5.72182897021425, + "tokens_seen": 30867456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004682539682539683, + "loss": 3.7489, + "theoretical_loss": 5.719879071652615, + "tokens_seen": 30932992 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046924603174603173, + "loss": 3.9044, + "theoretical_loss": 5.717934453794454, + "tokens_seen": 30998528 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047023809523809523, + "loss": 3.9157, + "theoretical_loss": 5.715995091219584, + "tokens_seen": 31064064 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047123015873015874, + "loss": 3.8305, + "theoretical_loss": 5.7140609586834925, + "tokens_seen": 31129600 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047222222222222224, + "loss": 3.8432, + "theoretical_loss": 5.712132031115749, + "tokens_seen": 31195136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004732142857142857, + "loss": 3.8101, + "theoretical_loss": 5.710208283618447, + "tokens_seen": 31260672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004742063492063492, + "loss": 4.0189, + "theoretical_loss": 5.708289691464657, + "tokens_seen": 31326208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004751984126984127, + "loss": 4.0465, + "theoretical_loss": 5.706376230096897, + "tokens_seen": 31391744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004761904761904762, + "loss": 3.8214, + "theoretical_loss": 5.704467875125619, + "tokens_seen": 31457280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004771825396825397, + "loss": 3.8767, + "theoretical_loss": 5.702564602327719, + "tokens_seen": 31522816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004781746031746032, + "loss": 3.748, + "theoretical_loss": 5.700666387645049, + "tokens_seen": 31588352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004791666666666667, + "loss": 4.0285, + "theoretical_loss": 5.698773207182963, + "tokens_seen": 31653888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004801587301587302, + "loss": 3.8935, + "theoretical_loss": 5.696885037208865, + "tokens_seen": 31719424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004811507936507937, + "loss": 3.852, + "theoretical_loss": 5.695001854150782, + "tokens_seen": 31784960 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048214285714285715, + "loss": 3.7945, + "theoretical_loss": 5.693123634595938, + "tokens_seen": 31850496 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048313492063492065, + "loss": 3.901, + "theoretical_loss": 5.69125035528937, + "tokens_seen": 31916032 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048412698412698415, + "loss": 3.854, + "theoretical_loss": 5.689381993132532, + "tokens_seen": 31981568 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048511904761904765, + "loss": 3.8865, + "theoretical_loss": 5.687518525181924, + "tokens_seen": 32047104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004861111111111111, + "loss": 3.8573, + "theoretical_loss": 5.685659928647743, + "tokens_seen": 32112640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004871031746031746, + "loss": 4.0027, + "theoretical_loss": 5.68380618089254, + "tokens_seen": 32178176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004880952380952381, + "loss": 3.8024, + "theoretical_loss": 5.6819572594298915, + "tokens_seen": 32243712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004890873015873016, + "loss": 3.7735, + "theoretical_loss": 5.680113141923088, + "tokens_seen": 32309248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004900793650793651, + "loss": 3.946, + "theoretical_loss": 5.67827380618384, + "tokens_seen": 32374784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004910714285714286, + "loss": 3.8417, + "theoretical_loss": 5.6764392301709865, + "tokens_seen": 32440320 + }, + { + "epoch": 0.01, + "learning_rate": 0.000492063492063492, + "loss": 3.749, + "theoretical_loss": 5.674609391989232, + "tokens_seen": 32505856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004930555555555556, + "loss": 3.9908, + "theoretical_loss": 5.672784269887885, + "tokens_seen": 32571392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004940476190476191, + "loss": 3.9758, + "theoretical_loss": 5.670963842259615, + "tokens_seen": 32636928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004950396825396826, + "loss": 3.995, + "theoretical_loss": 5.669148087639222, + "tokens_seen": 32702464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 19239, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.939499855041504, + "objective/train/theoretical_loss": 5.66733698470242, + "objective/train/tokens_used": 53228000, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.000496031746031746, + "loss": 3.9324, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004970238095238095, + "loss": 3.8437, + "theoretical_loss": 5.665530512264631, + "tokens_seen": 32833536 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498015873015873, + "loss": 3.8955, + "theoretical_loss": 5.663728649279793, + "tokens_seen": 32899072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990079365079365, + "loss": 3.7376, + "theoretical_loss": 5.6619313748391855, + "tokens_seen": 32964608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005, + "loss": 3.9219, + "theoretical_loss": 5.6601386681702515, + "tokens_seen": 33030144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999899699097292, + "loss": 3.8692, + "theoretical_loss": 5.658350508635454, + "tokens_seen": 33095680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999799398194584, + "loss": 3.7663, + "theoretical_loss": 5.656566875731127, + "tokens_seen": 33161216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999699097291876, + "loss": 3.7634, + "theoretical_loss": 5.654787749086345, + "tokens_seen": 33226752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999598796389167, + "loss": 3.8123, + "theoretical_loss": 5.6530131084618045, + "tokens_seen": 33292288 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499949849548646, + "loss": 3.8034, + "theoretical_loss": 5.651242933748718, + "tokens_seen": 33357824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999398194583751, + "loss": 3.8073, + "theoretical_loss": 5.649477204967717, + "tokens_seen": 33423360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999297893681044, + "loss": 3.9116, + "theoretical_loss": 5.647715902267764, + "tokens_seen": 33488896 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999197592778335, + "loss": 3.8111, + "theoretical_loss": 5.645959005925086, + "tokens_seen": 33554432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004999097291875627, + "loss": 3.7961, + "theoretical_loss": 5.644206496342103, + "tokens_seen": 33619968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998996990972919, + "loss": 3.8473, + "theoretical_loss": 5.642458354046386, + "tokens_seen": 33685504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998896690070211, + "loss": 3.7876, + "theoretical_loss": 5.6407145596896076, + "tokens_seen": 33751040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998796389167503, + "loss": 3.8586, + "theoretical_loss": 5.638975094046518, + "tokens_seen": 33816576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998696088264795, + "loss": 3.6675, + "theoretical_loss": 5.637239938013924, + "tokens_seen": 33882112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998595787362087, + "loss": 3.7223, + "theoretical_loss": 5.6355090726096755, + "tokens_seen": 33947648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998495486459378, + "loss": 3.7047, + "theoretical_loss": 5.633782478971672, + "tokens_seen": 34013184 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499839518555667, + "loss": 3.759, + "theoretical_loss": 5.632060138356873, + "tokens_seen": 34078720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998294884653962, + "loss": 3.6811, + "theoretical_loss": 5.630342032140315, + "tokens_seen": 34144256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998194583751254, + "loss": 3.8712, + "theoretical_loss": 5.628628141814151, + "tokens_seen": 34209792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004998094282848546, + "loss": 3.7287, + "theoretical_loss": 5.62691844898668, + "tokens_seen": 34275328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997993981945837, + "loss": 3.6715, + "theoretical_loss": 5.625212935381407, + "tokens_seen": 34340864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997893681043129, + "loss": 3.6806, + "theoretical_loss": 5.6235115828361035, + "tokens_seen": 34406400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997793380140421, + "loss": 3.5933, + "theoretical_loss": 5.621814373301866, + "tokens_seen": 34471936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997693079237714, + "loss": 3.7875, + "theoretical_loss": 5.620121288842209, + "tokens_seen": 34537472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997592778335005, + "loss": 3.722, + "theoretical_loss": 5.618432311632143, + "tokens_seen": 34603008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997492477432298, + "loss": 3.689, + "theoretical_loss": 5.616747423957278, + "tokens_seen": 34668544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997392176529588, + "loss": 3.697, + "theoretical_loss": 5.615066608212926, + "tokens_seen": 34734080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997291875626881, + "loss": 3.7821, + "theoretical_loss": 5.613389846903216, + "tokens_seen": 34799616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997191574724173, + "loss": 3.7247, + "theoretical_loss": 5.611717122640225, + "tokens_seen": 34865152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004997091273821465, + "loss": 3.6408, + "theoretical_loss": 5.610048418143098, + "tokens_seen": 34930688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996990972918757, + "loss": 3.6505, + "theoretical_loss": 5.6083837162372, + "tokens_seen": 34996224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996890672016048, + "loss": 3.7016, + "theoretical_loss": 5.606722999853261, + "tokens_seen": 35061760 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499679037111334, + "loss": 3.6174, + "theoretical_loss": 5.605066252026537, + "tokens_seen": 35127296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996690070210632, + "loss": 3.708, + "theoretical_loss": 5.603413455895973, + "tokens_seen": 35192832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996589769307924, + "loss": 3.6511, + "theoretical_loss": 5.601764594703377, + "tokens_seen": 35258368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996489468405216, + "loss": 3.4649, + "theoretical_loss": 5.600119651792607, + "tokens_seen": 35323904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996389167502507, + "loss": 3.566, + "theoretical_loss": 5.598478610608758, + "tokens_seen": 35389440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996288866599799, + "loss": 3.5815, + "theoretical_loss": 5.596841454697365, + "tokens_seen": 35454976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996188565697091, + "loss": 3.7303, + "theoretical_loss": 5.595208167703601, + "tokens_seen": 35520512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004996088264794383, + "loss": 3.6208, + "theoretical_loss": 5.5935787333714995, + "tokens_seen": 35586048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995987963891675, + "loss": 3.9222, + "theoretical_loss": 5.59195313554317, + "tokens_seen": 35651584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995887662988968, + "loss": 3.7724, + "theoretical_loss": 5.590331358158036, + "tokens_seen": 35717120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995787362086258, + "loss": 3.8137, + "theoretical_loss": 5.58871338525206, + "tokens_seen": 35782656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995687061183551, + "loss": 3.6252, + "theoretical_loss": 5.587099200956995, + "tokens_seen": 35848192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995586760280842, + "loss": 3.8119, + "theoretical_loss": 5.585488789499637, + "tokens_seen": 35913728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995486459378135, + "loss": 3.8026, + "theoretical_loss": 5.583882135201078, + "tokens_seen": 35979264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 20470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.579683303833008, + "objective/train/theoretical_loss": 5.582279222475973, + "objective/train/tokens_used": 56504800, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995386158475427, + "loss": 3.5523, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995285857572718, + "loss": 3.8757, + "theoretical_loss": 5.580680035831821, + "tokens_seen": 36110336 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499518555667001, + "loss": 3.6668, + "theoretical_loss": 5.579084559868232, + "tokens_seen": 36175872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004995085255767302, + "loss": 3.7736, + "theoretical_loss": 5.577492779276222, + "tokens_seen": 36241408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994984954864594, + "loss": 3.6003, + "theoretical_loss": 5.575904678837501, + "tokens_seen": 36306944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994884653961886, + "loss": 3.7828, + "theoretical_loss": 5.574320243423783, + "tokens_seen": 36372480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994784353059178, + "loss": 3.5222, + "theoretical_loss": 5.572739457996079, + "tokens_seen": 36438016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994684052156469, + "loss": 3.5318, + "theoretical_loss": 5.5711623076040215, + "tokens_seen": 36503552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994583751253761, + "loss": 3.8119, + "theoretical_loss": 5.569588777385177, + "tokens_seen": 36569088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994483450351053, + "loss": 3.6884, + "theoretical_loss": 5.568018852564379, + "tokens_seen": 36634624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994383149448345, + "loss": 3.6669, + "theoretical_loss": 5.566452518453051, + "tokens_seen": 36700160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994282848545637, + "loss": 3.4946, + "theoretical_loss": 5.564889760448561, + "tokens_seen": 36765696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994182547642928, + "loss": 3.6685, + "theoretical_loss": 5.5633305640335475, + "tokens_seen": 36831232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004994082246740221, + "loss": 3.6604, + "theoretical_loss": 5.561774914775286, + "tokens_seen": 36896768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993981945837512, + "loss": 3.6834, + "theoretical_loss": 5.5602227983250385, + "tokens_seen": 36962304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993881644934805, + "loss": 3.6271, + "theoretical_loss": 5.558674200417419, + "tokens_seen": 37027840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993781344032096, + "loss": 3.7549, + "theoretical_loss": 5.557129106869764, + "tokens_seen": 37093376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993681043129389, + "loss": 3.6475, + "theoretical_loss": 5.555587503581506, + "tokens_seen": 37158912 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499358074222668, + "loss": 3.7243, + "theoretical_loss": 5.554049376533554, + "tokens_seen": 37224448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993480441323972, + "loss": 3.6905, + "theoretical_loss": 5.552514711787683, + "tokens_seen": 37289984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993380140421264, + "loss": 3.6042, + "theoretical_loss": 5.550983495485925, + "tokens_seen": 37355520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993279839518556, + "loss": 3.8043, + "theoretical_loss": 5.549455713849963, + "tokens_seen": 37421056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993179538615848, + "loss": 3.7664, + "theoretical_loss": 5.547931353180543, + "tokens_seen": 37486592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004993079237713139, + "loss": 3.6224, + "theoretical_loss": 5.546410399856875, + "tokens_seen": 37552128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992978936810431, + "loss": 3.6067, + "theoretical_loss": 5.544892840336053, + "tokens_seen": 37617664 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992878635907723, + "loss": 3.6362, + "theoretical_loss": 5.543378661152472, + "tokens_seen": 37683200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992778335005015, + "loss": 3.6053, + "theoretical_loss": 5.5418678489172555, + "tokens_seen": 37748736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992678034102307, + "loss": 3.6984, + "theoretical_loss": 5.540360390317682, + "tokens_seen": 37814272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992577733199598, + "loss": 3.5072, + "theoretical_loss": 5.538856272116626, + "tokens_seen": 37879808 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499247743229689, + "loss": 3.482, + "theoretical_loss": 5.5373554811519945, + "tokens_seen": 37945344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992377131394183, + "loss": 3.5633, + "theoretical_loss": 5.535858004336175, + "tokens_seen": 38010880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992276830491475, + "loss": 3.6206, + "theoretical_loss": 5.534363828655488, + "tokens_seen": 38076416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992176529588767, + "loss": 3.5855, + "theoretical_loss": 5.532872941169639, + "tokens_seen": 38141952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004992076228686059, + "loss": 3.5556, + "theoretical_loss": 5.531385329011181, + "tokens_seen": 38207488 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499197592778335, + "loss": 3.645, + "theoretical_loss": 5.529900979384985, + "tokens_seen": 38273024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991875626880642, + "loss": 3.4956, + "theoretical_loss": 5.5284198795677, + "tokens_seen": 38338560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991775325977934, + "loss": 3.4295, + "theoretical_loss": 5.526942016907242, + "tokens_seen": 38404096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991675025075226, + "loss": 3.6309, + "theoretical_loss": 5.5254673788222615, + "tokens_seen": 38469632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991574724172518, + "loss": 3.5755, + "theoretical_loss": 5.523995952801634, + "tokens_seen": 38535168 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499147442326981, + "loss": 3.4162, + "theoretical_loss": 5.522527726403953, + "tokens_seen": 38600704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991374122367101, + "loss": 3.7008, + "theoretical_loss": 5.521062687257016, + "tokens_seen": 38666240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991273821464393, + "loss": 3.6047, + "theoretical_loss": 5.519600823057332, + "tokens_seen": 38731776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991173520561685, + "loss": 3.5591, + "theoretical_loss": 5.518142121569613, + "tokens_seen": 38797312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004991073219658977, + "loss": 3.5061, + "theoretical_loss": 5.516686570626293, + "tokens_seen": 38862848 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499097291875627, + "loss": 3.6794, + "theoretical_loss": 5.515234158127032, + "tokens_seen": 38928384 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499087261785356, + "loss": 3.7009, + "theoretical_loss": 5.5137848720382365, + "tokens_seen": 38993920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990772316950853, + "loss": 3.5366, + "theoretical_loss": 5.5123387003925775, + "tokens_seen": 39059456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990672016048144, + "loss": 3.5568, + "theoretical_loss": 5.510895631288515, + "tokens_seen": 39124992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990571715145437, + "loss": 3.5371, + "theoretical_loss": 5.50945565288983, + "tokens_seen": 39190528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990471414242729, + "loss": 3.5054, + "theoretical_loss": 5.508018753425152, + "tokens_seen": 39256064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 21760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.261561870574951, + "objective/train/theoretical_loss": 5.506584921187503, + "objective/train/tokens_used": 59781600, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.000499037111334002, + "loss": 3.4088, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990270812437312, + "loss": 3.5617, + "theoretical_loss": 5.505154144533835, + "tokens_seen": 39387136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990170511534604, + "loss": 3.6353, + "theoretical_loss": 5.503726411884568, + "tokens_seen": 39452672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004990070210631896, + "loss": 3.6269, + "theoretical_loss": 5.502301711723156, + "tokens_seen": 39518208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989969909729188, + "loss": 3.4697, + "theoretical_loss": 5.5008800325956235, + "tokens_seen": 39583744 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498986960882648, + "loss": 3.6587, + "theoretical_loss": 5.499461363110136, + "tokens_seen": 39649280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989769307923771, + "loss": 3.4003, + "theoretical_loss": 5.498045691936548, + "tokens_seen": 39714816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989669007021063, + "loss": 3.5999, + "theoretical_loss": 5.496633007805977, + "tokens_seen": 39780352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989568706118355, + "loss": 3.5044, + "theoretical_loss": 5.495223299510372, + "tokens_seen": 39845888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989468405215647, + "loss": 3.5133, + "theoretical_loss": 5.493816555902079, + "tokens_seen": 39911424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989368104312939, + "loss": 3.6526, + "theoretical_loss": 5.492412765893421, + "tokens_seen": 39976960 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498926780341023, + "loss": 3.5973, + "theoretical_loss": 5.49101191845628, + "tokens_seen": 40042496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989167502507523, + "loss": 3.457, + "theoretical_loss": 5.489614002621677, + "tokens_seen": 40108032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004989067201604814, + "loss": 3.5232, + "theoretical_loss": 5.4882190074793575, + "tokens_seen": 40173568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988966900702107, + "loss": 3.5784, + "theoretical_loss": 5.486826922177387, + "tokens_seen": 40239104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988866599799398, + "loss": 3.481, + "theoretical_loss": 5.4854377359217406, + "tokens_seen": 40304640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988766298896691, + "loss": 3.5402, + "theoretical_loss": 5.484051437975901, + "tokens_seen": 40370176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988665997993982, + "loss": 3.4356, + "theoretical_loss": 5.482668017660462, + "tokens_seen": 40435712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988565697091274, + "loss": 3.6824, + "theoretical_loss": 5.481287464352727, + "tokens_seen": 40501248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988465396188566, + "loss": 3.5418, + "theoretical_loss": 5.479909767486326, + "tokens_seen": 40566784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988365095285858, + "loss": 3.5266, + "theoretical_loss": 5.478534916550817, + "tokens_seen": 40632320 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498826479438315, + "loss": 3.7795, + "theoretical_loss": 5.477162901091306, + "tokens_seen": 40697856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988164493480441, + "loss": 3.4777, + "theoretical_loss": 5.475793710708061, + "tokens_seen": 40763392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004988064192577733, + "loss": 3.6463, + "theoretical_loss": 5.474427335056138, + "tokens_seen": 40828928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987963891675025, + "loss": 3.5159, + "theoretical_loss": 5.473063763844998, + "tokens_seen": 40894464 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987863590772317, + "loss": 3.6649, + "theoretical_loss": 5.471702986838139, + "tokens_seen": 40960000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987763289869609, + "loss": 3.6289, + "theoretical_loss": 5.470344993852729, + "tokens_seen": 41025536 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049876629889669, + "loss": 3.5264, + "theoretical_loss": 5.46898977475923, + "tokens_seen": 41091072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987562688064192, + "loss": 3.4805, + "theoretical_loss": 5.467637319481042, + "tokens_seen": 41156608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987462387161484, + "loss": 3.4954, + "theoretical_loss": 5.466287617994144, + "tokens_seen": 41222144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987362086258777, + "loss": 3.6776, + "theoretical_loss": 5.464940660326731, + "tokens_seen": 41287680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987261785356068, + "loss": 3.4983, + "theoretical_loss": 5.463596436558861, + "tokens_seen": 41353216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987161484453361, + "loss": 3.4561, + "theoretical_loss": 5.462254936822105, + "tokens_seen": 41418752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004987061183550651, + "loss": 3.3689, + "theoretical_loss": 5.4609161512992, + "tokens_seen": 41484288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986960882647944, + "loss": 3.5209, + "theoretical_loss": 5.4595800702237005, + "tokens_seen": 41549824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986860581745236, + "loss": 3.4534, + "theoretical_loss": 5.458246683879636, + "tokens_seen": 41615360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986760280842528, + "loss": 3.4917, + "theoretical_loss": 5.456915982601173, + "tokens_seen": 41680896 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498665997993982, + "loss": 3.5129, + "theoretical_loss": 5.455587956772277, + "tokens_seen": 41746432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986559679037111, + "loss": 3.5424, + "theoretical_loss": 5.454262596826377, + "tokens_seen": 41811968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986459378134403, + "loss": 3.6033, + "theoretical_loss": 5.452939893246034, + "tokens_seen": 41877504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986359077231695, + "loss": 3.4227, + "theoretical_loss": 5.451619836562617, + "tokens_seen": 41943040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986258776328987, + "loss": 3.3481, + "theoretical_loss": 5.450302417355973, + "tokens_seen": 42008576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004986158475426279, + "loss": 3.4214, + "theoretical_loss": 5.448987626254097, + "tokens_seen": 42074112 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498605817452357, + "loss": 3.5868, + "theoretical_loss": 5.447675453932829, + "tokens_seen": 42139648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985957873620862, + "loss": 3.4632, + "theoretical_loss": 5.446365891115516, + "tokens_seen": 42205184 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985857572718154, + "loss": 3.4932, + "theoretical_loss": 5.445058928572712, + "tokens_seen": 42270720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985757271815446, + "loss": 3.2718, + "theoretical_loss": 5.443754557121853, + "tokens_seen": 42336256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985656970912738, + "loss": 3.5228, + "theoretical_loss": 5.442452767626957, + "tokens_seen": 42401792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985556670010031, + "loss": 3.44, + "theoretical_loss": 5.441153550998311, + "tokens_seen": 42467328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985456369107321, + "loss": 3.6274, + "theoretical_loss": 5.439856898192163, + "tokens_seen": 42532864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 22938, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.533298969268799, + "objective/train/theoretical_loss": 5.438562800210427, + "objective/train/tokens_used": 63058400, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985356068204614, + "loss": 3.5661, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985255767301905, + "loss": 3.4244, + "theoretical_loss": 5.437271248100375, + "tokens_seen": 42663936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004985155466399198, + "loss": 3.3441, + "theoretical_loss": 5.435982232954342, + "tokens_seen": 42729472 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498505516549649, + "loss": 3.4363, + "theoretical_loss": 5.434695745909433, + "tokens_seen": 42795008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984954864593782, + "loss": 3.4464, + "theoretical_loss": 5.433411778147226, + "tokens_seen": 42860544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984854563691073, + "loss": 3.3311, + "theoretical_loss": 5.432130320893483, + "tokens_seen": 42926080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984754262788365, + "loss": 3.4539, + "theoretical_loss": 5.430851365417862, + "tokens_seen": 42991616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984653961885657, + "loss": 3.3607, + "theoretical_loss": 5.429574903033627, + "tokens_seen": 43057152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984553660982949, + "loss": 3.4861, + "theoretical_loss": 5.428300925097373, + "tokens_seen": 43122688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984453360080241, + "loss": 3.3869, + "theoretical_loss": 5.427029423008732, + "tokens_seen": 43188224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984353059177532, + "loss": 3.3839, + "theoretical_loss": 5.425760388210108, + "tokens_seen": 43253760 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984252758274825, + "loss": 3.497, + "theoretical_loss": 5.424493812186386, + "tokens_seen": 43319296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984152457372116, + "loss": 3.3457, + "theoretical_loss": 5.423229686464667, + "tokens_seen": 43384832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984052156469409, + "loss": 3.4126, + "theoretical_loss": 5.421968002613992, + "tokens_seen": 43450368 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049839518555667, + "loss": 3.6707, + "theoretical_loss": 5.420708752245071, + "tokens_seen": 43515904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983851554663993, + "loss": 3.5024, + "theoretical_loss": 5.419451927010014, + "tokens_seen": 43581440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983751253761284, + "loss": 3.5329, + "theoretical_loss": 5.418197518602068, + "tokens_seen": 43646976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983650952858576, + "loss": 3.5019, + "theoretical_loss": 5.41694551875535, + "tokens_seen": 43712512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983550651955868, + "loss": 3.4759, + "theoretical_loss": 5.415695919244586, + "tokens_seen": 43778048 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498345035105316, + "loss": 3.434, + "theoretical_loss": 5.41444871188485, + "tokens_seen": 43843584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983350050150452, + "loss": 3.2489, + "theoretical_loss": 5.413203888531308, + "tokens_seen": 43909120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983249749247743, + "loss": 3.3276, + "theoretical_loss": 5.411961441078963, + "tokens_seen": 43974656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983149448345035, + "loss": 3.4232, + "theoretical_loss": 5.410721361462395, + "tokens_seen": 44040192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004983049147442327, + "loss": 3.6092, + "theoretical_loss": 5.409483641655516, + "tokens_seen": 44105728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982948846539619, + "loss": 3.3943, + "theoretical_loss": 5.408248273671317, + "tokens_seen": 44171264 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982848545636911, + "loss": 3.2324, + "theoretical_loss": 5.407015249561624, + "tokens_seen": 44236800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982748244734202, + "loss": 3.4505, + "theoretical_loss": 5.405784561416844, + "tokens_seen": 44302336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982647943831494, + "loss": 3.3048, + "theoretical_loss": 5.404556201365727, + "tokens_seen": 44367872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982547642928786, + "loss": 3.3843, + "theoretical_loss": 5.403330161575123, + "tokens_seen": 44433408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982447342026079, + "loss": 3.3767, + "theoretical_loss": 5.402106434249742, + "tokens_seen": 44498944 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498234704112337, + "loss": 3.3667, + "theoretical_loss": 5.400885011631912, + "tokens_seen": 44564480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982246740220663, + "loss": 3.4477, + "theoretical_loss": 5.399665886001349, + "tokens_seen": 44630016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982146439317953, + "loss": 3.3393, + "theoretical_loss": 5.398449049674916, + "tokens_seen": 44695552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004982046138415246, + "loss": 3.3861, + "theoretical_loss": 5.397234495006392, + "tokens_seen": 44761088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981945837512538, + "loss": 3.4267, + "theoretical_loss": 5.396022214386246, + "tokens_seen": 44826624 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498184553660983, + "loss": 3.2004, + "theoretical_loss": 5.394812200241393, + "tokens_seen": 44892160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981745235707122, + "loss": 3.26, + "theoretical_loss": 5.393604445034988, + "tokens_seen": 44957696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981644934804413, + "loss": 3.3881, + "theoretical_loss": 5.392398941266183, + "tokens_seen": 45023232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981544633901705, + "loss": 3.2602, + "theoretical_loss": 5.391195681469908, + "tokens_seen": 45088768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981444332998997, + "loss": 3.3899, + "theoretical_loss": 5.389994658216651, + "tokens_seen": 45154304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981344032096289, + "loss": 3.3587, + "theoretical_loss": 5.388795864112235, + "tokens_seen": 45219840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981243731193581, + "loss": 3.2707, + "theoretical_loss": 5.387599291797599, + "tokens_seen": 45285376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981143430290873, + "loss": 3.3527, + "theoretical_loss": 5.386404933948584, + "tokens_seen": 45350912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004981043129388164, + "loss": 3.2937, + "theoretical_loss": 5.38521278327571, + "tokens_seen": 45416448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980942828485456, + "loss": 3.1938, + "theoretical_loss": 5.384022832523971, + "tokens_seen": 45481984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980842527582748, + "loss": 3.3112, + "theoretical_loss": 5.382835074472616, + "tokens_seen": 45547520 + }, + { + "epoch": 0.01, + "learning_rate": 0.000498074222668004, + "loss": 3.4723, + "theoretical_loss": 5.381649501934939, + "tokens_seen": 45613056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980641925777333, + "loss": 3.4063, + "theoretical_loss": 5.3804661077580755, + "tokens_seen": 45678592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980541624874623, + "loss": 3.3298, + "theoretical_loss": 5.379284884822787, + "tokens_seen": 45744128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980441323971916, + "loss": 3.4842, + "theoretical_loss": 5.37810582604326, + "tokens_seen": 45809664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 24173, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.244412899017334, + "objective/train/theoretical_loss": 5.376928924366901, + "objective/train/tokens_used": 66335200, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980341023069207, + "loss": 3.2732, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00049802407221665, + "loss": 3.2808, + "theoretical_loss": 5.37575417277413, + "tokens_seen": 45940736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980140421263792, + "loss": 3.4211, + "theoretical_loss": 5.374581564278184, + "tokens_seen": 46006272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004980040120361084, + "loss": 3.2733, + "theoretical_loss": 5.373411091924918, + "tokens_seen": 46071808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979939819458375, + "loss": 3.2797, + "theoretical_loss": 5.372242748792598, + "tokens_seen": 46137344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979839518555667, + "loss": 3.3182, + "theoretical_loss": 5.371076527991711, + "tokens_seen": 46202880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979739217652959, + "loss": 3.3615, + "theoretical_loss": 5.36991242266477, + "tokens_seen": 46268416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979638916750251, + "loss": 3.1471, + "theoretical_loss": 5.36875042598612, + "tokens_seen": 46333952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979538615847543, + "loss": 3.2799, + "theoretical_loss": 5.3675905311617385, + "tokens_seen": 46399488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979438314944834, + "loss": 3.3191, + "theoretical_loss": 5.366432731429054, + "tokens_seen": 46465024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979338014042126, + "loss": 3.5641, + "theoretical_loss": 5.365277020056752, + "tokens_seen": 46530560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979237713139418, + "loss": 3.2144, + "theoretical_loss": 5.364123390344582, + "tokens_seen": 46596096 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497913741223671, + "loss": 3.4544, + "theoretical_loss": 5.362971835623181, + "tokens_seen": 46661632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004979037111334002, + "loss": 3.3118, + "theoretical_loss": 5.361822349253881, + "tokens_seen": 46727168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978936810431293, + "loss": 3.3674, + "theoretical_loss": 5.360674924628524, + "tokens_seen": 46792704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978836509528586, + "loss": 3.3344, + "theoretical_loss": 5.359529555169287, + "tokens_seen": 46858240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978736208625877, + "loss": 3.3403, + "theoretical_loss": 5.358386234328487, + "tokens_seen": 46923776 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497863590772317, + "loss": 3.2124, + "theoretical_loss": 5.357244955588417, + "tokens_seen": 46989312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978535606820461, + "loss": 3.3922, + "theoretical_loss": 5.356105712461154, + "tokens_seen": 47054848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978435305917754, + "loss": 3.4126, + "theoretical_loss": 5.354968498488391, + "tokens_seen": 47120384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978335005015045, + "loss": 3.1662, + "theoretical_loss": 5.353833307241256, + "tokens_seen": 47185920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978234704112337, + "loss": 3.3329, + "theoretical_loss": 5.352700132320135, + "tokens_seen": 47251456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978134403209629, + "loss": 3.3241, + "theoretical_loss": 5.351568967354504, + "tokens_seen": 47316992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004978034102306921, + "loss": 3.475, + "theoretical_loss": 5.350439806002757, + "tokens_seen": 47382528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977933801404213, + "loss": 3.1461, + "theoretical_loss": 5.349312641952025, + "tokens_seen": 47448064 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977833500501504, + "loss": 3.1719, + "theoretical_loss": 5.348187468918022, + "tokens_seen": 47513600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977733199598796, + "loss": 3.3345, + "theoretical_loss": 5.347064280644861, + "tokens_seen": 47579136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977632898696088, + "loss": 3.2581, + "theoretical_loss": 5.345943070904897, + "tokens_seen": 47644672 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497753259779338, + "loss": 3.3514, + "theoretical_loss": 5.344823833498557, + "tokens_seen": 47710208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977432296890672, + "loss": 3.1913, + "theoretical_loss": 5.343706562254177, + "tokens_seen": 47775744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977331995987965, + "loss": 3.2925, + "theoretical_loss": 5.342591251027833, + "tokens_seen": 47841280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977231695085255, + "loss": 3.3166, + "theoretical_loss": 5.3414778937031855, + "tokens_seen": 47906816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004977131394182548, + "loss": 3.2012, + "theoretical_loss": 5.340366484191316, + "tokens_seen": 47972352 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497703109327984, + "loss": 3.1705, + "theoretical_loss": 5.339257016430563, + "tokens_seen": 48037888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976930792377132, + "loss": 3.1063, + "theoretical_loss": 5.338149484386367, + "tokens_seen": 48103424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976830491474424, + "loss": 3.0419, + "theoretical_loss": 5.337043882051111, + "tokens_seen": 48168960 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976730190571715, + "loss": 3.4246, + "theoretical_loss": 5.3359402034439665, + "tokens_seen": 48234496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976629889669007, + "loss": 3.2783, + "theoretical_loss": 5.334838442610732, + "tokens_seen": 48300032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976529588766299, + "loss": 3.1279, + "theoretical_loss": 5.333738593623686, + "tokens_seen": 48365568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976429287863591, + "loss": 3.1068, + "theoretical_loss": 5.332640650581424, + "tokens_seen": 48431104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976328986960883, + "loss": 3.1669, + "theoretical_loss": 5.331544607608718, + "tokens_seen": 48496640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976228686058175, + "loss": 3.1997, + "theoretical_loss": 5.330450458856356, + "tokens_seen": 48562176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976128385155466, + "loss": 3.283, + "theoretical_loss": 5.329358198500994, + "tokens_seen": 48627712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004976028084252758, + "loss": 3.3073, + "theoretical_loss": 5.328267820745011, + "tokens_seen": 48693248 + }, + { + "epoch": 0.01, + "learning_rate": 0.000497592778335005, + "loss": 3.2016, + "theoretical_loss": 5.327179319816354, + "tokens_seen": 48758784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975827482447342, + "loss": 3.1588, + "theoretical_loss": 5.326092689968397, + "tokens_seen": 48824320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975727181544635, + "loss": 3.2771, + "theoretical_loss": 5.3250079254797935, + "tokens_seen": 48889856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975626880641925, + "loss": 3.2783, + "theoretical_loss": 5.323925020654329, + "tokens_seen": 48955392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975526579739218, + "loss": 3.2774, + "theoretical_loss": 5.322843969820781, + "tokens_seen": 49020928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975426278836509, + "loss": 3.3671, + "theoretical_loss": 5.321764767332772, + "tokens_seen": 49086464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 25394, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.2616124153137207, + "objective/train/theoretical_loss": 5.320687407568633, + "objective/train/tokens_used": 69612000, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975325977933802, + "loss": 3.2469, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975225677031094, + "loss": 3.2161, + "theoretical_loss": 5.3196118849312555, + "tokens_seen": 49217536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975125376128386, + "loss": 3.0953, + "theoretical_loss": 5.31853819384796, + "tokens_seen": 49283072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004975025075225677, + "loss": 3.2303, + "theoretical_loss": 5.31746632877035, + "tokens_seen": 49348608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004974924774322969, + "loss": 3.202, + "theoretical_loss": 5.316396284174177, + "tokens_seen": 49414144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004974824473420261, + "loss": 2.9365, + "theoretical_loss": 5.3153280545592025, + "tokens_seen": 49479680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974724172517553, + "loss": 3.1697, + "theoretical_loss": 5.314261634449066, + "tokens_seen": 49545216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974623871614845, + "loss": 3.2332, + "theoretical_loss": 5.313197018391143, + "tokens_seen": 49610752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974523570712136, + "loss": 3.1275, + "theoretical_loss": 5.312134200956416, + "tokens_seen": 49676288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974423269809428, + "loss": 3.1183, + "theoretical_loss": 5.3110731767393435, + "tokens_seen": 49741824 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497432296890672, + "loss": 3.1224, + "theoretical_loss": 5.31001394035772, + "tokens_seen": 49807360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974222668004012, + "loss": 3.0423, + "theoretical_loss": 5.308956486452549, + "tokens_seen": 49872896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974122367101304, + "loss": 3.2029, + "theoretical_loss": 5.307900809687913, + "tokens_seen": 49938432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004974022066198595, + "loss": 3.1181, + "theoretical_loss": 5.306846904750847, + "tokens_seen": 50003968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973921765295888, + "loss": 3.0852, + "theoretical_loss": 5.305794766351202, + "tokens_seen": 50069504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973821464393179, + "loss": 3.1424, + "theoretical_loss": 5.304744389221523, + "tokens_seen": 50135040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973721163490472, + "loss": 3.1737, + "theoretical_loss": 5.303695768116921, + "tokens_seen": 50200576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973620862587763, + "loss": 3.0162, + "theoretical_loss": 5.302648897814946, + "tokens_seen": 50266112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973520561685056, + "loss": 3.0965, + "theoretical_loss": 5.301603773115465, + "tokens_seen": 50331648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973420260782347, + "loss": 3.1513, + "theoretical_loss": 5.300560388840532, + "tokens_seen": 50397184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973319959879639, + "loss": 3.0397, + "theoretical_loss": 5.299518739834273, + "tokens_seen": 50462720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973219658976931, + "loss": 3.0186, + "theoretical_loss": 5.298478820962751, + "tokens_seen": 50528256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973119358074223, + "loss": 3.1058, + "theoretical_loss": 5.297440627113859, + "tokens_seen": 50593792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004973019057171515, + "loss": 3.1252, + "theoretical_loss": 5.296404153197189, + "tokens_seen": 50659328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972918756268806, + "loss": 3.1342, + "theoretical_loss": 5.295369394143913, + "tokens_seen": 50724864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972818455366098, + "loss": 3.2945, + "theoretical_loss": 5.294336344906668, + "tokens_seen": 50790400 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497271815446339, + "loss": 3.2263, + "theoretical_loss": 5.293305000459435, + "tokens_seen": 50855936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972617853560682, + "loss": 3.0768, + "theoretical_loss": 5.292275355797421, + "tokens_seen": 50921472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972517552657974, + "loss": 3.1541, + "theoretical_loss": 5.291247405936941, + "tokens_seen": 50987008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972417251755266, + "loss": 3.0212, + "theoretical_loss": 5.290221145915307, + "tokens_seen": 51052544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972316950852557, + "loss": 3.0216, + "theoretical_loss": 5.289196570790708, + "tokens_seen": 51118080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972216649949849, + "loss": 3.1768, + "theoretical_loss": 5.288173675642094, + "tokens_seen": 51183616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972116349047142, + "loss": 2.9453, + "theoretical_loss": 5.287152455569069, + "tokens_seen": 51249152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004972016048144433, + "loss": 2.954, + "theoretical_loss": 5.286132905691774, + "tokens_seen": 51314688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971915747241726, + "loss": 3.0675, + "theoretical_loss": 5.285115021150775, + "tokens_seen": 51380224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971815446339017, + "loss": 3.0947, + "theoretical_loss": 5.284098797106948, + "tokens_seen": 51445760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971715145436309, + "loss": 3.1805, + "theoretical_loss": 5.283084228741377, + "tokens_seen": 51511296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971614844533601, + "loss": 3.0407, + "theoretical_loss": 5.282071311255235, + "tokens_seen": 51576832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971514543630893, + "loss": 2.8487, + "theoretical_loss": 5.2810600398696845, + "tokens_seen": 51642368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971414242728185, + "loss": 3.0642, + "theoretical_loss": 5.2800504098257575, + "tokens_seen": 51707904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971313941825477, + "loss": 2.9783, + "theoretical_loss": 5.279042416384257, + "tokens_seen": 51773440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971213640922768, + "loss": 3.1319, + "theoretical_loss": 5.278036054825645, + "tokens_seen": 51838976 + }, + { + "epoch": 0.02, + "learning_rate": 0.000497111334002006, + "loss": 3.0639, + "theoretical_loss": 5.277031320449941, + "tokens_seen": 51904512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004971013039117352, + "loss": 3.0766, + "theoretical_loss": 5.27602820857661, + "tokens_seen": 51970048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970912738214644, + "loss": 3.0661, + "theoretical_loss": 5.275026714544463, + "tokens_seen": 52035584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970812437311936, + "loss": 2.9486, + "theoretical_loss": 5.27402683371155, + "tokens_seen": 52101120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970712136409227, + "loss": 2.9227, + "theoretical_loss": 5.273028561455056, + "tokens_seen": 52166656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970611835506519, + "loss": 3.0415, + "theoretical_loss": 5.272031893171203, + "tokens_seen": 52232192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970511534603811, + "loss": 3.0331, + "theoretical_loss": 5.27103682427514, + "tokens_seen": 52297728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970411233701103, + "loss": 3.0387, + "theoretical_loss": 5.270043350200851, + "tokens_seen": 52363264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 26541, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.903182029724121, + "objective/train/theoretical_loss": 5.269051466401043, + "objective/train/tokens_used": 72888800, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970310932798396, + "loss": 2.8554, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970210631895686, + "loss": 2.9793, + "theoretical_loss": 5.268061168347058, + "tokens_seen": 52494336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970110330992979, + "loss": 3.0324, + "theoretical_loss": 5.267072451528764, + "tokens_seen": 52559872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004970010030090271, + "loss": 3.0227, + "theoretical_loss": 5.266085311454461, + "tokens_seen": 52625408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969909729187563, + "loss": 3.0175, + "theoretical_loss": 5.265099743650779, + "tokens_seen": 52690944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969809428284855, + "loss": 3.1, + "theoretical_loss": 5.264115743662588, + "tokens_seen": 52756480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969709127382147, + "loss": 3.0458, + "theoretical_loss": 5.26313330705289, + "tokens_seen": 52822016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969608826479438, + "loss": 2.9225, + "theoretical_loss": 5.262152429402734, + "tokens_seen": 52887552 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496950852557673, + "loss": 2.9116, + "theoretical_loss": 5.261173106311116, + "tokens_seen": 52953088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969408224674022, + "loss": 3.0762, + "theoretical_loss": 5.2601953333948765, + "tokens_seen": 53018624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969307923771314, + "loss": 2.9531, + "theoretical_loss": 5.25921910628862, + "tokens_seen": 53084160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969207622868606, + "loss": 2.8673, + "theoretical_loss": 5.258244420644611, + "tokens_seen": 53149696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004969107321965897, + "loss": 2.9323, + "theoretical_loss": 5.257271272132685, + "tokens_seen": 53215232 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496900702106319, + "loss": 2.8957, + "theoretical_loss": 5.256299656440154, + "tokens_seen": 53280768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968906720160481, + "loss": 2.8928, + "theoretical_loss": 5.255329569271717, + "tokens_seen": 53346304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968806419257774, + "loss": 2.8397, + "theoretical_loss": 5.254361006349367, + "tokens_seen": 53411840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968706118355065, + "loss": 2.9566, + "theoretical_loss": 5.2533939634122975, + "tokens_seen": 53477376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968605817452358, + "loss": 2.9, + "theoretical_loss": 5.252428436216821, + "tokens_seen": 53542912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968505516549649, + "loss": 2.9272, + "theoretical_loss": 5.251464420536269, + "tokens_seen": 53608448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968405215646941, + "loss": 2.9353, + "theoretical_loss": 5.250501912160911, + "tokens_seen": 53673984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968304914744233, + "loss": 3.0027, + "theoretical_loss": 5.24954090689786, + "tokens_seen": 53739520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968204613841525, + "loss": 3.1325, + "theoretical_loss": 5.248581400570989, + "tokens_seen": 53805056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968104312938817, + "loss": 2.9972, + "theoretical_loss": 5.247623389020841, + "tokens_seen": 53870592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004968004012036108, + "loss": 2.9275, + "theoretical_loss": 5.246666868104547, + "tokens_seen": 53936128 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049679037111334, + "loss": 2.9345, + "theoretical_loss": 5.245711833695731, + "tokens_seen": 54001664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967803410230692, + "loss": 3.0151, + "theoretical_loss": 5.244758281684433, + "tokens_seen": 54067200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967703109327984, + "loss": 2.9518, + "theoretical_loss": 5.24380620797702, + "tokens_seen": 54132736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967602808425276, + "loss": 2.6879, + "theoretical_loss": 5.242855608496102, + "tokens_seen": 54198272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967502507522568, + "loss": 2.9858, + "theoretical_loss": 5.2419064791804475, + "tokens_seen": 54263808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967402206619859, + "loss": 2.8926, + "theoretical_loss": 5.240958815984901, + "tokens_seen": 54329344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967301905717151, + "loss": 2.8316, + "theoretical_loss": 5.2400126148803, + "tokens_seen": 54394880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967201604814444, + "loss": 2.8804, + "theoretical_loss": 5.23906787185339, + "tokens_seen": 54460416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967101303911735, + "loss": 2.8869, + "theoretical_loss": 5.238124582906748, + "tokens_seen": 54525952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004967001003009028, + "loss": 2.9651, + "theoretical_loss": 5.237182744058696, + "tokens_seen": 54591488 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496690070210632, + "loss": 2.8806, + "theoretical_loss": 5.236242351343224, + "tokens_seen": 54657024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966800401203611, + "loss": 2.8032, + "theoretical_loss": 5.235303400809904, + "tokens_seen": 54722560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966700100300903, + "loss": 2.9231, + "theoretical_loss": 5.2343658885238185, + "tokens_seen": 54788096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966599799398195, + "loss": 2.8803, + "theoretical_loss": 5.2334298105654735, + "tokens_seen": 54853632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966499498495487, + "loss": 2.9421, + "theoretical_loss": 5.232495163030723, + "tokens_seen": 54919168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966399197592779, + "loss": 2.9557, + "theoretical_loss": 5.231561942030694, + "tokens_seen": 54984704 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496629889669007, + "loss": 2.8899, + "theoretical_loss": 5.230630143691698, + "tokens_seen": 55050240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966198595787362, + "loss": 3.0725, + "theoretical_loss": 5.22969976415517, + "tokens_seen": 55115776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004966098294884654, + "loss": 2.6245, + "theoretical_loss": 5.228770799577576, + "tokens_seen": 55181312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965997993981946, + "loss": 2.8727, + "theoretical_loss": 5.227843246130343, + "tokens_seen": 55246848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965897693079238, + "loss": 2.7802, + "theoretical_loss": 5.2269170999997865, + "tokens_seen": 55312384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965797392176529, + "loss": 2.8914, + "theoretical_loss": 5.225992357387031, + "tokens_seen": 55377920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965697091273821, + "loss": 2.8788, + "theoretical_loss": 5.225069014507933, + "tokens_seen": 55443456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965596790371113, + "loss": 2.6651, + "theoretical_loss": 5.224147067593012, + "tokens_seen": 55508992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965496489468405, + "loss": 2.8943, + "theoretical_loss": 5.223226512887375, + "tokens_seen": 55574528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965396188565698, + "loss": 2.8835, + "theoretical_loss": 5.222307346650636, + "tokens_seen": 55640064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 27758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.726249933242798, + "objective/train/theoretical_loss": 5.221389565156855, + "objective/train/tokens_used": 76165600, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965295887662988, + "loss": 2.7644, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965195586760281, + "loss": 3.0678, + "theoretical_loss": 5.2204731646944555, + "tokens_seen": 55771136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004965095285857573, + "loss": 2.7629, + "theoretical_loss": 5.219558141566156, + "tokens_seen": 55836672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964994984954865, + "loss": 2.8223, + "theoretical_loss": 5.218644492088899, + "tokens_seen": 55902208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964894684052157, + "loss": 2.7633, + "theoretical_loss": 5.217732212593778, + "tokens_seen": 55967744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964794383149449, + "loss": 2.8854, + "theoretical_loss": 5.216821299425968, + "tokens_seen": 56033280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496469408224674, + "loss": 2.8312, + "theoretical_loss": 5.215911748944652, + "tokens_seen": 56098816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964593781344032, + "loss": 2.8853, + "theoretical_loss": 5.215003557522959, + "tokens_seen": 56164352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964493480441324, + "loss": 2.8335, + "theoretical_loss": 5.21409672154788, + "tokens_seen": 56229888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964393179538616, + "loss": 2.6882, + "theoretical_loss": 5.213191237420215, + "tokens_seen": 56295424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964292878635908, + "loss": 2.8365, + "theoretical_loss": 5.212287101554496, + "tokens_seen": 56360960 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049641925777332, + "loss": 2.6733, + "theoretical_loss": 5.211384310378916, + "tokens_seen": 56426496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004964092276830491, + "loss": 2.899, + "theoretical_loss": 5.210482860335269, + "tokens_seen": 56492032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963991975927783, + "loss": 2.7717, + "theoretical_loss": 5.2095827478788745, + "tokens_seen": 56557568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963891675025075, + "loss": 2.8355, + "theoretical_loss": 5.208683969478518, + "tokens_seen": 56623104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963791374122367, + "loss": 2.7755, + "theoretical_loss": 5.2077865216163755, + "tokens_seen": 56688640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963691073219659, + "loss": 2.7045, + "theoretical_loss": 5.206890400787959, + "tokens_seen": 56754176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963590772316951, + "loss": 2.7248, + "theoretical_loss": 5.205995603502039, + "tokens_seen": 56819712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963490471414242, + "loss": 2.7907, + "theoretical_loss": 5.205102126280586, + "tokens_seen": 56885248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963390170511535, + "loss": 2.7334, + "theoretical_loss": 5.204209965658704, + "tokens_seen": 56950784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963289869608827, + "loss": 2.7713, + "theoretical_loss": 5.203319118184565, + "tokens_seen": 57016320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004963189568706119, + "loss": 2.7271, + "theoretical_loss": 5.202429580419343, + "tokens_seen": 57081856 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496308926780341, + "loss": 2.7677, + "theoretical_loss": 5.201541348937157, + "tokens_seen": 57147392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962988966900702, + "loss": 2.6903, + "theoretical_loss": 5.200654420324998, + "tokens_seen": 57212928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962888665997994, + "loss": 2.6241, + "theoretical_loss": 5.199768791182677, + "tokens_seen": 57278464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962788365095286, + "loss": 2.7922, + "theoretical_loss": 5.198884458122748, + "tokens_seen": 57344000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962688064192578, + "loss": 2.824, + "theoretical_loss": 5.198001417770461, + "tokens_seen": 57409536 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496258776328987, + "loss": 2.7535, + "theoretical_loss": 5.1971196667636885, + "tokens_seen": 57475072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962487462387161, + "loss": 2.746, + "theoretical_loss": 5.196239201752871, + "tokens_seen": 57540608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962387161484453, + "loss": 2.7526, + "theoretical_loss": 5.19536001940095, + "tokens_seen": 57606144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962286860581746, + "loss": 2.6523, + "theoretical_loss": 5.1944821163833135, + "tokens_seen": 57671680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004962186559679037, + "loss": 2.7136, + "theoretical_loss": 5.19360548938773, + "tokens_seen": 57737216 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496208625877633, + "loss": 2.6686, + "theoretical_loss": 5.192730135114292, + "tokens_seen": 57802752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961985957873621, + "loss": 2.8023, + "theoretical_loss": 5.191856050275353, + "tokens_seen": 57868288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961885656970913, + "loss": 2.8623, + "theoretical_loss": 5.190983231595473, + "tokens_seen": 57933824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961785356068205, + "loss": 2.6344, + "theoretical_loss": 5.1901116758113535, + "tokens_seen": 57999360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961685055165497, + "loss": 2.6295, + "theoretical_loss": 5.189241379671782, + "tokens_seen": 58064896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961584754262789, + "loss": 2.824, + "theoretical_loss": 5.188372339937576, + "tokens_seen": 58130432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961484453360081, + "loss": 2.8064, + "theoretical_loss": 5.187504553381521, + "tokens_seen": 58195968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961384152457372, + "loss": 2.6023, + "theoretical_loss": 5.1866380167883115, + "tokens_seen": 58261504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961283851554664, + "loss": 2.7624, + "theoretical_loss": 5.1857727269545, + "tokens_seen": 58327040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961183550651956, + "loss": 2.7573, + "theoretical_loss": 5.184908680688437, + "tokens_seen": 58392576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004961083249749248, + "loss": 2.6314, + "theoretical_loss": 5.18404587481021, + "tokens_seen": 58458112 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496098294884654, + "loss": 2.5928, + "theoretical_loss": 5.183184306151595, + "tokens_seen": 58523648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960882647943831, + "loss": 2.6044, + "theoretical_loss": 5.182323971555995, + "tokens_seen": 58589184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960782347041123, + "loss": 2.7901, + "theoretical_loss": 5.181464867878384, + "tokens_seen": 58654720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960682046138415, + "loss": 2.9129, + "theoretical_loss": 5.1806069919852575, + "tokens_seen": 58720256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960581745235707, + "loss": 2.6296, + "theoretical_loss": 5.17975034075457, + "tokens_seen": 58785792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960481444333, + "loss": 2.6992, + "theoretical_loss": 5.1788949110756874, + "tokens_seen": 58851328 + }, + { + "epoch": 0.02, + "learning_rate": 0.000496038114343029, + "loss": 2.5925, + "theoretical_loss": 5.178040699849327, + "tokens_seen": 58916864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 29071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.467764139175415, + "objective/train/theoretical_loss": 5.1771877039875065, + "objective/train/tokens_used": 79442400, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960280842527583, + "loss": 2.5666, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960180541624875, + "loss": 2.6329, + "theoretical_loss": 5.176335920413489, + "tokens_seen": 59047936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004960080240722167, + "loss": 2.739, + "theoretical_loss": 5.175485346061732, + "tokens_seen": 59113472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959979939819459, + "loss": 2.6209, + "theoretical_loss": 5.174635977877831, + "tokens_seen": 59179008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959879638916751, + "loss": 2.7303, + "theoretical_loss": 5.173787812818468, + "tokens_seen": 59244544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959779338014042, + "loss": 2.6576, + "theoretical_loss": 5.172940847851359, + "tokens_seen": 59310080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959679037111334, + "loss": 2.6681, + "theoretical_loss": 5.172095079955204, + "tokens_seen": 59375616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959578736208626, + "loss": 2.4987, + "theoretical_loss": 5.171250506119634, + "tokens_seen": 59441152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959478435305918, + "loss": 2.7807, + "theoretical_loss": 5.170407123345155, + "tokens_seen": 59506688 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495937813440321, + "loss": 2.6089, + "theoretical_loss": 5.169564928643103, + "tokens_seen": 59572224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959277833500501, + "loss": 2.6434, + "theoretical_loss": 5.168723919035591, + "tokens_seen": 59637760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959177532597793, + "loss": 2.5194, + "theoretical_loss": 5.167884091555456, + "tokens_seen": 59703296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004959077231695085, + "loss": 2.6549, + "theoretical_loss": 5.167045443246213, + "tokens_seen": 59768832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958976930792377, + "loss": 2.5936, + "theoretical_loss": 5.166207971162001, + "tokens_seen": 59834368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958876629889669, + "loss": 2.6061, + "theoretical_loss": 5.1653716723675345, + "tokens_seen": 59899904 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495877632898696, + "loss": 2.7264, + "theoretical_loss": 5.164536543938056, + "tokens_seen": 59965440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958676028084253, + "loss": 2.5703, + "theoretical_loss": 5.163702582959283, + "tokens_seen": 60030976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958575727181544, + "loss": 2.812, + "theoretical_loss": 5.162869786527363, + "tokens_seen": 60096512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958475426278837, + "loss": 2.7306, + "theoretical_loss": 5.1620381517488205, + "tokens_seen": 60162048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958375125376129, + "loss": 2.6976, + "theoretical_loss": 5.161207675740515, + "tokens_seen": 60227584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958274824473421, + "loss": 2.6215, + "theoretical_loss": 5.160378355629584, + "tokens_seen": 60293120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958174523570712, + "loss": 2.6476, + "theoretical_loss": 5.159550188553405, + "tokens_seen": 60358656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004958074222668004, + "loss": 2.6968, + "theoretical_loss": 5.15872317165954, + "tokens_seen": 60424192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957973921765296, + "loss": 2.6204, + "theoretical_loss": 5.1578973021056935, + "tokens_seen": 60489728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957873620862588, + "loss": 2.595, + "theoretical_loss": 5.157072577059662, + "tokens_seen": 60555264 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495777331995988, + "loss": 2.7449, + "theoretical_loss": 5.15624899369929, + "tokens_seen": 60620800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957673019057172, + "loss": 2.6633, + "theoretical_loss": 5.155426549212422, + "tokens_seen": 60686336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957572718154463, + "loss": 2.6701, + "theoretical_loss": 5.154605240796855, + "tokens_seen": 60751872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957472417251755, + "loss": 2.5858, + "theoretical_loss": 5.153785065660296, + "tokens_seen": 60817408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957372116349047, + "loss": 2.7155, + "theoretical_loss": 5.152966021020315, + "tokens_seen": 60882944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957271815446339, + "loss": 2.6238, + "theoretical_loss": 5.152148104104297, + "tokens_seen": 60948480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957171514543631, + "loss": 2.5579, + "theoretical_loss": 5.1513313121494, + "tokens_seen": 61014016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004957071213640923, + "loss": 2.6649, + "theoretical_loss": 5.15051564240251, + "tokens_seen": 61079552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956970912738214, + "loss": 2.6694, + "theoretical_loss": 5.149701092120194, + "tokens_seen": 61145088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956870611835507, + "loss": 2.5664, + "theoretical_loss": 5.148887658568658, + "tokens_seen": 61210624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956770310932798, + "loss": 2.6691, + "theoretical_loss": 5.148075339023699, + "tokens_seen": 61276160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956670010030091, + "loss": 2.6029, + "theoretical_loss": 5.147264130770671, + "tokens_seen": 61341696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956569709127383, + "loss": 2.5932, + "theoretical_loss": 5.146454031104426, + "tokens_seen": 61407232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956469408224674, + "loss": 2.7343, + "theoretical_loss": 5.1456450373292855, + "tokens_seen": 61472768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956369107321966, + "loss": 2.6277, + "theoretical_loss": 5.144837146758987, + "tokens_seen": 61538304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956268806419258, + "loss": 2.6515, + "theoretical_loss": 5.144030356716646, + "tokens_seen": 61603840 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495616850551655, + "loss": 2.6539, + "theoretical_loss": 5.1432246645347135, + "tokens_seen": 61669376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004956068204613842, + "loss": 2.5245, + "theoretical_loss": 5.14242006755493, + "tokens_seen": 61734912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955967903711133, + "loss": 2.5762, + "theoretical_loss": 5.141616563128288, + "tokens_seen": 61800448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955867602808425, + "loss": 2.6381, + "theoretical_loss": 5.140814148614985, + "tokens_seen": 61865984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955767301905717, + "loss": 2.6046, + "theoretical_loss": 5.1400128213843885, + "tokens_seen": 61931520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955667001003009, + "loss": 2.7205, + "theoretical_loss": 5.139212578814986, + "tokens_seen": 61997056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955566700100301, + "loss": 2.6095, + "theoretical_loss": 5.13841341829435, + "tokens_seen": 62062592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955466399197592, + "loss": 2.5299, + "theoretical_loss": 5.137615337219098, + "tokens_seen": 62128128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955366098294884, + "loss": 2.6375, + "theoretical_loss": 5.136818332994844, + "tokens_seen": 62193664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 30276, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.826779365539551, + "objective/train/theoretical_loss": 5.136022403036167, + "objective/train/tokens_used": 82719200, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955265797392177, + "loss": 2.692, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955165496489468, + "loss": 2.5311, + "theoretical_loss": 5.135227544766565, + "tokens_seen": 62324736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004955065195586761, + "loss": 2.6723, + "theoretical_loss": 5.134433755618414, + "tokens_seen": 62390272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954964894684052, + "loss": 2.4407, + "theoretical_loss": 5.1336410330329345, + "tokens_seen": 62455808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954864593781344, + "loss": 2.573, + "theoretical_loss": 5.132849374460145, + "tokens_seen": 62521344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954764292878636, + "loss": 2.6037, + "theoretical_loss": 5.132058777358827, + "tokens_seen": 62586880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954663991975928, + "loss": 2.5329, + "theoretical_loss": 5.1312692391964845, + "tokens_seen": 62652416 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495456369107322, + "loss": 2.5088, + "theoretical_loss": 5.1304807574493, + "tokens_seen": 62717952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954463390170512, + "loss": 2.5646, + "theoretical_loss": 5.129693329602108, + "tokens_seen": 62783488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954363089267803, + "loss": 2.4945, + "theoretical_loss": 5.128906953148342, + "tokens_seen": 62849024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954262788365095, + "loss": 2.6021, + "theoretical_loss": 5.128121625590007, + "tokens_seen": 62914560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954162487462387, + "loss": 2.6082, + "theoretical_loss": 5.127337344437633, + "tokens_seen": 62980096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004954062186559679, + "loss": 2.6673, + "theoretical_loss": 5.126554107210245, + "tokens_seen": 63045632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953961885656971, + "loss": 2.5752, + "theoretical_loss": 5.125771911435321, + "tokens_seen": 63111168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953861584754263, + "loss": 2.5591, + "theoretical_loss": 5.124990754648754, + "tokens_seen": 63176704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953761283851555, + "loss": 2.5316, + "theoretical_loss": 5.124210634394814, + "tokens_seen": 63242240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953660982948846, + "loss": 2.4536, + "theoretical_loss": 5.1234315482261135, + "tokens_seen": 63307776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953560682046139, + "loss": 2.4678, + "theoretical_loss": 5.122653493703571, + "tokens_seen": 63373312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953460381143431, + "loss": 2.5655, + "theoretical_loss": 5.121876468396373, + "tokens_seen": 63438848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953360080240723, + "loss": 2.5653, + "theoretical_loss": 5.121100469881933, + "tokens_seen": 63504384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953259779338014, + "loss": 2.454, + "theoretical_loss": 5.120325495745863, + "tokens_seen": 63569920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953159478435306, + "loss": 2.4739, + "theoretical_loss": 5.119551543581933, + "tokens_seen": 63635456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004953059177532598, + "loss": 2.5628, + "theoretical_loss": 5.118778610992038, + "tokens_seen": 63700992 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495295887662989, + "loss": 2.55, + "theoretical_loss": 5.118006695586154, + "tokens_seen": 63766528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952858575727182, + "loss": 2.4941, + "theoretical_loss": 5.1172357949823155, + "tokens_seen": 63832064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952758274824474, + "loss": 2.5923, + "theoretical_loss": 5.116465906806567, + "tokens_seen": 63897600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952657973921765, + "loss": 2.5904, + "theoretical_loss": 5.11569702869294, + "tokens_seen": 63963136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952557673019057, + "loss": 2.4012, + "theoretical_loss": 5.114929158283407, + "tokens_seen": 64028672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952457372116349, + "loss": 2.5246, + "theoretical_loss": 5.114162293227853, + "tokens_seen": 64094208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952357071213641, + "loss": 2.5447, + "theoretical_loss": 5.113396431184043, + "tokens_seen": 64159744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952256770310933, + "loss": 2.4555, + "theoretical_loss": 5.112631569817578, + "tokens_seen": 64225280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952156469408225, + "loss": 2.492, + "theoretical_loss": 5.111867706801872, + "tokens_seen": 64290816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004952056168505516, + "loss": 2.4369, + "theoretical_loss": 5.111104839818113, + "tokens_seen": 64356352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951955867602809, + "loss": 2.5594, + "theoretical_loss": 5.1103429665552245, + "tokens_seen": 64421888 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049518555667001, + "loss": 2.5938, + "theoretical_loss": 5.109582084709841, + "tokens_seen": 64487424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951755265797393, + "loss": 2.5882, + "theoretical_loss": 5.108822191986269, + "tokens_seen": 64552960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951654964894685, + "loss": 2.4155, + "theoretical_loss": 5.108063286096453, + "tokens_seen": 64618496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951554663991976, + "loss": 2.4994, + "theoretical_loss": 5.107305364759943, + "tokens_seen": 64684032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951454363089268, + "loss": 2.458, + "theoretical_loss": 5.106548425703869, + "tokens_seen": 64749568 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495135406218656, + "loss": 2.6211, + "theoretical_loss": 5.1057924666628915, + "tokens_seen": 64815104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951253761283852, + "loss": 2.5657, + "theoretical_loss": 5.105037485379189, + "tokens_seen": 64880640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951153460381144, + "loss": 2.6163, + "theoretical_loss": 5.104283479602407, + "tokens_seen": 64946176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004951053159478435, + "loss": 2.5555, + "theoretical_loss": 5.103530447089642, + "tokens_seen": 65011712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950952858575727, + "loss": 2.5246, + "theoretical_loss": 5.102778385605397, + "tokens_seen": 65077248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950852557673019, + "loss": 2.5383, + "theoretical_loss": 5.102027292921557, + "tokens_seen": 65142784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950752256770311, + "loss": 2.5669, + "theoretical_loss": 5.101277166817351, + "tokens_seen": 65208320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950651955867603, + "loss": 2.4133, + "theoretical_loss": 5.100528005079328, + "tokens_seen": 65273856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950551654964894, + "loss": 2.5005, + "theoretical_loss": 5.099779805501318, + "tokens_seen": 65339392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950451354062186, + "loss": 2.5157, + "theoretical_loss": 5.099032565884409, + "tokens_seen": 65404928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950351053159479, + "loss": 2.5642, + "theoretical_loss": 5.098286284036908, + "tokens_seen": 65470464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 31526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.6855673789978027, + "objective/train/theoretical_loss": 5.097540957774312, + "objective/train/tokens_used": 85996000, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.000495025075225677, + "loss": 2.601, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950150451354063, + "loss": 2.5722, + "theoretical_loss": 5.096796584919282, + "tokens_seen": 65601536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004950050150451354, + "loss": 2.5923, + "theoretical_loss": 5.096053163301608, + "tokens_seen": 65667072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949949849548646, + "loss": 2.6337, + "theoretical_loss": 5.095310690758177, + "tokens_seen": 65732608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949849548645938, + "loss": 2.5041, + "theoretical_loss": 5.094569165132949, + "tokens_seen": 65798144 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494974924774323, + "loss": 2.4249, + "theoretical_loss": 5.09382858427692, + "tokens_seen": 65863680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949648946840522, + "loss": 2.5546, + "theoretical_loss": 5.0930889460481, + "tokens_seen": 65929216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949548645937814, + "loss": 2.5086, + "theoretical_loss": 5.092350248311471, + "tokens_seen": 65994752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949448345035105, + "loss": 2.4848, + "theoretical_loss": 5.091612488938971, + "tokens_seen": 66060288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949348044132397, + "loss": 2.4684, + "theoretical_loss": 5.090875665809454, + "tokens_seen": 66125824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949247743229689, + "loss": 2.4157, + "theoretical_loss": 5.09013977680867, + "tokens_seen": 66191360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949147442326981, + "loss": 2.4993, + "theoretical_loss": 5.089404819829225, + "tokens_seen": 66256896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004949047141424273, + "loss": 2.5073, + "theoretical_loss": 5.088670792770563, + "tokens_seen": 66322432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948946840521565, + "loss": 2.4551, + "theoretical_loss": 5.087937693538927, + "tokens_seen": 66387968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948846539618856, + "loss": 2.4576, + "theoretical_loss": 5.0872055200473385, + "tokens_seen": 66453504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948746238716148, + "loss": 2.4314, + "theoretical_loss": 5.0864742702155645, + "tokens_seen": 66519040 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494864593781344, + "loss": 2.5878, + "theoretical_loss": 5.0857439419700885, + "tokens_seen": 66584576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948545636910733, + "loss": 2.4488, + "theoretical_loss": 5.085014533244088, + "tokens_seen": 66650112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948445336008024, + "loss": 2.4594, + "theoretical_loss": 5.084286041977396, + "tokens_seen": 66715648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948345035105316, + "loss": 2.4788, + "theoretical_loss": 5.083558466116487, + "tokens_seen": 66781184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948244734202607, + "loss": 2.3486, + "theoretical_loss": 5.082831803614433, + "tokens_seen": 66846720 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049481444332999, + "loss": 2.5019, + "theoretical_loss": 5.082106052430891, + "tokens_seen": 66912256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004948044132397192, + "loss": 2.443, + "theoretical_loss": 5.081381210532063, + "tokens_seen": 66977792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947943831494484, + "loss": 2.5506, + "theoretical_loss": 5.08065727589068, + "tokens_seen": 67043328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947843530591776, + "loss": 2.4703, + "theoretical_loss": 5.079934246485963, + "tokens_seen": 67108864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947743229689067, + "loss": 2.6505, + "theoretical_loss": 5.079212120303606, + "tokens_seen": 67174400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947642928786359, + "loss": 2.4162, + "theoretical_loss": 5.078490895335744, + "tokens_seen": 67239936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947542627883651, + "loss": 2.4924, + "theoretical_loss": 5.077770569580924, + "tokens_seen": 67305472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947442326980943, + "loss": 2.3657, + "theoretical_loss": 5.077051141044086, + "tokens_seen": 67371008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947342026078235, + "loss": 2.4142, + "theoretical_loss": 5.076332607736527, + "tokens_seen": 67436544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947241725175527, + "loss": 2.5191, + "theoretical_loss": 5.0756149676758815, + "tokens_seen": 67502080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947141424272818, + "loss": 2.6382, + "theoretical_loss": 5.0748982188860925, + "tokens_seen": 67567616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004947041123370111, + "loss": 2.5011, + "theoretical_loss": 5.074182359397387, + "tokens_seen": 67633152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946940822467402, + "loss": 2.3405, + "theoretical_loss": 5.073467387246247, + "tokens_seen": 67698688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946840521564695, + "loss": 2.4512, + "theoretical_loss": 5.072753300475385, + "tokens_seen": 67764224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946740220661987, + "loss": 2.453, + "theoretical_loss": 5.072040097133718, + "tokens_seen": 67829760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946639919759278, + "loss": 2.3675, + "theoretical_loss": 5.0713277752763455, + "tokens_seen": 67895296 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494653961885657, + "loss": 2.4094, + "theoretical_loss": 5.070616332964518, + "tokens_seen": 67960832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946439317953862, + "loss": 2.4472, + "theoretical_loss": 5.069905768265615, + "tokens_seen": 68026368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946339017051154, + "loss": 2.3908, + "theoretical_loss": 5.069196079253119, + "tokens_seen": 68091904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946238716148446, + "loss": 2.2599, + "theoretical_loss": 5.068487264006592, + "tokens_seen": 68157440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946138415245737, + "loss": 2.3426, + "theoretical_loss": 5.0677793206116455, + "tokens_seen": 68222976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004946038114343029, + "loss": 2.4983, + "theoretical_loss": 5.067072247159923, + "tokens_seen": 68288512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945937813440321, + "loss": 2.398, + "theoretical_loss": 5.06636604174907, + "tokens_seen": 68354048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945837512537613, + "loss": 2.4806, + "theoretical_loss": 5.065660702482711, + "tokens_seen": 68419584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945737211634905, + "loss": 2.4044, + "theoretical_loss": 5.064956227470425, + "tokens_seen": 68485120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945636910732196, + "loss": 2.4816, + "theoretical_loss": 5.06425261482772, + "tokens_seen": 68550656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945536609829488, + "loss": 2.5079, + "theoretical_loss": 5.063549862676011, + "tokens_seen": 68616192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945436308926781, + "loss": 2.3715, + "theoretical_loss": 5.062847969142591, + "tokens_seen": 68681728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945336008024072, + "loss": 2.4807, + "theoretical_loss": 5.062146932360616, + "tokens_seen": 68747264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 32629, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.361605167388916, + "objective/train/theoretical_loss": 5.061446750469072, + "objective/train/tokens_used": 89272800, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945235707121365, + "loss": 2.4421, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945135406218656, + "loss": 2.5426, + "theoretical_loss": 5.060747421612758, + "tokens_seen": 68878336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004945035105315948, + "loss": 2.5104, + "theoretical_loss": 5.060048943942254, + "tokens_seen": 68943872 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494493480441324, + "loss": 2.3779, + "theoretical_loss": 5.059351315613908, + "tokens_seen": 69009408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944834503510532, + "loss": 2.3525, + "theoretical_loss": 5.058654534789804, + "tokens_seen": 69074944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944734202607824, + "loss": 2.4548, + "theoretical_loss": 5.057958599637746, + "tokens_seen": 69140480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944633901705116, + "loss": 2.4524, + "theoretical_loss": 5.057263508331225, + "tokens_seen": 69206016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944533600802407, + "loss": 2.3883, + "theoretical_loss": 5.0565692590494065, + "tokens_seen": 69271552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944433299899699, + "loss": 2.3282, + "theoretical_loss": 5.055875849977101, + "tokens_seen": 69337088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944332998996991, + "loss": 2.5257, + "theoretical_loss": 5.055183279304744, + "tokens_seen": 69402624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944232698094283, + "loss": 2.4369, + "theoretical_loss": 5.0544915452283705, + "tokens_seen": 69468160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944132397191575, + "loss": 2.3919, + "theoretical_loss": 5.053800645949597, + "tokens_seen": 69533696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004944032096288867, + "loss": 2.3596, + "theoretical_loss": 5.053110579675593, + "tokens_seen": 69599232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943931795386158, + "loss": 2.4108, + "theoretical_loss": 5.052421344619064, + "tokens_seen": 69664768 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494383149448345, + "loss": 2.4562, + "theoretical_loss": 5.051732938998227, + "tokens_seen": 69730304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943731193580742, + "loss": 2.3607, + "theoretical_loss": 5.051045361036787, + "tokens_seen": 69795840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943630892678035, + "loss": 2.3991, + "theoretical_loss": 5.0503586089639185, + "tokens_seen": 69861376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943530591775326, + "loss": 2.376, + "theoretical_loss": 5.04967268101424, + "tokens_seen": 69926912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943430290872618, + "loss": 2.3931, + "theoretical_loss": 5.048987575427793, + "tokens_seen": 69992448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943329989969909, + "loss": 2.4497, + "theoretical_loss": 5.048303290450023, + "tokens_seen": 70057984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943229689067202, + "loss": 2.5376, + "theoretical_loss": 5.047619824331753, + "tokens_seen": 70123520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943129388164494, + "loss": 2.4006, + "theoretical_loss": 5.046937175329167, + "tokens_seen": 70189056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004943029087261786, + "loss": 2.47, + "theoretical_loss": 5.046255341703786, + "tokens_seen": 70254592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942928786359078, + "loss": 2.4115, + "theoretical_loss": 5.045574321722445, + "tokens_seen": 70320128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942828485456369, + "loss": 2.3817, + "theoretical_loss": 5.0448941136572785, + "tokens_seen": 70385664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942728184553661, + "loss": 2.4669, + "theoretical_loss": 5.044214715785689, + "tokens_seen": 70451200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942627883650953, + "loss": 2.3834, + "theoretical_loss": 5.043536126390334, + "tokens_seen": 70516736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942527582748245, + "loss": 2.2787, + "theoretical_loss": 5.042858343759107, + "tokens_seen": 70582272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942427281845537, + "loss": 2.3397, + "theoretical_loss": 5.0421813661851065, + "tokens_seen": 70647808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942326980942828, + "loss": 2.4088, + "theoretical_loss": 5.041505191966625, + "tokens_seen": 70713344 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494222668004012, + "loss": 2.4917, + "theoretical_loss": 5.040829819407123, + "tokens_seen": 70778880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942126379137412, + "loss": 2.4219, + "theoretical_loss": 5.040155246815212, + "tokens_seen": 70844416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004942026078234704, + "loss": 2.526, + "theoretical_loss": 5.039481472504633, + "tokens_seen": 70909952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941925777331996, + "loss": 2.4536, + "theoretical_loss": 5.038808494794232, + "tokens_seen": 70975488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941825476429289, + "loss": 2.3223, + "theoretical_loss": 5.038136312007948, + "tokens_seen": 71041024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941725175526579, + "loss": 2.3294, + "theoretical_loss": 5.037464922474786, + "tokens_seen": 71106560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941624874623872, + "loss": 2.3593, + "theoretical_loss": 5.036794324528802, + "tokens_seen": 71172096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941524573721163, + "loss": 2.3318, + "theoretical_loss": 5.036124516509076, + "tokens_seen": 71237632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941424272818456, + "loss": 2.5027, + "theoretical_loss": 5.035455496759701, + "tokens_seen": 71303168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941323971915748, + "loss": 2.4607, + "theoretical_loss": 5.034787263629758, + "tokens_seen": 71368704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941223671013039, + "loss": 2.4538, + "theoretical_loss": 5.0341198154732965, + "tokens_seen": 71434240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941123370110331, + "loss": 2.4948, + "theoretical_loss": 5.033453150649318, + "tokens_seen": 71499776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004941023069207623, + "loss": 2.3555, + "theoretical_loss": 5.032787267521752, + "tokens_seen": 71565312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940922768304915, + "loss": 2.414, + "theoretical_loss": 5.032122164459443, + "tokens_seen": 71630848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940822467402207, + "loss": 2.4486, + "theoretical_loss": 5.031457839836122, + "tokens_seen": 71696384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940722166499498, + "loss": 2.2785, + "theoretical_loss": 5.0307942920304, + "tokens_seen": 71761920 + }, + { + "epoch": 0.02, + "learning_rate": 0.000494062186559679, + "loss": 2.3769, + "theoretical_loss": 5.030131519425737, + "tokens_seen": 71827456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940521564694082, + "loss": 2.4637, + "theoretical_loss": 5.02946952041043, + "tokens_seen": 71892992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940421263791374, + "loss": 2.4214, + "theoretical_loss": 5.028808293377589, + "tokens_seen": 71958528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940320962888666, + "loss": 2.401, + "theoretical_loss": 5.028147836725127, + "tokens_seen": 72024064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 33987, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.3899569511413574, + "objective/train/theoretical_loss": 5.0274881488557295, + "objective/train/tokens_used": 92549600, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940220661985958, + "loss": 2.2586, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940120361083249, + "loss": 2.3959, + "theoretical_loss": 5.026829228176848, + "tokens_seen": 72155136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004940020060180542, + "loss": 2.3961, + "theoretical_loss": 5.026171073100672, + "tokens_seen": 72220672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939919759277834, + "loss": 2.3641, + "theoretical_loss": 5.025513682044114, + "tokens_seen": 72286208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939819458375126, + "loss": 2.5316, + "theoretical_loss": 5.024857053428793, + "tokens_seen": 72351744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939719157472418, + "loss": 2.3744, + "theoretical_loss": 5.024201185681015, + "tokens_seen": 72417280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493961885656971, + "loss": 2.4531, + "theoretical_loss": 5.023546077231752, + "tokens_seen": 72482816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939518555667001, + "loss": 2.3488, + "theoretical_loss": 5.0228917265166295, + "tokens_seen": 72548352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939418254764293, + "loss": 2.1993, + "theoretical_loss": 5.022238131975905, + "tokens_seen": 72613888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939317953861585, + "loss": 2.3809, + "theoretical_loss": 5.021585292054446, + "tokens_seen": 72679424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939217652958877, + "loss": 2.3737, + "theoretical_loss": 5.020933205201725, + "tokens_seen": 72744960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004939117352056169, + "loss": 2.4281, + "theoretical_loss": 5.020281869871787, + "tokens_seen": 72810496 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493901705115346, + "loss": 2.3805, + "theoretical_loss": 5.019631284523241, + "tokens_seen": 72876032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938916750250752, + "loss": 2.2917, + "theoretical_loss": 5.01898144761924, + "tokens_seen": 72941568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938816449348044, + "loss": 2.2502, + "theoretical_loss": 5.018332357627462, + "tokens_seen": 73007104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938716148445337, + "loss": 2.3468, + "theoretical_loss": 5.017684013020096, + "tokens_seen": 73072640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938615847542628, + "loss": 2.4905, + "theoretical_loss": 5.017036412273823, + "tokens_seen": 73138176 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493851554663992, + "loss": 2.4197, + "theoretical_loss": 5.016389553869796, + "tokens_seen": 73203712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938415245737211, + "loss": 2.4077, + "theoretical_loss": 5.0157434362936275, + "tokens_seen": 73269248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938314944834504, + "loss": 2.4081, + "theoretical_loss": 5.01509805803537, + "tokens_seen": 73334784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938214643931796, + "loss": 2.3962, + "theoretical_loss": 5.0144534175894995, + "tokens_seen": 73400320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004938114343029088, + "loss": 2.1944, + "theoretical_loss": 5.013809513454898, + "tokens_seen": 73465856 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493801404212638, + "loss": 2.2799, + "theoretical_loss": 5.013166344134838, + "tokens_seen": 73531392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937913741223671, + "loss": 2.3609, + "theoretical_loss": 5.012523908136963, + "tokens_seen": 73596928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937813440320963, + "loss": 2.3996, + "theoretical_loss": 5.011882203973277, + "tokens_seen": 73662464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937713139418255, + "loss": 2.2832, + "theoretical_loss": 5.011241230160119, + "tokens_seen": 73728000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937612838515547, + "loss": 2.5292, + "theoretical_loss": 5.0106009852181534, + "tokens_seen": 73793536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937512537612839, + "loss": 2.4843, + "theoretical_loss": 5.009961467672353, + "tokens_seen": 73859072 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493741223671013, + "loss": 2.3916, + "theoretical_loss": 5.009322676051976, + "tokens_seen": 73924608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937311935807422, + "loss": 2.3794, + "theoretical_loss": 5.008684608890562, + "tokens_seen": 73990144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937211634904714, + "loss": 2.4458, + "theoretical_loss": 5.008047264725901, + "tokens_seen": 74055680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937111334002006, + "loss": 2.2998, + "theoretical_loss": 5.007410642100032, + "tokens_seen": 74121216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004937011033099298, + "loss": 2.5725, + "theoretical_loss": 5.006774739559216, + "tokens_seen": 74186752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936910732196591, + "loss": 2.4183, + "theoretical_loss": 5.00613955565392, + "tokens_seen": 74252288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936810431293881, + "loss": 2.2647, + "theoretical_loss": 5.005505088938815, + "tokens_seen": 74317824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936710130391174, + "loss": 2.4314, + "theoretical_loss": 5.004871337972742, + "tokens_seen": 74383360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936609829488465, + "loss": 2.3269, + "theoretical_loss": 5.004238301318706, + "tokens_seen": 74448896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936509528585758, + "loss": 2.3726, + "theoretical_loss": 5.003605977543861, + "tokens_seen": 74514432 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493640922768305, + "loss": 2.3283, + "theoretical_loss": 5.00297436521949, + "tokens_seen": 74579968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936308926780341, + "loss": 2.283, + "theoretical_loss": 5.002343462920992, + "tokens_seen": 74645504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936208625877633, + "loss": 2.4802, + "theoretical_loss": 5.001713269227867, + "tokens_seen": 74711040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936108324974925, + "loss": 2.3634, + "theoretical_loss": 5.0010837827237005, + "tokens_seen": 74776576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004936008024072217, + "loss": 2.3726, + "theoretical_loss": 5.000455001996143, + "tokens_seen": 74842112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935907723169509, + "loss": 2.4063, + "theoretical_loss": 4.999826925636905, + "tokens_seen": 74907648 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049358074222668, + "loss": 2.4284, + "theoretical_loss": 4.999199552241732, + "tokens_seen": 74973184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935707121364092, + "loss": 2.519, + "theoretical_loss": 4.998572880410396, + "tokens_seen": 75038720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935606820461384, + "loss": 2.2857, + "theoretical_loss": 4.997946908746677, + "tokens_seen": 75104256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935506519558676, + "loss": 2.3496, + "theoretical_loss": 4.997321635858347, + "tokens_seen": 75169792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935406218655968, + "loss": 2.2919, + "theoretical_loss": 4.9966970603571585, + "tokens_seen": 75235328 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493530591775326, + "loss": 2.4327, + "theoretical_loss": 4.996073180858831, + "tokens_seen": 75300864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 35047, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.3895986080169678, + "objective/train/theoretical_loss": 4.995449995983028, + "objective/train/tokens_used": 95826400, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935205616850551, + "loss": 2.4051, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935105315947844, + "loss": 2.3766, + "theoretical_loss": 4.994827504353353, + "tokens_seen": 75431936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004935005015045135, + "loss": 2.3935, + "theoretical_loss": 4.994205704597325, + "tokens_seen": 75497472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934904714142428, + "loss": 2.2525, + "theoretical_loss": 4.993584595346373, + "tokens_seen": 75563008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934804413239719, + "loss": 2.4622, + "theoretical_loss": 4.992964175235814, + "tokens_seen": 75628544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934704112337011, + "loss": 2.4227, + "theoretical_loss": 4.992344442904843, + "tokens_seen": 75694080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934603811434303, + "loss": 2.2414, + "theoretical_loss": 4.991725396996515, + "tokens_seen": 75759616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934503510531595, + "loss": 2.2249, + "theoretical_loss": 4.991107036157738, + "tokens_seen": 75825152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934403209628887, + "loss": 2.4939, + "theoretical_loss": 4.990489359039248, + "tokens_seen": 75890688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934302908726179, + "loss": 2.4551, + "theoretical_loss": 4.989872364295605, + "tokens_seen": 75956224 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493420260782347, + "loss": 2.3287, + "theoretical_loss": 4.989256050585174, + "tokens_seen": 76021760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934102306920762, + "loss": 2.4987, + "theoretical_loss": 4.9886404165701075, + "tokens_seen": 76087296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004934002006018054, + "loss": 2.4824, + "theoretical_loss": 4.98802546091634, + "tokens_seen": 76152832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933901705115346, + "loss": 2.3886, + "theoretical_loss": 4.9874111822935685, + "tokens_seen": 76218368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933801404212638, + "loss": 2.4323, + "theoretical_loss": 4.986797579375239, + "tokens_seen": 76283904 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493370110330993, + "loss": 2.4224, + "theoretical_loss": 4.986184650838532, + "tokens_seen": 76349440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933600802407221, + "loss": 2.3331, + "theoretical_loss": 4.985572395364354, + "tokens_seen": 76414976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933500501504513, + "loss": 2.345, + "theoretical_loss": 4.98496081163732, + "tokens_seen": 76480512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933400200601805, + "loss": 2.3124, + "theoretical_loss": 4.984349898345735, + "tokens_seen": 76546048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933299899699098, + "loss": 2.3043, + "theoretical_loss": 4.98373965418159, + "tokens_seen": 76611584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933199598796389, + "loss": 2.3849, + "theoretical_loss": 4.983130077840543, + "tokens_seen": 76677120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004933099297893682, + "loss": 2.3238, + "theoretical_loss": 4.982521168021906, + "tokens_seen": 76742656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932998996990972, + "loss": 2.2674, + "theoretical_loss": 4.981912923428634, + "tokens_seen": 76808192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932898696088265, + "loss": 2.3456, + "theoretical_loss": 4.981305342767307, + "tokens_seen": 76873728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932798395185557, + "loss": 2.4043, + "theoretical_loss": 4.980698424748122, + "tokens_seen": 76939264 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932698094282849, + "loss": 2.3046, + "theoretical_loss": 4.980092168084878, + "tokens_seen": 77004800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932597793380141, + "loss": 2.3472, + "theoretical_loss": 4.97948657149496, + "tokens_seen": 77070336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932497492477432, + "loss": 2.3848, + "theoretical_loss": 4.978881633699331, + "tokens_seen": 77135872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932397191574724, + "loss": 2.2723, + "theoretical_loss": 4.9782773534225155, + "tokens_seen": 77201408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932296890672016, + "loss": 2.3407, + "theoretical_loss": 4.977673729392586, + "tokens_seen": 77266944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004932196589769308, + "loss": 2.305, + "theoretical_loss": 4.977070760341153, + "tokens_seen": 77332480 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049320962888666, + "loss": 2.3549, + "theoretical_loss": 4.976468445003352, + "tokens_seen": 77398016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931995987963893, + "loss": 2.3669, + "theoretical_loss": 4.975866782117828, + "tokens_seen": 77463552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931895687061183, + "loss": 2.2699, + "theoretical_loss": 4.975265770426724, + "tokens_seen": 77529088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931795386158476, + "loss": 2.3732, + "theoretical_loss": 4.974665408675667, + "tokens_seen": 77594624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931695085255767, + "loss": 2.3558, + "theoretical_loss": 4.974065695613763, + "tokens_seen": 77660160 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493159478435306, + "loss": 2.3124, + "theoretical_loss": 4.973466629993574, + "tokens_seen": 77725696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931494483450352, + "loss": 2.2116, + "theoretical_loss": 4.972868210571109, + "tokens_seen": 77791232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931394182547643, + "loss": 2.534, + "theoretical_loss": 4.972270436105815, + "tokens_seen": 77856768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931293881644935, + "loss": 2.2791, + "theoretical_loss": 4.971673305360563, + "tokens_seen": 77922304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931193580742227, + "loss": 2.2846, + "theoretical_loss": 4.971076817101633, + "tokens_seen": 77987840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004931093279839519, + "loss": 2.2618, + "theoretical_loss": 4.970480970098703, + "tokens_seen": 78053376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930992978936811, + "loss": 2.2019, + "theoretical_loss": 4.96988576312484, + "tokens_seen": 78118912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930892678034102, + "loss": 2.3872, + "theoretical_loss": 4.969291194956485, + "tokens_seen": 78184448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930792377131394, + "loss": 2.34, + "theoretical_loss": 4.968697264373438, + "tokens_seen": 78249984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930692076228686, + "loss": 2.1887, + "theoretical_loss": 4.968103970158854, + "tokens_seen": 78315520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930591775325978, + "loss": 2.3373, + "theoretical_loss": 4.967511311099223, + "tokens_seen": 78381056 + }, + { + "epoch": 0.02, + "learning_rate": 0.000493049147442327, + "loss": 2.1876, + "theoretical_loss": 4.96691928598436, + "tokens_seen": 78446592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930391173520562, + "loss": 2.3258, + "theoretical_loss": 4.966327893607398, + "tokens_seen": 78512128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930290872617853, + "loss": 2.3239, + "theoretical_loss": 4.96573713276477, + "tokens_seen": 78577664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 36308, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.390918254852295, + "objective/train/theoretical_loss": 4.965147002256201, + "objective/train/tokens_used": 99103200, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930190571715146, + "loss": 2.4039, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004930090270812437, + "loss": 2.3214, + "theoretical_loss": 4.964557500884695, + "tokens_seen": 78708736 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492998996990973, + "loss": 2.4351, + "theoretical_loss": 4.963968627456523, + "tokens_seen": 78774272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929889669007021, + "loss": 2.1843, + "theoretical_loss": 4.96338038078121, + "tokens_seen": 78839808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929789368104313, + "loss": 2.2595, + "theoretical_loss": 4.962792759671528, + "tokens_seen": 78905344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929689067201605, + "loss": 2.2838, + "theoretical_loss": 4.962205762943478, + "tokens_seen": 78970880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929588766298897, + "loss": 2.2167, + "theoretical_loss": 4.961619389416288, + "tokens_seen": 79036416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929488465396189, + "loss": 2.2225, + "theoretical_loss": 4.96103363791239, + "tokens_seen": 79101952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929388164493481, + "loss": 2.3226, + "theoretical_loss": 4.960448507257416, + "tokens_seen": 79167488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929287863590773, + "loss": 2.3252, + "theoretical_loss": 4.959863996280188, + "tokens_seen": 79233024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929187562688064, + "loss": 2.4479, + "theoretical_loss": 4.959280103812697, + "tokens_seen": 79298560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004929087261785356, + "loss": 2.2975, + "theoretical_loss": 4.958696828690105, + "tokens_seen": 79364096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928986960882648, + "loss": 2.2864, + "theoretical_loss": 4.958114169750726, + "tokens_seen": 79429632 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492888665997994, + "loss": 2.3081, + "theoretical_loss": 4.957532125836012, + "tokens_seen": 79495168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928786359077232, + "loss": 2.3915, + "theoretical_loss": 4.956950695790551, + "tokens_seen": 79560704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928686058174523, + "loss": 2.3031, + "theoretical_loss": 4.95636987846205, + "tokens_seen": 79626240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928585757271815, + "loss": 2.329, + "theoretical_loss": 4.955789672701323, + "tokens_seen": 79691776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928485456369107, + "loss": 2.3324, + "theoretical_loss": 4.955210077362285, + "tokens_seen": 79757312 + }, + { + "epoch": 0.02, + "learning_rate": 0.00049283851554664, + "loss": 2.415, + "theoretical_loss": 4.954631091301934, + "tokens_seen": 79822848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928284854563691, + "loss": 2.247, + "theoretical_loss": 4.954052713380349, + "tokens_seen": 79888384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928184553660984, + "loss": 2.2082, + "theoretical_loss": 4.953474942460673, + "tokens_seen": 79953920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004928084252758274, + "loss": 2.2591, + "theoretical_loss": 4.952897777409104, + "tokens_seen": 80019456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927983951855567, + "loss": 2.3059, + "theoretical_loss": 4.952321217094882, + "tokens_seen": 80084992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927883650952859, + "loss": 2.2099, + "theoretical_loss": 4.951745260390284, + "tokens_seen": 80150528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927783350050151, + "loss": 2.2248, + "theoretical_loss": 4.951169906170607, + "tokens_seen": 80216064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927683049147443, + "loss": 2.3071, + "theoretical_loss": 4.950595153314161, + "tokens_seen": 80281600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927582748244734, + "loss": 2.3146, + "theoretical_loss": 4.950021000702261, + "tokens_seen": 80347136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927482447342026, + "loss": 2.2434, + "theoretical_loss": 4.949447447219208, + "tokens_seen": 80412672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927382146439318, + "loss": 2.2631, + "theoretical_loss": 4.948874491752287, + "tokens_seen": 80478208 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492728184553661, + "loss": 2.1726, + "theoretical_loss": 4.9483021331917545, + "tokens_seen": 80543744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927181544633902, + "loss": 2.1731, + "theoretical_loss": 4.9477303704308255, + "tokens_seen": 80609280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004927081243731193, + "loss": 2.2811, + "theoretical_loss": 4.947159202365664, + "tokens_seen": 80674816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926980942828485, + "loss": 2.2465, + "theoretical_loss": 4.946588627895375, + "tokens_seen": 80740352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926880641925777, + "loss": 2.2483, + "theoretical_loss": 4.946018645921994, + "tokens_seen": 80805888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926780341023069, + "loss": 2.362, + "theoretical_loss": 4.945449255350473, + "tokens_seen": 80871424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926680040120361, + "loss": 2.3309, + "theoretical_loss": 4.944880455088677, + "tokens_seen": 80936960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926579739217654, + "loss": 2.3405, + "theoretical_loss": 4.944312244047365, + "tokens_seen": 81002496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926479438314944, + "loss": 2.2101, + "theoretical_loss": 4.943744621140189, + "tokens_seen": 81068032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926379137412237, + "loss": 2.1922, + "theoretical_loss": 4.94317758528368, + "tokens_seen": 81133568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926278836509528, + "loss": 2.3443, + "theoretical_loss": 4.942611135397234, + "tokens_seen": 81199104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926178535606821, + "loss": 2.3674, + "theoretical_loss": 4.942045270403112, + "tokens_seen": 81264640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004926078234704113, + "loss": 2.2895, + "theoretical_loss": 4.941479989226421, + "tokens_seen": 81330176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925977933801404, + "loss": 2.4029, + "theoretical_loss": 4.940915290795106, + "tokens_seen": 81395712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925877632898696, + "loss": 2.3193, + "theoretical_loss": 4.940351174039947, + "tokens_seen": 81461248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925777331995988, + "loss": 2.3438, + "theoretical_loss": 4.939787637894537, + "tokens_seen": 81526784 + }, + { + "epoch": 0.02, + "learning_rate": 0.000492567703109328, + "loss": 2.1573, + "theoretical_loss": 4.939224681295286, + "tokens_seen": 81592320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925576730190572, + "loss": 2.3019, + "theoretical_loss": 4.938662303181399, + "tokens_seen": 81657856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925476429287864, + "loss": 2.4629, + "theoretical_loss": 4.938100502494878, + "tokens_seen": 81723392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925376128385155, + "loss": 2.382, + "theoretical_loss": 4.9375392781805, + "tokens_seen": 81788928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925275827482447, + "loss": 2.271, + "theoretical_loss": 4.936978629185818, + "tokens_seen": 81854464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 37569, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.225572347640991, + "objective/train/theoretical_loss": 4.936418554461147, + "objective/train/tokens_used": 102380000, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925175526579739, + "loss": 2.2991, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004925075225677031, + "loss": 2.3343, + "theoretical_loss": 4.935859052959554, + "tokens_seen": 81985536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924974924774323, + "loss": 2.158, + "theoretical_loss": 4.935300123636851, + "tokens_seen": 82051072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924874623871615, + "loss": 2.3066, + "theoretical_loss": 4.93474176545158, + "tokens_seen": 82116608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924774322968907, + "loss": 2.2955, + "theoretical_loss": 4.934183977365015, + "tokens_seen": 82182144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924674022066199, + "loss": 2.1906, + "theoretical_loss": 4.9336267583411395, + "tokens_seen": 82247680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924573721163491, + "loss": 2.1904, + "theoretical_loss": 4.933070107346644, + "tokens_seen": 82313216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924473420260783, + "loss": 2.228, + "theoretical_loss": 4.932514023350921, + "tokens_seen": 82378752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0004924373119358075, + "loss": 2.3377, + "theoretical_loss": 4.931958505326042, + "tokens_seen": 82444288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004924272818455366, + "loss": 2.2382, + "theoretical_loss": 4.931403552246769, + "tokens_seen": 82509824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004924172517552658, + "loss": 2.2852, + "theoretical_loss": 4.930849163090521, + "tokens_seen": 82575360 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492407221664995, + "loss": 2.2623, + "theoretical_loss": 4.930295336837387, + "tokens_seen": 82640896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923971915747242, + "loss": 2.2127, + "theoretical_loss": 4.9297420724701055, + "tokens_seen": 82706432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923871614844534, + "loss": 2.2778, + "theoretical_loss": 4.929189368974054, + "tokens_seen": 82771968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923771313941825, + "loss": 2.2017, + "theoretical_loss": 4.928637225337246, + "tokens_seen": 82837504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923671013039117, + "loss": 2.3392, + "theoretical_loss": 4.9280856405503215, + "tokens_seen": 82903040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923570712136409, + "loss": 2.3687, + "theoretical_loss": 4.927534613606534, + "tokens_seen": 82968576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923470411233702, + "loss": 2.1749, + "theoretical_loss": 4.926984143501745, + "tokens_seen": 83034112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923370110330993, + "loss": 2.2705, + "theoretical_loss": 4.926434229234411, + "tokens_seen": 83099648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923269809428286, + "loss": 2.2625, + "theoretical_loss": 4.925884869805586, + "tokens_seen": 83165184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923169508525576, + "loss": 2.1972, + "theoretical_loss": 4.925336064218893, + "tokens_seen": 83230720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004923069207622869, + "loss": 2.3238, + "theoretical_loss": 4.924787811480538, + "tokens_seen": 83296256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922968906720161, + "loss": 2.307, + "theoretical_loss": 4.9242401105992855, + "tokens_seen": 83361792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922868605817453, + "loss": 2.3206, + "theoretical_loss": 4.923692960586452, + "tokens_seen": 83427328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922768304914745, + "loss": 2.0572, + "theoretical_loss": 4.923146360455906, + "tokens_seen": 83492864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922668004012036, + "loss": 2.1891, + "theoretical_loss": 4.92260030922405, + "tokens_seen": 83558400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922567703109328, + "loss": 2.4244, + "theoretical_loss": 4.922054805909816, + "tokens_seen": 83623936 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492246740220662, + "loss": 2.2269, + "theoretical_loss": 4.921509849534658, + "tokens_seen": 83689472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922367101303912, + "loss": 2.2411, + "theoretical_loss": 4.920965439122538, + "tokens_seen": 83755008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922266800401204, + "loss": 2.2088, + "theoretical_loss": 4.9204215736999295, + "tokens_seen": 83820544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922166499498495, + "loss": 2.1803, + "theoretical_loss": 4.919878252295793, + "tokens_seen": 83886080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004922066198595787, + "loss": 2.302, + "theoretical_loss": 4.919335473941581, + "tokens_seen": 83951616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921965897693079, + "loss": 2.3513, + "theoretical_loss": 4.918793237671224, + "tokens_seen": 84017152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921865596790371, + "loss": 2.2708, + "theoretical_loss": 4.918251542521121, + "tokens_seen": 84082688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921765295887663, + "loss": 2.3117, + "theoretical_loss": 4.9177103875301365, + "tokens_seen": 84148224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921664994984956, + "loss": 2.2367, + "theoretical_loss": 4.917169771739586, + "tokens_seen": 84213760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921564694082246, + "loss": 2.1806, + "theoretical_loss": 4.916629694193233, + "tokens_seen": 84279296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921464393179539, + "loss": 2.2496, + "theoretical_loss": 4.916090153937276, + "tokens_seen": 84344832 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492136409227683, + "loss": 2.3679, + "theoretical_loss": 4.915551150020346, + "tokens_seen": 84410368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921263791374123, + "loss": 2.4162, + "theoretical_loss": 4.915012681493495, + "tokens_seen": 84475904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921163490471415, + "loss": 2.3295, + "theoretical_loss": 4.914474747410187, + "tokens_seen": 84541440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004921063189568706, + "loss": 2.1671, + "theoretical_loss": 4.913937346826293, + "tokens_seen": 84606976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920962888665998, + "loss": 2.2696, + "theoretical_loss": 4.91340047880008, + "tokens_seen": 84672512 + }, + { + "epoch": 0.03, + "learning_rate": 0.000492086258776329, + "loss": 2.2297, + "theoretical_loss": 4.9128641423922055, + "tokens_seen": 84738048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920762286860582, + "loss": 2.2356, + "theoretical_loss": 4.912328336665711, + "tokens_seen": 84803584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920661985957874, + "loss": 2.2053, + "theoretical_loss": 4.911793060686007, + "tokens_seen": 84869120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920561685055166, + "loss": 2.3342, + "theoretical_loss": 4.9112583135208725, + "tokens_seen": 84934656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920461384152457, + "loss": 2.3374, + "theoretical_loss": 4.910724094240447, + "tokens_seen": 85000192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920361083249749, + "loss": 2.2217, + "theoretical_loss": 4.910190401917214, + "tokens_seen": 85065728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920260782347041, + "loss": 2.3589, + "theoretical_loss": 4.909657235626009, + "tokens_seen": 85131264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 38678, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.262509346008301, + "objective/train/theoretical_loss": 4.909124594443993, + "objective/train/tokens_used": 105656800, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920160481444333, + "loss": 2.2546, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004920060180541625, + "loss": 2.2233, + "theoretical_loss": 4.908592477450659, + "tokens_seen": 85262336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919959879638916, + "loss": 2.2851, + "theoretical_loss": 4.908060883727821, + "tokens_seen": 85327872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919859578736209, + "loss": 2.2246, + "theoretical_loss": 4.9075298123596, + "tokens_seen": 85393408 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049197592778335, + "loss": 2.1563, + "theoretical_loss": 4.906999262432427, + "tokens_seen": 85458944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919658976930793, + "loss": 2.2046, + "theoretical_loss": 4.906469233035025, + "tokens_seen": 85524480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919558676028084, + "loss": 2.3753, + "theoretical_loss": 4.905939723258408, + "tokens_seen": 85590016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919458375125377, + "loss": 2.2753, + "theoretical_loss": 4.905410732195875, + "tokens_seen": 85655552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919358074222668, + "loss": 2.2728, + "theoretical_loss": 4.904882258942994, + "tokens_seen": 85721088 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491925777331996, + "loss": 2.1419, + "theoretical_loss": 4.904354302597603, + "tokens_seen": 85786624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919157472417252, + "loss": 2.3265, + "theoretical_loss": 4.903826862259798, + "tokens_seen": 85852160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004919057171514544, + "loss": 2.2675, + "theoretical_loss": 4.90329993703193, + "tokens_seen": 85917696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918956870611836, + "loss": 2.2532, + "theoretical_loss": 4.902773526018591, + "tokens_seen": 85983232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918856569709127, + "loss": 2.2299, + "theoretical_loss": 4.902247628326613, + "tokens_seen": 86048768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918756268806419, + "loss": 2.2988, + "theoretical_loss": 4.9017222430650556, + "tokens_seen": 86114304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918655967903711, + "loss": 2.2983, + "theoretical_loss": 4.901197369345205, + "tokens_seen": 86179840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918555667001003, + "loss": 2.1239, + "theoretical_loss": 4.9006730062805595, + "tokens_seen": 86245376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918455366098295, + "loss": 2.2667, + "theoretical_loss": 4.9001491529868275, + "tokens_seen": 86310912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918355065195586, + "loss": 2.1387, + "theoretical_loss": 4.899625808581918, + "tokens_seen": 86376448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918254764292878, + "loss": 2.3589, + "theoretical_loss": 4.899102972185936, + "tokens_seen": 86441984 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491815446339017, + "loss": 2.1358, + "theoretical_loss": 4.898580642921171, + "tokens_seen": 86507520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004918054162487463, + "loss": 2.196, + "theoretical_loss": 4.898058819912093, + "tokens_seen": 86573056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917953861584754, + "loss": 2.3319, + "theoretical_loss": 4.897537502285347, + "tokens_seen": 86638592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917853560682047, + "loss": 2.3595, + "theoretical_loss": 4.897016689169741, + "tokens_seen": 86704128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917753259779337, + "loss": 2.1427, + "theoretical_loss": 4.896496379696243, + "tokens_seen": 86769664 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491765295887663, + "loss": 2.2472, + "theoretical_loss": 4.895976572997973, + "tokens_seen": 86835200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917552657973922, + "loss": 2.2138, + "theoretical_loss": 4.895457268210196, + "tokens_seen": 86900736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917452357071214, + "loss": 2.2447, + "theoretical_loss": 4.8949384644703136, + "tokens_seen": 86966272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917352056168506, + "loss": 2.1805, + "theoretical_loss": 4.894420160917861, + "tokens_seen": 87031808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917251755265797, + "loss": 2.3462, + "theoretical_loss": 4.893902356694495, + "tokens_seen": 87097344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917151454363089, + "loss": 2.1486, + "theoretical_loss": 4.8933850509439925, + "tokens_seen": 87162880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004917051153460381, + "loss": 2.2719, + "theoretical_loss": 4.89286824281224, + "tokens_seen": 87228416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916950852557673, + "loss": 2.1561, + "theoretical_loss": 4.892351931447224, + "tokens_seen": 87293952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916850551654965, + "loss": 2.2226, + "theoretical_loss": 4.891836115999034, + "tokens_seen": 87359488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916750250752258, + "loss": 2.1581, + "theoretical_loss": 4.891320795619847, + "tokens_seen": 87425024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916649949849548, + "loss": 2.1783, + "theoretical_loss": 4.890805969463923, + "tokens_seen": 87490560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916549648946841, + "loss": 2.2491, + "theoretical_loss": 4.8902916366876, + "tokens_seen": 87556096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916449348044132, + "loss": 2.17, + "theoretical_loss": 4.889777796449285, + "tokens_seen": 87621632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916349047141425, + "loss": 2.1803, + "theoretical_loss": 4.889264447909451, + "tokens_seen": 87687168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916248746238717, + "loss": 2.1781, + "theoretical_loss": 4.888751590230626, + "tokens_seen": 87752704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004916148445336008, + "loss": 2.1858, + "theoretical_loss": 4.888239222577389, + "tokens_seen": 87818240 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049160481444333, + "loss": 2.31, + "theoretical_loss": 4.887727344116364, + "tokens_seen": 87883776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915947843530592, + "loss": 2.2325, + "theoretical_loss": 4.8872159540162095, + "tokens_seen": 87949312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915847542627884, + "loss": 2.2435, + "theoretical_loss": 4.88670505144762, + "tokens_seen": 88014848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915747241725176, + "loss": 2.2201, + "theoretical_loss": 4.8861946355833075, + "tokens_seen": 88080384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915646940822468, + "loss": 2.2519, + "theoretical_loss": 4.885684705598009, + "tokens_seen": 88145920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915546639919759, + "loss": 2.1957, + "theoretical_loss": 4.885175260668468, + "tokens_seen": 88211456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915446339017051, + "loss": 2.1849, + "theoretical_loss": 4.884666299973436, + "tokens_seen": 88276992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915346038114343, + "loss": 2.3223, + "theoretical_loss": 4.884157822693663, + "tokens_seen": 88342528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915245737211635, + "loss": 2.1226, + "theoretical_loss": 4.88364982801189, + "tokens_seen": 88408064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 39491, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1248512268066406, + "objective/train/theoretical_loss": 4.8831423151128455, + "objective/train/tokens_used": 108933600, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915145436308927, + "loss": 2.2258, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004915045135406218, + "loss": 2.2099, + "theoretical_loss": 4.8826352831832365, + "tokens_seen": 88539136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914944834503511, + "loss": 2.261, + "theoretical_loss": 4.882128731411745, + "tokens_seen": 88604672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914844533600802, + "loss": 2.2904, + "theoretical_loss": 4.881622658989022, + "tokens_seen": 88670208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914744232698095, + "loss": 2.1795, + "theoretical_loss": 4.881117065107675, + "tokens_seen": 88735744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914643931795386, + "loss": 2.1256, + "theoretical_loss": 4.880611948962271, + "tokens_seen": 88801280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914543630892679, + "loss": 2.2691, + "theoretical_loss": 4.880107309749322, + "tokens_seen": 88866816 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491444332998997, + "loss": 2.2427, + "theoretical_loss": 4.879603146667284, + "tokens_seen": 88932352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914343029087262, + "loss": 2.243, + "theoretical_loss": 4.879099458916553, + "tokens_seen": 88997888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914242728184554, + "loss": 2.1452, + "theoretical_loss": 4.878596245699449, + "tokens_seen": 89063424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914142427281846, + "loss": 2.3737, + "theoretical_loss": 4.878093506220221, + "tokens_seen": 89128960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004914042126379138, + "loss": 2.2804, + "theoretical_loss": 4.877591239685035, + "tokens_seen": 89194496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913941825476429, + "loss": 2.1363, + "theoretical_loss": 4.877089445301969, + "tokens_seen": 89260032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913841524573721, + "loss": 2.1141, + "theoretical_loss": 4.876588122281007, + "tokens_seen": 89325568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913741223671013, + "loss": 2.2339, + "theoretical_loss": 4.876087269834034, + "tokens_seen": 89391104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913640922768305, + "loss": 2.227, + "theoretical_loss": 4.875586887174828, + "tokens_seen": 89456640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913540621865597, + "loss": 2.2483, + "theoretical_loss": 4.875086973519058, + "tokens_seen": 89522176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913440320962888, + "loss": 2.2427, + "theoretical_loss": 4.874587528084271, + "tokens_seen": 89587712 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491334002006018, + "loss": 2.2148, + "theoretical_loss": 4.874088550089896, + "tokens_seen": 89653248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913239719157472, + "loss": 2.2367, + "theoretical_loss": 4.873590038757229, + "tokens_seen": 89718784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913139418254765, + "loss": 2.0677, + "theoretical_loss": 4.87309199330943, + "tokens_seen": 89784320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004913039117352056, + "loss": 2.0532, + "theoretical_loss": 4.872594412971521, + "tokens_seen": 89849856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912938816449349, + "loss": 2.2485, + "theoretical_loss": 4.872097296970376, + "tokens_seen": 89915392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912838515546639, + "loss": 2.2631, + "theoretical_loss": 4.871600644534717, + "tokens_seen": 89980928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912738214643932, + "loss": 2.1852, + "theoretical_loss": 4.871104454895106, + "tokens_seen": 90046464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912637913741224, + "loss": 2.2026, + "theoretical_loss": 4.870608727283942, + "tokens_seen": 90112000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912537612838516, + "loss": 2.468, + "theoretical_loss": 4.870113460935455, + "tokens_seen": 90177536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912437311935808, + "loss": 2.172, + "theoretical_loss": 4.869618655085697, + "tokens_seen": 90243072 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049123370110331, + "loss": 2.1622, + "theoretical_loss": 4.869124308972543, + "tokens_seen": 90308608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912236710130391, + "loss": 2.2891, + "theoretical_loss": 4.868630421835677, + "tokens_seen": 90374144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912136409227683, + "loss": 2.2721, + "theoretical_loss": 4.868136992916593, + "tokens_seen": 90439680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004912036108324975, + "loss": 2.2764, + "theoretical_loss": 4.867644021458585, + "tokens_seen": 90505216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911935807422267, + "loss": 2.0747, + "theoretical_loss": 4.8671515067067475, + "tokens_seen": 90570752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911835506519559, + "loss": 2.307, + "theoretical_loss": 4.866659447907962, + "tokens_seen": 90636288 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491173520561685, + "loss": 2.1087, + "theoretical_loss": 4.866167844310896, + "tokens_seen": 90701824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911634904714142, + "loss": 2.1647, + "theoretical_loss": 4.865676695165996, + "tokens_seen": 90767360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911534603811434, + "loss": 2.1439, + "theoretical_loss": 4.865185999725487, + "tokens_seen": 90832896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911434302908726, + "loss": 2.2423, + "theoretical_loss": 4.864695757243359, + "tokens_seen": 90898432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911334002006019, + "loss": 2.3074, + "theoretical_loss": 4.864205966975364, + "tokens_seen": 90963968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911233701103309, + "loss": 2.1423, + "theoretical_loss": 4.8637166281790165, + "tokens_seen": 91029504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911133400200602, + "loss": 2.2531, + "theoretical_loss": 4.86322774011358, + "tokens_seen": 91095040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004911033099297893, + "loss": 2.2293, + "theoretical_loss": 4.862739302040067, + "tokens_seen": 91160576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910932798395186, + "loss": 2.1055, + "theoretical_loss": 4.862251313221231, + "tokens_seen": 91226112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910832497492478, + "loss": 2.3372, + "theoretical_loss": 4.861763772921561, + "tokens_seen": 91291648 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491073219658977, + "loss": 2.2232, + "theoretical_loss": 4.861276680407279, + "tokens_seen": 91357184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910631895687061, + "loss": 2.2952, + "theoretical_loss": 4.860790034946331, + "tokens_seen": 91422720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910531594784353, + "loss": 2.2261, + "theoretical_loss": 4.860303835808384, + "tokens_seen": 91488256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910431293881645, + "loss": 2.1302, + "theoretical_loss": 4.859818082264822, + "tokens_seen": 91553792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910330992978937, + "loss": 2.1862, + "theoretical_loss": 4.859332773588736, + "tokens_seen": 91619328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910230692076229, + "loss": 2.1627, + "theoretical_loss": 4.858847909054923, + "tokens_seen": 91684864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 40748, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0513906478881836, + "objective/train/theoretical_loss": 4.858363487939879, + "objective/train/tokens_used": 112210400, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.000491013039117352, + "loss": 2.1639, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004910030090270812, + "loss": 2.1745, + "theoretical_loss": 4.857879509521796, + "tokens_seen": 91815936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909929789368104, + "loss": 2.1931, + "theoretical_loss": 4.857395973080552, + "tokens_seen": 91881472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909829488465397, + "loss": 2.238, + "theoretical_loss": 4.856912877897713, + "tokens_seen": 91947008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909729187562688, + "loss": 2.1722, + "theoretical_loss": 4.85643022325652, + "tokens_seen": 92012544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909628886659981, + "loss": 2.1344, + "theoretical_loss": 4.855948008441891, + "tokens_seen": 92078080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909528585757272, + "loss": 2.2092, + "theoretical_loss": 4.85546623274041, + "tokens_seen": 92143616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909428284854564, + "loss": 2.0946, + "theoretical_loss": 4.854984895440325, + "tokens_seen": 92209152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909327983951856, + "loss": 2.1771, + "theoretical_loss": 4.854503995831545, + "tokens_seen": 92274688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909227683049148, + "loss": 2.3036, + "theoretical_loss": 4.854023533205627, + "tokens_seen": 92340224 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490912738214644, + "loss": 2.2324, + "theoretical_loss": 4.853543506855784, + "tokens_seen": 92405760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004909027081243731, + "loss": 2.3006, + "theoretical_loss": 4.853063916076865, + "tokens_seen": 92471296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908926780341023, + "loss": 2.0872, + "theoretical_loss": 4.852584760165362, + "tokens_seen": 92536832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908826479438315, + "loss": 2.0797, + "theoretical_loss": 4.8521060384194, + "tokens_seen": 92602368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908726178535607, + "loss": 2.1653, + "theoretical_loss": 4.851627750138731, + "tokens_seen": 92667904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908625877632899, + "loss": 2.2153, + "theoretical_loss": 4.851149894624731, + "tokens_seen": 92733440 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490852557673019, + "loss": 2.2606, + "theoretical_loss": 4.8506724711803955, + "tokens_seen": 92798976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908425275827482, + "loss": 2.0095, + "theoretical_loss": 4.850195479110333, + "tokens_seen": 92864512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908324974924774, + "loss": 2.2392, + "theoretical_loss": 4.849718917720762, + "tokens_seen": 92930048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908224674022067, + "loss": 2.2521, + "theoretical_loss": 4.849242786319503, + "tokens_seen": 92995584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908124373119358, + "loss": 2.1778, + "theoretical_loss": 4.848767084215979, + "tokens_seen": 93061120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004908024072216651, + "loss": 2.1339, + "theoretical_loss": 4.848291810721204, + "tokens_seen": 93126656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907923771313941, + "loss": 2.1599, + "theoretical_loss": 4.847816965147783, + "tokens_seen": 93192192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907823470411234, + "loss": 2.2041, + "theoretical_loss": 4.847342546809907, + "tokens_seen": 93257728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907723169508526, + "loss": 2.117, + "theoretical_loss": 4.846868555023345, + "tokens_seen": 93323264 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907622868605818, + "loss": 2.2127, + "theoretical_loss": 4.846394989105444, + "tokens_seen": 93388800 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490752256770311, + "loss": 2.1791, + "theoretical_loss": 4.845921848375118, + "tokens_seen": 93454336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907422266800401, + "loss": 2.1601, + "theoretical_loss": 4.845449132152851, + "tokens_seen": 93519872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907321965897693, + "loss": 2.206, + "theoretical_loss": 4.844976839760683, + "tokens_seen": 93585408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907221664994985, + "loss": 2.1621, + "theoretical_loss": 4.844504970522217, + "tokens_seen": 93650944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907121364092277, + "loss": 2.1446, + "theoretical_loss": 4.844033523762603, + "tokens_seen": 93716480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004907021063189569, + "loss": 2.1451, + "theoretical_loss": 4.843562498808538, + "tokens_seen": 93782016 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490692076228686, + "loss": 2.3297, + "theoretical_loss": 4.843091894988265, + "tokens_seen": 93847552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906820461384152, + "loss": 2.0289, + "theoretical_loss": 4.842621711631564, + "tokens_seen": 93913088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906720160481444, + "loss": 2.2873, + "theoretical_loss": 4.842151948069745, + "tokens_seen": 93978624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906619859578736, + "loss": 2.0635, + "theoretical_loss": 4.84168260363565, + "tokens_seen": 94044160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906519558676028, + "loss": 2.1303, + "theoretical_loss": 4.841213677663648, + "tokens_seen": 94109696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906419257773321, + "loss": 2.1397, + "theoretical_loss": 4.840745169489621, + "tokens_seen": 94175232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906318956870611, + "loss": 2.1282, + "theoretical_loss": 4.840277078450969, + "tokens_seen": 94240768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906218655967904, + "loss": 2.173, + "theoretical_loss": 4.839809403886606, + "tokens_seen": 94306304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906118355065195, + "loss": 2.0226, + "theoretical_loss": 4.839342145136947, + "tokens_seen": 94371840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004906018054162488, + "loss": 2.1261, + "theoretical_loss": 4.838875301543913, + "tokens_seen": 94437376 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490591775325978, + "loss": 2.1357, + "theoretical_loss": 4.838408872450919, + "tokens_seen": 94502912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905817452357072, + "loss": 2.2417, + "theoretical_loss": 4.837942857202876, + "tokens_seen": 94568448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905717151454363, + "loss": 2.0694, + "theoretical_loss": 4.837477255146178, + "tokens_seen": 94633984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905616850551655, + "loss": 2.1705, + "theoretical_loss": 4.837012065628711, + "tokens_seen": 94699520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905516549648947, + "loss": 2.0234, + "theoretical_loss": 4.836547287999833, + "tokens_seen": 94765056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905416248746239, + "loss": 2.1074, + "theoretical_loss": 4.836082921610381, + "tokens_seen": 94830592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905315947843531, + "loss": 2.1587, + "theoretical_loss": 4.835618965812664, + "tokens_seen": 94896128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905215646940822, + "loss": 2.1655, + "theoretical_loss": 4.835155419960453, + "tokens_seen": 94961664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 42039, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9614694118499756, + "objective/train/theoretical_loss": 4.8346922834089865, + "objective/train/tokens_used": 115487200, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905115346038114, + "loss": 2.0634, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004905015045135406, + "loss": 2.1116, + "theoretical_loss": 4.834229555514956, + "tokens_seen": 95092736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904914744232698, + "loss": 2.1224, + "theoretical_loss": 4.833767235636511, + "tokens_seen": 95158272 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490481444332999, + "loss": 2.1066, + "theoretical_loss": 4.833305323133246, + "tokens_seen": 95223808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904714142427281, + "loss": 2.1791, + "theoretical_loss": 4.832843817366204, + "tokens_seen": 95289344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904613841524574, + "loss": 2.1526, + "theoretical_loss": 4.832382717697866, + "tokens_seen": 95354880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904513540621865, + "loss": 2.1619, + "theoretical_loss": 4.831922023492153, + "tokens_seen": 95420416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904413239719158, + "loss": 2.1402, + "theoretical_loss": 4.831461734114415, + "tokens_seen": 95485952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904312938816449, + "loss": 2.2027, + "theoretical_loss": 4.831001848931429, + "tokens_seen": 95551488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904212637913742, + "loss": 2.1762, + "theoretical_loss": 4.830542367311402, + "tokens_seen": 95617024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904112337011033, + "loss": 2.236, + "theoretical_loss": 4.830083288623955, + "tokens_seen": 95682560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004904012036108325, + "loss": 2.124, + "theoretical_loss": 4.829624612240126, + "tokens_seen": 95748096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903911735205617, + "loss": 2.2121, + "theoretical_loss": 4.829166337532367, + "tokens_seen": 95813632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903811434302909, + "loss": 2.123, + "theoretical_loss": 4.828708463874532, + "tokens_seen": 95879168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903711133400201, + "loss": 2.1793, + "theoretical_loss": 4.828250990641886, + "tokens_seen": 95944704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903610832497492, + "loss": 2.1689, + "theoretical_loss": 4.827793917211083, + "tokens_seen": 96010240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903510531594784, + "loss": 2.2186, + "theoretical_loss": 4.827337242960181, + "tokens_seen": 96075776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903410230692076, + "loss": 1.9967, + "theoretical_loss": 4.8268809672686235, + "tokens_seen": 96141312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903309929789368, + "loss": 2.0479, + "theoretical_loss": 4.826425089517245, + "tokens_seen": 96206848 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490320962888666, + "loss": 2.1069, + "theoretical_loss": 4.825969609088258, + "tokens_seen": 96272384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903109327983952, + "loss": 2.2513, + "theoretical_loss": 4.825514525365258, + "tokens_seen": 96337920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004903009027081243, + "loss": 2.1097, + "theoretical_loss": 4.825059837733214, + "tokens_seen": 96403456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902908726178535, + "loss": 2.12, + "theoretical_loss": 4.824605545578463, + "tokens_seen": 96468992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902808425275828, + "loss": 2.121, + "theoretical_loss": 4.824151648288714, + "tokens_seen": 96534528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902708124373119, + "loss": 2.0901, + "theoretical_loss": 4.823698145253035, + "tokens_seen": 96600064 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902607823470412, + "loss": 2.2418, + "theoretical_loss": 4.823245035861852, + "tokens_seen": 96665600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902507522567703, + "loss": 2.2021, + "theoretical_loss": 4.822792319506952, + "tokens_seen": 96731136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902407221664995, + "loss": 2.1308, + "theoretical_loss": 4.822339995581466, + "tokens_seen": 96796672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902306920762287, + "loss": 2.1137, + "theoretical_loss": 4.821888063479874, + "tokens_seen": 96862208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902206619859579, + "loss": 2.2492, + "theoretical_loss": 4.8214365225980025, + "tokens_seen": 96927744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902106318956871, + "loss": 2.1254, + "theoretical_loss": 4.8209853723330145, + "tokens_seen": 96993280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004902006018054163, + "loss": 2.0455, + "theoretical_loss": 4.820534612083408, + "tokens_seen": 97058816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901905717151454, + "loss": 2.198, + "theoretical_loss": 4.820084241249013, + "tokens_seen": 97124352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901805416248746, + "loss": 2.1138, + "theoretical_loss": 4.819634259230991, + "tokens_seen": 97189888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901705115346038, + "loss": 2.0338, + "theoretical_loss": 4.819184665431821, + "tokens_seen": 97255424 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490160481444333, + "loss": 2.0277, + "theoretical_loss": 4.818735459255306, + "tokens_seen": 97320960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901504513540623, + "loss": 2.177, + "theoretical_loss": 4.818286640106564, + "tokens_seen": 97386496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901404212637913, + "loss": 2.3246, + "theoretical_loss": 4.817838207392027, + "tokens_seen": 97452032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901303911735206, + "loss": 2.2574, + "theoretical_loss": 4.817390160519435, + "tokens_seen": 97517568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901203610832497, + "loss": 2.2758, + "theoretical_loss": 4.816942498897831, + "tokens_seen": 97583104 + }, + { + "epoch": 0.03, + "learning_rate": 0.000490110330992979, + "loss": 2.4278, + "theoretical_loss": 4.816495221937563, + "tokens_seen": 97648640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004901003009027082, + "loss": 2.1682, + "theoretical_loss": 4.816048329050273, + "tokens_seen": 97714176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900902708124374, + "loss": 2.2299, + "theoretical_loss": 4.815601819648897, + "tokens_seen": 97779712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900802407221665, + "loss": 2.0681, + "theoretical_loss": 4.815155693147666, + "tokens_seen": 97845248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900702106318957, + "loss": 2.1605, + "theoretical_loss": 4.81470994896209, + "tokens_seen": 97910784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900601805416249, + "loss": 2.1837, + "theoretical_loss": 4.8142645865089655, + "tokens_seen": 97976320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900501504513541, + "loss": 2.1709, + "theoretical_loss": 4.813819605206369, + "tokens_seen": 98041856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900401203610833, + "loss": 2.0897, + "theoretical_loss": 4.813375004473651, + "tokens_seen": 98107392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900300902708124, + "loss": 2.0329, + "theoretical_loss": 4.812930783731433, + "tokens_seen": 98172928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900200601805416, + "loss": 1.9842, + "theoretical_loss": 4.8124869424016055, + "tokens_seen": 98238464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 43181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1590049266815186, + "objective/train/theoretical_loss": 4.812043479907323, + "objective/train/tokens_used": 118764000, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004900100300902708, + "loss": 2.1595, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.00049, + "loss": 2.0954, + "theoretical_loss": 4.8116003956730005, + "tokens_seen": 98369536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899899699097292, + "loss": 2.1232, + "theoretical_loss": 4.811157689124311, + "tokens_seen": 98435072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899799398194583, + "loss": 1.989, + "theoretical_loss": 4.8107153596881815, + "tokens_seen": 98500608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899699097291876, + "loss": 2.016, + "theoretical_loss": 4.810273406792788, + "tokens_seen": 98566144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899598796389167, + "loss": 2.1247, + "theoretical_loss": 4.809831829867555, + "tokens_seen": 98631680 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489949849548646, + "loss": 2.2352, + "theoretical_loss": 4.809390628343146, + "tokens_seen": 98697216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899398194583751, + "loss": 2.1375, + "theoretical_loss": 4.808949801651468, + "tokens_seen": 98762752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899297893681044, + "loss": 2.1334, + "theoretical_loss": 4.8085093492256625, + "tokens_seen": 98828288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899197592778335, + "loss": 2.0008, + "theoretical_loss": 4.808069270500102, + "tokens_seen": 98893824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004899097291875627, + "loss": 2.0579, + "theoretical_loss": 4.80762956491039, + "tokens_seen": 98959360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898996990972919, + "loss": 2.1631, + "theoretical_loss": 4.807190231893355, + "tokens_seen": 99024896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898896690070211, + "loss": 2.2276, + "theoretical_loss": 4.806751270887048, + "tokens_seen": 99090432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898796389167503, + "loss": 2.1152, + "theoretical_loss": 4.806312681330736, + "tokens_seen": 99155968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898696088264794, + "loss": 2.0636, + "theoretical_loss": 4.805874462664903, + "tokens_seen": 99221504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898595787362086, + "loss": 2.1501, + "theoretical_loss": 4.805436614331247, + "tokens_seen": 99287040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898495486459378, + "loss": 2.1961, + "theoretical_loss": 4.804999135772668, + "tokens_seen": 99352576 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489839518555667, + "loss": 2.2069, + "theoretical_loss": 4.804562026433278, + "tokens_seen": 99418112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898294884653962, + "loss": 2.1267, + "theoretical_loss": 4.804125285758383, + "tokens_seen": 99483648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898194583751254, + "loss": 2.148, + "theoretical_loss": 4.803688913194495, + "tokens_seen": 99549184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004898094282848545, + "loss": 2.1239, + "theoretical_loss": 4.8032529081893145, + "tokens_seen": 99614720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897993981945837, + "loss": 2.1014, + "theoretical_loss": 4.802817270191736, + "tokens_seen": 99680256 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489789368104313, + "loss": 2.1279, + "theoretical_loss": 4.802381998651841, + "tokens_seen": 99745792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897793380140421, + "loss": 2.0333, + "theoretical_loss": 4.801947093020896, + "tokens_seen": 99811328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897693079237714, + "loss": 2.2686, + "theoretical_loss": 4.801512552751349, + "tokens_seen": 99876864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897592778335005, + "loss": 2.0119, + "theoretical_loss": 4.801078377296827, + "tokens_seen": 99942400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897492477432297, + "loss": 2.0609, + "theoretical_loss": 4.800644566112128, + "tokens_seen": 100007936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897392176529589, + "loss": 2.2979, + "theoretical_loss": 4.800211118653225, + "tokens_seen": 100073472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897291875626881, + "loss": 2.1458, + "theoretical_loss": 4.799778034377258, + "tokens_seen": 100139008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897191574724173, + "loss": 2.1482, + "theoretical_loss": 4.799345312742531, + "tokens_seen": 100204544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004897091273821465, + "loss": 2.1256, + "theoretical_loss": 4.798912953208511, + "tokens_seen": 100270080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896990972918756, + "loss": 2.1053, + "theoretical_loss": 4.798480955235822, + "tokens_seen": 100335616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896890672016048, + "loss": 2.3129, + "theoretical_loss": 4.798049318286243, + "tokens_seen": 100401152 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489679037111334, + "loss": 2.2135, + "theoretical_loss": 4.797618041822707, + "tokens_seen": 100466688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896690070210632, + "loss": 1.9871, + "theoretical_loss": 4.797187125309293, + "tokens_seen": 100532224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896589769307924, + "loss": 2.1374, + "theoretical_loss": 4.796756568211227, + "tokens_seen": 100597760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896489468405215, + "loss": 2.0421, + "theoretical_loss": 4.796326369994876, + "tokens_seen": 100663296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896389167502507, + "loss": 2.0846, + "theoretical_loss": 4.795896530127747, + "tokens_seen": 100728832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896288866599799, + "loss": 2.2051, + "theoretical_loss": 4.795467048078484, + "tokens_seen": 100794368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896188565697091, + "loss": 2.0185, + "theoretical_loss": 4.79503792331686, + "tokens_seen": 100859904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004896088264794384, + "loss": 2.1444, + "theoretical_loss": 4.794609155313782, + "tokens_seen": 100925440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895987963891674, + "loss": 2.0898, + "theoretical_loss": 4.79418074354128, + "tokens_seen": 100990976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895887662988967, + "loss": 2.2007, + "theoretical_loss": 4.79375268747251, + "tokens_seen": 101056512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895787362086259, + "loss": 2.2705, + "theoretical_loss": 4.793324986581746, + "tokens_seen": 101122048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895687061183551, + "loss": 2.0447, + "theoretical_loss": 4.79289764034438, + "tokens_seen": 101187584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895586760280843, + "loss": 2.0152, + "theoretical_loss": 4.792470648236918, + "tokens_seen": 101253120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895486459378135, + "loss": 2.0745, + "theoretical_loss": 4.792044009736978, + "tokens_seen": 101318656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895386158475426, + "loss": 2.119, + "theoretical_loss": 4.791617724323283, + "tokens_seen": 101384192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895285857572718, + "loss": 2.1461, + "theoretical_loss": 4.791191791475663, + "tokens_seen": 101449728 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489518555667001, + "loss": 2.1156, + "theoretical_loss": 4.79076621067505, + "tokens_seen": 101515264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 44310, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.096851110458374, + "objective/train/theoretical_loss": 4.790340981403474, + "objective/train/tokens_used": 122040800, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004895085255767302, + "loss": 2.0514, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894984954864594, + "loss": 2.1227, + "theoretical_loss": 4.7899161031440585, + "tokens_seen": 101646336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894884653961885, + "loss": 2.0795, + "theoretical_loss": 4.7894915753810245, + "tokens_seen": 101711872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894784353059178, + "loss": 2.009, + "theoretical_loss": 4.789067397599677, + "tokens_seen": 101777408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894684052156469, + "loss": 2.1219, + "theoretical_loss": 4.788643569286411, + "tokens_seen": 101842944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894583751253762, + "loss": 2.0349, + "theoretical_loss": 4.788220089928705, + "tokens_seen": 101908480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894483450351053, + "loss": 2.0492, + "theoretical_loss": 4.787796959015118, + "tokens_seen": 101974016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894383149448346, + "loss": 2.1157, + "theoretical_loss": 4.787374176035287, + "tokens_seen": 102039552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894282848545637, + "loss": 2.2446, + "theoretical_loss": 4.78695174047992, + "tokens_seen": 102105088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894182547642929, + "loss": 2.1509, + "theoretical_loss": 4.786529651840802, + "tokens_seen": 102170624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004894082246740221, + "loss": 2.0141, + "theoretical_loss": 4.786107909610784, + "tokens_seen": 102236160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893981945837513, + "loss": 1.9962, + "theoretical_loss": 4.785686513283783, + "tokens_seen": 102301696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893881644934805, + "loss": 2.0077, + "theoretical_loss": 4.785265462354779, + "tokens_seen": 102367232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893781344032096, + "loss": 1.9921, + "theoretical_loss": 4.7848447563198135, + "tokens_seen": 102432768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893681043129388, + "loss": 2.0961, + "theoretical_loss": 4.784424394675983, + "tokens_seen": 102498304 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489358074222668, + "loss": 2.129, + "theoretical_loss": 4.784004376921439, + "tokens_seen": 102563840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893480441323972, + "loss": 2.052, + "theoretical_loss": 4.783584702555387, + "tokens_seen": 102629376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893380140421264, + "loss": 2.0835, + "theoretical_loss": 4.783165371078075, + "tokens_seen": 102694912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893279839518556, + "loss": 2.0564, + "theoretical_loss": 4.782746381990803, + "tokens_seen": 102760448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893179538615847, + "loss": 2.2044, + "theoretical_loss": 4.78232773479591, + "tokens_seen": 102825984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004893079237713139, + "loss": 2.1646, + "theoretical_loss": 4.7819094289967765, + "tokens_seen": 102891520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892978936810432, + "loss": 2.0919, + "theoretical_loss": 4.781491464097819, + "tokens_seen": 102957056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892878635907723, + "loss": 1.9718, + "theoretical_loss": 4.781073839604488, + "tokens_seen": 103022592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892778335005016, + "loss": 2.0772, + "theoretical_loss": 4.780656555023268, + "tokens_seen": 103088128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892678034102307, + "loss": 2.1245, + "theoretical_loss": 4.780239609861669, + "tokens_seen": 103153664 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892577733199599, + "loss": 2.1008, + "theoretical_loss": 4.77982300362823, + "tokens_seen": 103219200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892477432296891, + "loss": 2.1242, + "theoretical_loss": 4.779406735832509, + "tokens_seen": 103284736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892377131394183, + "loss": 2.1415, + "theoretical_loss": 4.778990805985089, + "tokens_seen": 103350272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892276830491475, + "loss": 2.0164, + "theoretical_loss": 4.778575213597568, + "tokens_seen": 103415808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892176529588767, + "loss": 2.0987, + "theoretical_loss": 4.778159958182557, + "tokens_seen": 103481344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004892076228686058, + "loss": 1.96, + "theoretical_loss": 4.777745039253682, + "tokens_seen": 103546880 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489197592778335, + "loss": 2.0463, + "theoretical_loss": 4.777330456325579, + "tokens_seen": 103612416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891875626880642, + "loss": 2.1013, + "theoretical_loss": 4.776916208913887, + "tokens_seen": 103677952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891775325977934, + "loss": 2.1226, + "theoretical_loss": 4.776502296535252, + "tokens_seen": 103743488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891675025075226, + "loss": 2.0624, + "theoretical_loss": 4.776088718707317, + "tokens_seen": 103809024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891574724172517, + "loss": 2.0187, + "theoretical_loss": 4.77567547494873, + "tokens_seen": 103874560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891474423269809, + "loss": 2.1736, + "theoretical_loss": 4.775262564779128, + "tokens_seen": 103940096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891374122367101, + "loss": 2.2307, + "theoretical_loss": 4.774849987719145, + "tokens_seen": 104005632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891273821464393, + "loss": 2.2168, + "theoretical_loss": 4.774437743290404, + "tokens_seen": 104071168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891173520561686, + "loss": 2.1519, + "theoretical_loss": 4.774025831015517, + "tokens_seen": 104136704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004891073219658976, + "loss": 1.9428, + "theoretical_loss": 4.7736142504180785, + "tokens_seen": 104202240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890972918756269, + "loss": 1.8914, + "theoretical_loss": 4.773203001022668, + "tokens_seen": 104267776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890872617853561, + "loss": 2.1539, + "theoretical_loss": 4.772792082354843, + "tokens_seen": 104333312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890772316950853, + "loss": 2.1066, + "theoretical_loss": 4.77238149394114, + "tokens_seen": 104398848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890672016048145, + "loss": 2.1423, + "theoretical_loss": 4.771971235309065, + "tokens_seen": 104464384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890571715145437, + "loss": 2.0385, + "theoretical_loss": 4.771561305987104, + "tokens_seen": 104529920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890471414242728, + "loss": 2.1267, + "theoretical_loss": 4.771151705504707, + "tokens_seen": 104595456 + }, + { + "epoch": 0.03, + "learning_rate": 0.000489037111334002, + "loss": 2.1033, + "theoretical_loss": 4.770742433392289, + "tokens_seen": 104660992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890270812437312, + "loss": 1.996, + "theoretical_loss": 4.770333489181233, + "tokens_seen": 104726528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890170511534604, + "loss": 1.9585, + "theoretical_loss": 4.769924872403883, + "tokens_seen": 104792064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 45444, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.2567241191864014, + "objective/train/theoretical_loss": 4.769516582593539, + "objective/train/tokens_used": 125317600, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004890070210631896, + "loss": 2.1496, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889969909729187, + "loss": 2.0591, + "theoretical_loss": 4.769108619284461, + "tokens_seen": 104923136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889869608826479, + "loss": 1.9202, + "theoretical_loss": 4.768700982011859, + "tokens_seen": 104988672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889769307923771, + "loss": 2.0797, + "theoretical_loss": 4.768293670311899, + "tokens_seen": 105054208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889669007021063, + "loss": 1.9994, + "theoretical_loss": 4.7678866837216916, + "tokens_seen": 105119744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889568706118355, + "loss": 2.063, + "theoretical_loss": 4.767480021779294, + "tokens_seen": 105185280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889468405215647, + "loss": 2.0418, + "theoretical_loss": 4.76707368402371, + "tokens_seen": 105250816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889368104312939, + "loss": 2.0357, + "theoretical_loss": 4.766667669994881, + "tokens_seen": 105316352 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488926780341023, + "loss": 2.0333, + "theoretical_loss": 4.76626197923369, + "tokens_seen": 105381888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889167502507523, + "loss": 2.0549, + "theoretical_loss": 4.765856611281953, + "tokens_seen": 105447424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004889067201604815, + "loss": 2.0288, + "theoretical_loss": 4.765451565682423, + "tokens_seen": 105512960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888966900702107, + "loss": 2.1411, + "theoretical_loss": 4.765046841978783, + "tokens_seen": 105578496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888866599799398, + "loss": 2.034, + "theoretical_loss": 4.764642439715642, + "tokens_seen": 105644032 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488876629889669, + "loss": 2.2775, + "theoretical_loss": 4.764238358438539, + "tokens_seen": 105709568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888665997993982, + "loss": 2.1421, + "theoretical_loss": 4.763834597693935, + "tokens_seen": 105775104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888565697091274, + "loss": 1.9869, + "theoretical_loss": 4.763431157029213, + "tokens_seen": 105840640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888465396188566, + "loss": 1.9878, + "theoretical_loss": 4.763028035992675, + "tokens_seen": 105906176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888365095285858, + "loss": 2.1662, + "theoretical_loss": 4.762625234133537, + "tokens_seen": 105971712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888264794383149, + "loss": 2.036, + "theoretical_loss": 4.762222751001933, + "tokens_seen": 106037248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888164493480441, + "loss": 1.975, + "theoretical_loss": 4.761820586148906, + "tokens_seen": 106102784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004888064192577733, + "loss": 2.1059, + "theoretical_loss": 4.761418739126407, + "tokens_seen": 106168320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887963891675025, + "loss": 1.9887, + "theoretical_loss": 4.761017209487298, + "tokens_seen": 106233856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887863590772317, + "loss": 2.0692, + "theoretical_loss": 4.760615996785342, + "tokens_seen": 106299392 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488776328986961, + "loss": 2.1139, + "theoretical_loss": 4.760215100575204, + "tokens_seen": 106364928 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048876629889669, + "loss": 2.109, + "theoretical_loss": 4.759814520412451, + "tokens_seen": 106430464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887562688064193, + "loss": 2.0608, + "theoretical_loss": 4.759414255853547, + "tokens_seen": 106496000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887462387161484, + "loss": 2.1165, + "theoretical_loss": 4.759014306455848, + "tokens_seen": 106561536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887362086258777, + "loss": 2.0127, + "theoretical_loss": 4.758614671777604, + "tokens_seen": 106627072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887261785356069, + "loss": 2.0511, + "theoretical_loss": 4.758215351377956, + "tokens_seen": 106692608 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488716148445336, + "loss": 1.9893, + "theoretical_loss": 4.757816344816933, + "tokens_seen": 106758144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004887061183550652, + "loss": 2.0739, + "theoretical_loss": 4.75741765165545, + "tokens_seen": 106823680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886960882647944, + "loss": 2.099, + "theoretical_loss": 4.7570192714553015, + "tokens_seen": 106889216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886860581745236, + "loss": 2.0104, + "theoretical_loss": 4.756621203779169, + "tokens_seen": 106954752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886760280842528, + "loss": 2.1024, + "theoretical_loss": 4.756223448190605, + "tokens_seen": 107020288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886659979939819, + "loss": 2.182, + "theoretical_loss": 4.755826004254046, + "tokens_seen": 107085824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886559679037111, + "loss": 2.0969, + "theoretical_loss": 4.755428871534798, + "tokens_seen": 107151360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886459378134403, + "loss": 2.0316, + "theoretical_loss": 4.755032049599039, + "tokens_seen": 107216896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886359077231695, + "loss": 2.0812, + "theoretical_loss": 4.754635538013817, + "tokens_seen": 107282432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886258776328988, + "loss": 2.0803, + "theoretical_loss": 4.754239336347048, + "tokens_seen": 107347968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886158475426278, + "loss": 2.0529, + "theoretical_loss": 4.753843444167513, + "tokens_seen": 107413504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004886058174523571, + "loss": 2.0632, + "theoretical_loss": 4.753447861044853, + "tokens_seen": 107479040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885957873620863, + "loss": 2.1088, + "theoretical_loss": 4.753052586549572, + "tokens_seen": 107544576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885857572718155, + "loss": 2.0756, + "theoretical_loss": 4.75265762025303, + "tokens_seen": 107610112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885757271815447, + "loss": 2.0652, + "theoretical_loss": 4.752262961727446, + "tokens_seen": 107675648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885656970912739, + "loss": 2.1281, + "theoretical_loss": 4.75186861054589, + "tokens_seen": 107741184 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488555667001003, + "loss": 2.0189, + "theoretical_loss": 4.751474566282285, + "tokens_seen": 107806720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885456369107322, + "loss": 2.1799, + "theoretical_loss": 4.751080828511402, + "tokens_seen": 107872256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885356068204614, + "loss": 2.1386, + "theoretical_loss": 4.750687396808859, + "tokens_seen": 107937792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885255767301906, + "loss": 1.9656, + "theoretical_loss": 4.7502942707511195, + "tokens_seen": 108003328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004885155466399198, + "loss": 2.0848, + "theoretical_loss": 4.749901449915489, + "tokens_seen": 108068864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 46682, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.244960069656372, + "objective/train/theoretical_loss": 4.749508933880117, + "objective/train/tokens_used": 128594400, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488505516549649, + "loss": 2.1502, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884954864593781, + "loss": 2.0584, + "theoretical_loss": 4.7491167222239845, + "tokens_seen": 108199936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884854563691073, + "loss": 2.047, + "theoretical_loss": 4.7487248145269145, + "tokens_seen": 108265472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884754262788365, + "loss": 2.0879, + "theoretical_loss": 4.7483332103695615, + "tokens_seen": 108331008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884653961885657, + "loss": 2.1495, + "theoretical_loss": 4.747941909333411, + "tokens_seen": 108396544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884553660982949, + "loss": 2.0389, + "theoretical_loss": 4.747550911000781, + "tokens_seen": 108462080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884453360080241, + "loss": 2.0023, + "theoretical_loss": 4.747160214954813, + "tokens_seen": 108527616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884353059177532, + "loss": 2.1185, + "theoretical_loss": 4.746769820779477, + "tokens_seen": 108593152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884252758274825, + "loss": 2.0583, + "theoretical_loss": 4.746379728059566, + "tokens_seen": 108658688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884152457372117, + "loss": 2.1069, + "theoretical_loss": 4.745989936380692, + "tokens_seen": 108724224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004884052156469409, + "loss": 2.1241, + "theoretical_loss": 4.745600445329288, + "tokens_seen": 108789760 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048839518555667, + "loss": 2.0836, + "theoretical_loss": 4.745211254492602, + "tokens_seen": 108855296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883851554663992, + "loss": 2.0603, + "theoretical_loss": 4.744822363458699, + "tokens_seen": 108920832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883751253761284, + "loss": 2.0709, + "theoretical_loss": 4.744433771816455, + "tokens_seen": 108986368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883650952858576, + "loss": 2.1458, + "theoretical_loss": 4.744045479155557, + "tokens_seen": 109051904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883550651955868, + "loss": 2.0432, + "theoretical_loss": 4.7436574850665005, + "tokens_seen": 109117440 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488345035105316, + "loss": 1.9805, + "theoretical_loss": 4.743269789140587, + "tokens_seen": 109182976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883350050150451, + "loss": 1.9949, + "theoretical_loss": 4.742882390969925, + "tokens_seen": 109248512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883249749247743, + "loss": 2.0617, + "theoretical_loss": 4.742495290147421, + "tokens_seen": 109314048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883149448345035, + "loss": 1.941, + "theoretical_loss": 4.742108486266783, + "tokens_seen": 109379584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004883049147442327, + "loss": 2.072, + "theoretical_loss": 4.74172197892252, + "tokens_seen": 109445120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882948846539619, + "loss": 2.0183, + "theoretical_loss": 4.741335767709934, + "tokens_seen": 109510656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882848545636911, + "loss": 2.0642, + "theoretical_loss": 4.740949852225122, + "tokens_seen": 109576192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882748244734203, + "loss": 2.0999, + "theoretical_loss": 4.740564232064973, + "tokens_seen": 109641728 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048826479438314946, + "loss": 2.127, + "theoretical_loss": 4.7401789068271665, + "tokens_seen": 109707264 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048825476429287864, + "loss": 2.0709, + "theoretical_loss": 4.73979387611017, + "tokens_seen": 109772800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882447342026078, + "loss": 1.9825, + "theoretical_loss": 4.739409139513235, + "tokens_seen": 109838336 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488234704112337, + "loss": 2.095, + "theoretical_loss": 4.739024696636401, + "tokens_seen": 109903872 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048822467402206624, + "loss": 2.0993, + "theoretical_loss": 4.738640547080485, + "tokens_seen": 109969408 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048821464393179536, + "loss": 1.9991, + "theoretical_loss": 4.738256690447088, + "tokens_seen": 110034944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004882046138415246, + "loss": 2.1215, + "theoretical_loss": 4.7378731263385845, + "tokens_seen": 110100480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881945837512537, + "loss": 2.041, + "theoretical_loss": 4.737489854358129, + "tokens_seen": 110166016 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048818455366098296, + "loss": 1.9741, + "theoretical_loss": 4.737106874109649, + "tokens_seen": 110231552 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048817452357071214, + "loss": 2.0566, + "theoretical_loss": 4.736724185197844, + "tokens_seen": 110297088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881644934804413, + "loss": 2.1576, + "theoretical_loss": 4.7363417872281826, + "tokens_seen": 110362624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881544633901705, + "loss": 2.102, + "theoretical_loss": 4.735959679806903, + "tokens_seen": 110428160 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048814443329989974, + "loss": 2.0778, + "theoretical_loss": 4.735577862541007, + "tokens_seen": 110493696 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048813440320962887, + "loss": 2.0332, + "theoretical_loss": 4.735196335038264, + "tokens_seen": 110559232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004881243731193581, + "loss": 2.1867, + "theoretical_loss": 4.734815096907203, + "tokens_seen": 110624768 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048811434302908723, + "loss": 1.961, + "theoretical_loss": 4.734434147757117, + "tokens_seen": 110690304 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048810431293881646, + "loss": 2.0434, + "theoretical_loss": 4.734053487198054, + "tokens_seen": 110755840 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048809428284854564, + "loss": 2.0175, + "theoretical_loss": 4.733673114840817, + "tokens_seen": 110821376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880842527582748, + "loss": 2.0581, + "theoretical_loss": 4.73329303029697, + "tokens_seen": 110886912 + }, + { + "epoch": 0.03, + "learning_rate": 0.000488074222668004, + "loss": 2.1314, + "theoretical_loss": 4.732913233178825, + "tokens_seen": 110952448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880641925777332, + "loss": 1.966, + "theoretical_loss": 4.732533723099444, + "tokens_seen": 111017984 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048805416248746237, + "loss": 2.0607, + "theoretical_loss": 4.732154499672642, + "tokens_seen": 111083520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004880441323971916, + "loss": 1.9928, + "theoretical_loss": 4.731775562512977, + "tokens_seen": 111149056 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048803410230692073, + "loss": 2.0046, + "theoretical_loss": 4.7313969112357555, + "tokens_seen": 111214592 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048802407221664997, + "loss": 1.9087, + "theoretical_loss": 4.731018545457024, + "tokens_seen": 111280128 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048801404212637915, + "loss": 2.0779, + "theoretical_loss": 4.730640464793574, + "tokens_seen": 111345664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 47788, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0118637084960938, + "objective/train/theoretical_loss": 4.730262668862931, + "objective/train/tokens_used": 131871200, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048800401203610833, + "loss": 2.0503, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879939819458375, + "loss": 2.1219, + "theoretical_loss": 4.729885157283366, + "tokens_seen": 111476736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879839518555667, + "loss": 2.0115, + "theoretical_loss": 4.729507929673879, + "tokens_seen": 111542272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879739217652959, + "loss": 2.0365, + "theoretical_loss": 4.729130985654208, + "tokens_seen": 111607808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879638916750251, + "loss": 2.0448, + "theoretical_loss": 4.72875432484482, + "tokens_seen": 111673344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879538615847543, + "loss": 2.0932, + "theoretical_loss": 4.728377946866915, + "tokens_seen": 111738880 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048794383149448347, + "loss": 2.13, + "theoretical_loss": 4.72800185134242, + "tokens_seen": 111804416 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048793380140421265, + "loss": 2.1258, + "theoretical_loss": 4.727626037893989, + "tokens_seen": 111869952 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048792377131394183, + "loss": 2.0047, + "theoretical_loss": 4.727250506145001, + "tokens_seen": 111935488 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048791374122367107, + "loss": 2.0298, + "theoretical_loss": 4.726875255719557, + "tokens_seen": 112001024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004879037111334002, + "loss": 2.2066, + "theoretical_loss": 4.726500286242482, + "tokens_seen": 112066560 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048789368104312943, + "loss": 2.0767, + "theoretical_loss": 4.7261255973393155, + "tokens_seen": 112132096 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048788365095285856, + "loss": 2.0733, + "theoretical_loss": 4.725751188636319, + "tokens_seen": 112197632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004878736208625878, + "loss": 2.0302, + "theoretical_loss": 4.725377059760468, + "tokens_seen": 112263168 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048786359077231697, + "loss": 2.0246, + "theoretical_loss": 4.725003210339452, + "tokens_seen": 112328704 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048785356068204615, + "loss": 2.1013, + "theoretical_loss": 4.724629640001672, + "tokens_seen": 112394240 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048784353059177533, + "loss": 2.051, + "theoretical_loss": 4.724256348376242, + "tokens_seen": 112459776 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048783350050150457, + "loss": 1.9062, + "theoretical_loss": 4.723883335092982, + "tokens_seen": 112525312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004878234704112337, + "loss": 2.1067, + "theoretical_loss": 4.723510599782419, + "tokens_seen": 112590848 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048781344032096293, + "loss": 2.0956, + "theoretical_loss": 4.723138142075788, + "tokens_seen": 112656384 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048780341023069206, + "loss": 2.0313, + "theoretical_loss": 4.722765961605024, + "tokens_seen": 112721920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877933801404213, + "loss": 2.1148, + "theoretical_loss": 4.722394058002765, + "tokens_seen": 112787456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877833500501505, + "loss": 2.1289, + "theoretical_loss": 4.7220224309023475, + "tokens_seen": 112852992 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048777331995987966, + "loss": 1.9976, + "theoretical_loss": 4.721651079937811, + "tokens_seen": 112918528 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048776328986960884, + "loss": 2.1319, + "theoretical_loss": 4.7212800047438845, + "tokens_seen": 112984064 + }, + { + "epoch": 0.03, + "learning_rate": 0.000487753259779338, + "loss": 2.0292, + "theoretical_loss": 4.720909204955996, + "tokens_seen": 113049600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877432296890672, + "loss": 2.0156, + "theoretical_loss": 4.720538680210268, + "tokens_seen": 113115136 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048773319959879644, + "loss": 2.0573, + "theoretical_loss": 4.720168430143508, + "tokens_seen": 113180672 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048772316950852556, + "loss": 2.0379, + "theoretical_loss": 4.719798454393219, + "tokens_seen": 113246208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877131394182548, + "loss": 2.078, + "theoretical_loss": 4.719428752597588, + "tokens_seen": 113311744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004877031093279839, + "loss": 1.9686, + "theoretical_loss": 4.719059324395493, + "tokens_seen": 113377280 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048769307923771316, + "loss": 2.0286, + "theoretical_loss": 4.71869016942649, + "tokens_seen": 113442816 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048768304914744234, + "loss": 2.0351, + "theoretical_loss": 4.718321287330823, + "tokens_seen": 113508352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876730190571715, + "loss": 2.0825, + "theoretical_loss": 4.7179526777494125, + "tokens_seen": 113573888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876629889669007, + "loss": 2.0425, + "theoretical_loss": 4.717584340323864, + "tokens_seen": 113639424 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048765295887662994, + "loss": 1.9456, + "theoretical_loss": 4.7172162746964545, + "tokens_seen": 113704960 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048764292878635907, + "loss": 2.1003, + "theoretical_loss": 4.716848480510142, + "tokens_seen": 113770496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004876328986960883, + "loss": 2.0178, + "theoretical_loss": 4.716480957408556, + "tokens_seen": 113836032 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048762286860581743, + "loss": 1.9103, + "theoretical_loss": 4.716113705035999, + "tokens_seen": 113901568 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048761283851554666, + "loss": 1.9907, + "theoretical_loss": 4.715746723037448, + "tokens_seen": 113967104 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048760280842527584, + "loss": 2.0775, + "theoretical_loss": 4.715380011058542, + "tokens_seen": 114032640 + }, + { + "epoch": 0.03, + "learning_rate": 0.000487592778335005, + "loss": 1.9564, + "theoretical_loss": 4.715013568745597, + "tokens_seen": 114098176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875827482447342, + "loss": 2.0277, + "theoretical_loss": 4.714647395745587, + "tokens_seen": 114163712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875727181544634, + "loss": 1.989, + "theoretical_loss": 4.714281491706154, + "tokens_seen": 114229248 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048756268806419257, + "loss": 2.0658, + "theoretical_loss": 4.713915856275603, + "tokens_seen": 114294784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875526579739218, + "loss": 1.9711, + "theoretical_loss": 4.713550489102901, + "tokens_seen": 114360320 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048754262788365093, + "loss": 1.9212, + "theoretical_loss": 4.7131853898376725, + "tokens_seen": 114425856 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048753259779338017, + "loss": 2.0382, + "theoretical_loss": 4.7128205581302005, + "tokens_seen": 114491392 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048752256770310935, + "loss": 2.0236, + "theoretical_loss": 4.712455993631425, + "tokens_seen": 114556928 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048751253761283853, + "loss": 2.0987, + "theoretical_loss": 4.712091695992941, + "tokens_seen": 114622464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 48994, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.027988910675049, + "objective/train/theoretical_loss": 4.711727664866995, + "objective/train/tokens_used": 135148000, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004875025075225677, + "loss": 1.9839, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874924774322969, + "loss": 2.0128, + "theoretical_loss": 4.7113638999064875, + "tokens_seen": 114753536 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048748244734202607, + "loss": 2.0693, + "theoretical_loss": 4.711000400764967, + "tokens_seen": 114819072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874724172517553, + "loss": 1.9637, + "theoretical_loss": 4.71063716709663, + "tokens_seen": 114884608 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048746238716148443, + "loss": 2.1671, + "theoretical_loss": 4.7102741985563235, + "tokens_seen": 114950144 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048745235707121367, + "loss": 2.0567, + "theoretical_loss": 4.709911494799533, + "tokens_seen": 115015680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874423269809428, + "loss": 2.1491, + "theoretical_loss": 4.709549055482395, + "tokens_seen": 115081216 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048743229689067203, + "loss": 2.0667, + "theoretical_loss": 4.709186880261683, + "tokens_seen": 115146752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874222668004012, + "loss": 2.0219, + "theoretical_loss": 4.708824968794813, + "tokens_seen": 115212288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874122367101304, + "loss": 1.8244, + "theoretical_loss": 4.70846332073984, + "tokens_seen": 115277824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0004874022066198596, + "loss": 2.0155, + "theoretical_loss": 4.708101935755455, + "tokens_seen": 115343360 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048739217652958876, + "loss": 2.0304, + "theoretical_loss": 4.707740813500988, + "tokens_seen": 115408896 + }, + { + "epoch": 0.03, + "learning_rate": 0.00048738214643931794, + "loss": 2.1362, + "theoretical_loss": 4.707379953636399, + "tokens_seen": 115474432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048737211634904717, + "loss": 1.9377, + "theoretical_loss": 4.7070193558222835, + "tokens_seen": 115539968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873620862587763, + "loss": 2.0657, + "theoretical_loss": 4.7066590197198686, + "tokens_seen": 115605504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048735205616850553, + "loss": 2.0004, + "theoretical_loss": 4.706298944991008, + "tokens_seen": 115671040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873420260782347, + "loss": 1.9986, + "theoretical_loss": 4.705939131298189, + "tokens_seen": 115736576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873319959879639, + "loss": 2.1924, + "theoretical_loss": 4.705579578304519, + "tokens_seen": 115802112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004873219658976931, + "loss": 2.0684, + "theoretical_loss": 4.705220285673736, + "tokens_seen": 115867648 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048731193580742226, + "loss": 2.1415, + "theoretical_loss": 4.704861253070198, + "tokens_seen": 115933184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048730190571715144, + "loss": 2.0849, + "theoretical_loss": 4.704502480158887, + "tokens_seen": 115998720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872918756268807, + "loss": 2.162, + "theoretical_loss": 4.704143966605406, + "tokens_seen": 116064256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872818455366098, + "loss": 1.967, + "theoretical_loss": 4.703785712075973, + "tokens_seen": 116129792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048727181544633904, + "loss": 2.2589, + "theoretical_loss": 4.703427716237432, + "tokens_seen": 116195328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048726178535606816, + "loss": 2.0707, + "theoretical_loss": 4.703069978757233, + "tokens_seen": 116260864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872517552657974, + "loss": 2.0031, + "theoretical_loss": 4.702712499303448, + "tokens_seen": 116326400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872417251755266, + "loss": 2.1392, + "theoretical_loss": 4.7023552775447595, + "tokens_seen": 116391936 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048723169508525576, + "loss": 2.151, + "theoretical_loss": 4.701998313150462, + "tokens_seen": 116457472 + }, + { + "epoch": 0.04, + "learning_rate": 0.000487221664994985, + "loss": 1.9996, + "theoretical_loss": 4.701641605790459, + "tokens_seen": 116523008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004872116349047141, + "loss": 2.0861, + "theoretical_loss": 4.701285155135265, + "tokens_seen": 116588544 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048720160481444336, + "loss": 1.9912, + "theoretical_loss": 4.700928960856, + "tokens_seen": 116654080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048719157472417254, + "loss": 2.034, + "theoretical_loss": 4.70057302262439, + "tokens_seen": 116719616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871815446339017, + "loss": 1.9401, + "theoretical_loss": 4.700217340112767, + "tokens_seen": 116785152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871715145436309, + "loss": 2.0347, + "theoretical_loss": 4.699861912994063, + "tokens_seen": 116850688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048716148445336014, + "loss": 2.0246, + "theoretical_loss": 4.699506740941813, + "tokens_seen": 116916224 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048715145436308927, + "loss": 2.083, + "theoretical_loss": 4.699151823630155, + "tokens_seen": 116981760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871414242728185, + "loss": 1.8992, + "theoretical_loss": 4.698797160733818, + "tokens_seen": 117047296 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048713139418254763, + "loss": 2.0065, + "theoretical_loss": 4.6984427519281375, + "tokens_seen": 117112832 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048712136409227686, + "loss": 2.2515, + "theoretical_loss": 4.698088596889037, + "tokens_seen": 117178368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048711133400200604, + "loss": 1.978, + "theoretical_loss": 4.697734695293039, + "tokens_seen": 117243904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004871013039117352, + "loss": 2.1086, + "theoretical_loss": 4.697381046817256, + "tokens_seen": 117309440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870912738214644, + "loss": 1.9559, + "theoretical_loss": 4.697027651139393, + "tokens_seen": 117374976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870812437311936, + "loss": 2.0527, + "theoretical_loss": 4.696674507937746, + "tokens_seen": 117440512 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048707121364092277, + "loss": 1.8997, + "theoretical_loss": 4.696321616891199, + "tokens_seen": 117506048 + }, + { + "epoch": 0.04, + "learning_rate": 0.000487061183550652, + "loss": 2.0869, + "theoretical_loss": 4.695968977679221, + "tokens_seen": 117571584 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048705115346038113, + "loss": 2.0808, + "theoretical_loss": 4.695616589981871, + "tokens_seen": 117637120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048704112337011037, + "loss": 1.9676, + "theoretical_loss": 4.695264453479789, + "tokens_seen": 117702656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048703109327983955, + "loss": 1.9466, + "theoretical_loss": 4.6949125678542, + "tokens_seen": 117768192 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048702106318956873, + "loss": 2.1435, + "theoretical_loss": 4.69456093278691, + "tokens_seen": 117833728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870110330992979, + "loss": 2.0159, + "theoretical_loss": 4.694209547960304, + "tokens_seen": 117899264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 50256, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.115172863006592, + "objective/train/theoretical_loss": 4.693858413057347, + "objective/train/tokens_used": 138424800, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004870010030090271, + "loss": 2.089, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048699097291875627, + "loss": 2.1059, + "theoretical_loss": 4.693507527761584, + "tokens_seen": 118030336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869809428284855, + "loss": 2.0822, + "theoretical_loss": 4.6931568917571305, + "tokens_seen": 118095872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048697091273821463, + "loss": 1.9791, + "theoretical_loss": 4.692806504728683, + "tokens_seen": 118161408 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048696088264794387, + "loss": 2.1196, + "theoretical_loss": 4.692456366361506, + "tokens_seen": 118226944 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486950852557673, + "loss": 2.126, + "theoretical_loss": 4.692106476341438, + "tokens_seen": 118292480 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048694082246740223, + "loss": 1.9137, + "theoretical_loss": 4.691756834354891, + "tokens_seen": 118358016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869307923771314, + "loss": 2.0523, + "theoretical_loss": 4.691407440088843, + "tokens_seen": 118423552 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869207622868606, + "loss": 1.9622, + "theoretical_loss": 4.69105829323084, + "tokens_seen": 118489088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004869107321965898, + "loss": 2.086, + "theoretical_loss": 4.690709393468995, + "tokens_seen": 118554624 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048690070210631896, + "loss": 2.0097, + "theoretical_loss": 4.690360740491989, + "tokens_seen": 118620160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048689067201604814, + "loss": 1.9499, + "theoretical_loss": 4.690012333989063, + "tokens_seen": 118685696 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048688064192577737, + "loss": 2.0381, + "theoretical_loss": 4.689664173650023, + "tokens_seen": 118751232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868706118355065, + "loss": 2.0104, + "theoretical_loss": 4.689316259165235, + "tokens_seen": 118816768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048686058174523573, + "loss": 1.9984, + "theoretical_loss": 4.6889685902256275, + "tokens_seen": 118882304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868505516549649, + "loss": 1.9784, + "theoretical_loss": 4.688621166522684, + "tokens_seen": 118947840 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868405215646941, + "loss": 2.0523, + "theoretical_loss": 4.688273987748448, + "tokens_seen": 119013376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868304914744233, + "loss": 2.0843, + "theoretical_loss": 4.687927053595518, + "tokens_seen": 119078912 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048682046138415246, + "loss": 2.1189, + "theoretical_loss": 4.687580363757048, + "tokens_seen": 119144448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048681043129388164, + "loss": 1.987, + "theoretical_loss": 4.6872339179267435, + "tokens_seen": 119209984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004868004012036109, + "loss": 2.0367, + "theoretical_loss": 4.686887715798864, + "tokens_seen": 119275520 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048679037111334, + "loss": 1.9118, + "theoretical_loss": 4.6865417570682215, + "tokens_seen": 119341056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048678034102306924, + "loss": 1.9286, + "theoretical_loss": 4.686196041430172, + "tokens_seen": 119406592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048677031093279836, + "loss": 2.0501, + "theoretical_loss": 4.685850568580624, + "tokens_seen": 119472128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867602808425276, + "loss": 1.9388, + "theoretical_loss": 4.685505338216032, + "tokens_seen": 119537664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867502507522568, + "loss": 2.101, + "theoretical_loss": 4.685160350033398, + "tokens_seen": 119603200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048674022066198596, + "loss": 1.9148, + "theoretical_loss": 4.684815603730263, + "tokens_seen": 119668736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048673019057171514, + "loss": 1.9599, + "theoretical_loss": 4.684471099004718, + "tokens_seen": 119734272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867201604814443, + "loss": 1.9956, + "theoretical_loss": 4.684126835555389, + "tokens_seen": 119799808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004867101303911735, + "loss": 2.2209, + "theoretical_loss": 4.683782813081448, + "tokens_seen": 119865344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048670010030090274, + "loss": 2.0373, + "theoretical_loss": 4.683439031282606, + "tokens_seen": 119930880 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048669007021063187, + "loss": 2.0668, + "theoretical_loss": 4.683095489859107, + "tokens_seen": 119996416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004866800401203611, + "loss": 1.886, + "theoretical_loss": 4.682752188511737, + "tokens_seen": 120061952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004866700100300903, + "loss": 1.9773, + "theoretical_loss": 4.682409126941815, + "tokens_seen": 120127488 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048665997993981947, + "loss": 2.0758, + "theoretical_loss": 4.682066304851195, + "tokens_seen": 120193024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048664994984954865, + "loss": 2.0136, + "theoretical_loss": 4.681723721942264, + "tokens_seen": 120258560 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048663991975927783, + "loss": 2.1484, + "theoretical_loss": 4.6813813779179405, + "tokens_seen": 120324096 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486629889669007, + "loss": 2.0139, + "theoretical_loss": 4.681039272481674, + "tokens_seen": 120389632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048661985957873624, + "loss": 2.0905, + "theoretical_loss": 4.680697405337443, + "tokens_seen": 120455168 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048660982948846537, + "loss": 2.1135, + "theoretical_loss": 4.680355776189755, + "tokens_seen": 120520704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865997993981946, + "loss": 2.0328, + "theoretical_loss": 4.680014384743641, + "tokens_seen": 120586240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048658976930792373, + "loss": 1.8083, + "theoretical_loss": 4.679673230704664, + "tokens_seen": 120651776 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048657973921765297, + "loss": 2.0607, + "theoretical_loss": 4.679332313778904, + "tokens_seen": 120717312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048656970912738215, + "loss": 2.0935, + "theoretical_loss": 4.678991633672968, + "tokens_seen": 120782848 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048655967903711133, + "loss": 1.9801, + "theoretical_loss": 4.678651190093988, + "tokens_seen": 120848384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865496489468405, + "loss": 1.9543, + "theoretical_loss": 4.67831098274961, + "tokens_seen": 120913920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048653961885656975, + "loss": 2.168, + "theoretical_loss": 4.677971011348005, + "tokens_seen": 120979456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865295887662989, + "loss": 2.1068, + "theoretical_loss": 4.677631275597859, + "tokens_seen": 121044992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004865195586760281, + "loss": 1.9495, + "theoretical_loss": 4.677291775208378, + "tokens_seen": 121110528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048650952858575724, + "loss": 2.0618, + "theoretical_loss": 4.676952509889281, + "tokens_seen": 121176064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 51395, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0603079795837402, + "objective/train/theoretical_loss": 4.676613479350804, + "objective/train/tokens_used": 141701600, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048649949849548647, + "loss": 2.0547, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048648946840521565, + "loss": 1.9461, + "theoretical_loss": 4.6762746833036966, + "tokens_seen": 121307136 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048647943831494483, + "loss": 1.9894, + "theoretical_loss": 4.675936121459218, + "tokens_seen": 121372672 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048646940822467407, + "loss": 1.9208, + "theoretical_loss": 4.675597793529143, + "tokens_seen": 121438208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864593781344032, + "loss": 1.9807, + "theoretical_loss": 4.675259699225752, + "tokens_seen": 121503744 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048644934804413243, + "loss": 1.9913, + "theoretical_loss": 4.674921838261836, + "tokens_seen": 121569280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864393179538616, + "loss": 2.0445, + "theoretical_loss": 4.674584210350695, + "tokens_seen": 121634816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004864292878635908, + "loss": 1.995, + "theoretical_loss": 4.674246815206136, + "tokens_seen": 121700352 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048641925777332, + "loss": 2.07, + "theoretical_loss": 4.673909652542466, + "tokens_seen": 121765888 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048640922768304916, + "loss": 1.9318, + "theoretical_loss": 4.673572722074502, + "tokens_seen": 121831424 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048639919759277834, + "loss": 2.0544, + "theoretical_loss": 4.673236023517562, + "tokens_seen": 121896960 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048638916750250757, + "loss": 1.9322, + "theoretical_loss": 4.672899556587465, + "tokens_seen": 121962496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863791374122367, + "loss": 1.9983, + "theoretical_loss": 4.672563321000533, + "tokens_seen": 122028032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048636910732196593, + "loss": 1.9582, + "theoretical_loss": 4.672227316473584, + "tokens_seen": 122093568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863590772316951, + "loss": 2.004, + "theoretical_loss": 4.671891542723937, + "tokens_seen": 122159104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863490471414243, + "loss": 1.9713, + "theoretical_loss": 4.671555999469408, + "tokens_seen": 122224640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863390170511535, + "loss": 2.1866, + "theoretical_loss": 4.671220686428308, + "tokens_seen": 122290176 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048632898696088266, + "loss": 2.0042, + "theoretical_loss": 4.670885603319445, + "tokens_seen": 122355712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048631895687061184, + "loss": 1.9173, + "theoretical_loss": 4.670550749862119, + "tokens_seen": 122421248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004863089267803411, + "loss": 1.9072, + "theoretical_loss": 4.670216125776122, + "tokens_seen": 122486784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862988966900702, + "loss": 2.0523, + "theoretical_loss": 4.669881730781741, + "tokens_seen": 122552320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048628886659979944, + "loss": 1.9463, + "theoretical_loss": 4.66954756459975, + "tokens_seen": 122617856 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048627883650952857, + "loss": 1.9474, + "theoretical_loss": 4.669213626951413, + "tokens_seen": 122683392 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862688064192578, + "loss": 1.9968, + "theoretical_loss": 4.668879917558484, + "tokens_seen": 122748928 + }, + { + "epoch": 0.04, + "learning_rate": 0.000486258776328987, + "loss": 1.994, + "theoretical_loss": 4.6685464361432025, + "tokens_seen": 122814464 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048624874623871616, + "loss": 2.0271, + "theoretical_loss": 4.668213182428295, + "tokens_seen": 122880000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048623871614844534, + "loss": 2.1096, + "theoretical_loss": 4.667880156136971, + "tokens_seen": 122945536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862286860581745, + "loss": 2.0099, + "theoretical_loss": 4.667547356992927, + "tokens_seen": 123011072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004862186559679037, + "loss": 2.0032, + "theoretical_loss": 4.667214784720338, + "tokens_seen": 123076608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048620862587763294, + "loss": 1.9349, + "theoretical_loss": 4.666882439043865, + "tokens_seen": 123142144 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048619859578736207, + "loss": 2.0813, + "theoretical_loss": 4.666550319688646, + "tokens_seen": 123207680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861885656970913, + "loss": 2.0377, + "theoretical_loss": 4.6662184263803, + "tokens_seen": 123273216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861785356068205, + "loss": 1.9875, + "theoretical_loss": 4.665886758844923, + "tokens_seen": 123338752 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048616850551654967, + "loss": 2.0077, + "theoretical_loss": 4.66555531680909, + "tokens_seen": 123404288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048615847542627885, + "loss": 2.0744, + "theoretical_loss": 4.665224099999852, + "tokens_seen": 123469824 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048614844533600803, + "loss": 1.8992, + "theoretical_loss": 4.664893108144735, + "tokens_seen": 123535360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861384152457372, + "loss": 2.0361, + "theoretical_loss": 4.664562340971735, + "tokens_seen": 123600896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048612838515546644, + "loss": 1.9678, + "theoretical_loss": 4.664231798209327, + "tokens_seen": 123666432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048611835506519557, + "loss": 2.1891, + "theoretical_loss": 4.663901479586454, + "tokens_seen": 123731968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004861083249749248, + "loss": 1.957, + "theoretical_loss": 4.663571384832531, + "tokens_seen": 123797504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048609829488465393, + "loss": 1.9761, + "theoretical_loss": 4.663241513677443, + "tokens_seen": 123863040 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048608826479438317, + "loss": 2.0235, + "theoretical_loss": 4.662911865851541, + "tokens_seen": 123928576 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048607823470411235, + "loss": 1.934, + "theoretical_loss": 4.662582441085648, + "tokens_seen": 123994112 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048606820461384153, + "loss": 1.8964, + "theoretical_loss": 4.662253239111049, + "tokens_seen": 124059648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860581745235707, + "loss": 2.0257, + "theoretical_loss": 4.661924259659497, + "tokens_seen": 124125184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048604814443329995, + "loss": 2.017, + "theoretical_loss": 4.661595502463211, + "tokens_seen": 124190720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860381143430291, + "loss": 1.9626, + "theoretical_loss": 4.661266967254868, + "tokens_seen": 124256256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004860280842527583, + "loss": 1.9144, + "theoretical_loss": 4.660938653767614, + "tokens_seen": 124321792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048601805416248744, + "loss": 2.0844, + "theoretical_loss": 4.660610561735049, + "tokens_seen": 124387328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048600802407221667, + "loss": 2.0015, + "theoretical_loss": 4.66028269089124, + "tokens_seen": 124452864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 52572, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9419217109680176, + "objective/train/theoretical_loss": 4.659955040970709, + "objective/train/tokens_used": 144978400, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048599799398194585, + "loss": 1.9425, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048598796389167503, + "loss": 1.9502, + "theoretical_loss": 4.659627611708437, + "tokens_seen": 124583936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859779338014042, + "loss": 2.1001, + "theoretical_loss": 4.659300402839863, + "tokens_seen": 124649472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859679037111334, + "loss": 2.0167, + "theoretical_loss": 4.658973414100881, + "tokens_seen": 124715008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859578736208626, + "loss": 2.1285, + "theoretical_loss": 4.658646645227842, + "tokens_seen": 124780544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859478435305918, + "loss": 1.9959, + "theoretical_loss": 4.6583200959575475, + "tokens_seen": 124846080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048593781344032094, + "loss": 1.9928, + "theoretical_loss": 4.657993766027255, + "tokens_seen": 124911616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859277833500502, + "loss": 2.0752, + "theoretical_loss": 4.657667655174674, + "tokens_seen": 124977152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004859177532597793, + "loss": 2.0844, + "theoretical_loss": 4.657341763137963, + "tokens_seen": 125042688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048590772316950854, + "loss": 2.0637, + "theoretical_loss": 4.657016089655732, + "tokens_seen": 125108224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858976930792377, + "loss": 2.0728, + "theoretical_loss": 4.65669063446704, + "tokens_seen": 125173760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858876629889669, + "loss": 2.0352, + "theoretical_loss": 4.656365397311393, + "tokens_seen": 125239296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858776328986961, + "loss": 2.0233, + "theoretical_loss": 4.656040377928743, + "tokens_seen": 125304832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858676028084253, + "loss": 1.8639, + "theoretical_loss": 4.655715576059493, + "tokens_seen": 125370368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048585757271815444, + "loss": 2.0151, + "theoretical_loss": 4.655390991444484, + "tokens_seen": 125435904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858475426278837, + "loss": 2.0151, + "theoretical_loss": 4.6550666238250065, + "tokens_seen": 125501440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858375125376128, + "loss": 1.9687, + "theoretical_loss": 4.65474247294279, + "tokens_seen": 125566976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048582748244734204, + "loss": 1.9697, + "theoretical_loss": 4.654418538540009, + "tokens_seen": 125632512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858174523570712, + "loss": 2.0563, + "theoretical_loss": 4.654094820359278, + "tokens_seen": 125698048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004858074222668004, + "loss": 2.0584, + "theoretical_loss": 4.653771318143651, + "tokens_seen": 125763584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857973921765296, + "loss": 1.9804, + "theoretical_loss": 4.65344803163662, + "tokens_seen": 125829120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048578736208625877, + "loss": 1.9934, + "theoretical_loss": 4.653124960582118, + "tokens_seen": 125894656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048577733199598795, + "loss": 2.051, + "theoretical_loss": 4.652802104724513, + "tokens_seen": 125960192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857673019057172, + "loss": 2.0099, + "theoretical_loss": 4.652479463808609, + "tokens_seen": 126025728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857572718154463, + "loss": 1.9361, + "theoretical_loss": 4.652157037579647, + "tokens_seen": 126091264 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048574724172517554, + "loss": 1.9769, + "theoretical_loss": 4.6518348257833, + "tokens_seen": 126156800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048573721163490467, + "loss": 1.895, + "theoretical_loss": 4.651512828165675, + "tokens_seen": 126222336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004857271815446339, + "loss": 2.0541, + "theoretical_loss": 4.651191044473311, + "tokens_seen": 126287872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048571715145436314, + "loss": 1.9447, + "theoretical_loss": 4.65086947445318, + "tokens_seen": 126353408 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048570712136409227, + "loss": 1.7801, + "theoretical_loss": 4.650548117852681, + "tokens_seen": 126418944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856970912738215, + "loss": 2.0252, + "theoretical_loss": 4.650226974419646, + "tokens_seen": 126484480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856870611835507, + "loss": 2.0488, + "theoretical_loss": 4.649906043902332, + "tokens_seen": 126550016 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048567703109327987, + "loss": 2.0856, + "theoretical_loss": 4.6495853260494275, + "tokens_seen": 126615552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048566700100300905, + "loss": 1.9979, + "theoretical_loss": 4.649264820610043, + "tokens_seen": 126681088 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048565697091273823, + "loss": 1.9332, + "theoretical_loss": 4.648944527333717, + "tokens_seen": 126746624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004856469408224674, + "loss": 1.9909, + "theoretical_loss": 4.648624445970414, + "tokens_seen": 126812160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048563691073219664, + "loss": 1.9091, + "theoretical_loss": 4.648304576270519, + "tokens_seen": 126877696 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048562688064192577, + "loss": 2.0723, + "theoretical_loss": 4.647984917984841, + "tokens_seen": 126943232 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485616850551655, + "loss": 2.0723, + "theoretical_loss": 4.647665470864613, + "tokens_seen": 127008768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048560682046138413, + "loss": 1.9708, + "theoretical_loss": 4.6473462346614856, + "tokens_seen": 127074304 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048559679037111337, + "loss": 2.0581, + "theoretical_loss": 4.647027209127531, + "tokens_seen": 127139840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048558676028084255, + "loss": 1.9734, + "theoretical_loss": 4.646708394015242, + "tokens_seen": 127205376 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048557673019057173, + "loss": 1.9716, + "theoretical_loss": 4.646389789077524, + "tokens_seen": 127270912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855667001003009, + "loss": 2.0263, + "theoretical_loss": 4.646071394067707, + "tokens_seen": 127336448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048555667001003015, + "loss": 1.9559, + "theoretical_loss": 4.645753208739533, + "tokens_seen": 127401984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855466399197593, + "loss": 1.9377, + "theoretical_loss": 4.645435232847159, + "tokens_seen": 127467520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004855366098294885, + "loss": 1.9411, + "theoretical_loss": 4.645117466145157, + "tokens_seen": 127533056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048552657973921764, + "loss": 1.8651, + "theoretical_loss": 4.6447999083885145, + "tokens_seen": 127598592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048551654964894687, + "loss": 2.0505, + "theoretical_loss": 4.644482559332629, + "tokens_seen": 127664128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048550651955867605, + "loss": 2.0377, + "theoretical_loss": 4.644165418733312, + "tokens_seen": 127729664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 53741, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.041177749633789, + "objective/train/theoretical_loss": 4.6438484863467835, + "objective/train/tokens_used": 148255200, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048549648946840523, + "loss": 2.0956, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854864593781344, + "loss": 2.1179, + "theoretical_loss": 4.643531761929674, + "tokens_seen": 127860736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854764292878636, + "loss": 1.9199, + "theoretical_loss": 4.643215245239025, + "tokens_seen": 127926272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854663991975928, + "loss": 1.9031, + "theoretical_loss": 4.642898936032285, + "tokens_seen": 127991808 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485456369107322, + "loss": 1.9236, + "theoretical_loss": 4.6425828340673085, + "tokens_seen": 128057344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048544633901705114, + "loss": 2.0659, + "theoretical_loss": 4.642266939102357, + "tokens_seen": 128122880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854363089267804, + "loss": 1.9441, + "theoretical_loss": 4.641951250896098, + "tokens_seen": 128188416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854262788365095, + "loss": 1.9857, + "theoretical_loss": 4.641635769207603, + "tokens_seen": 128253952 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048541624874623874, + "loss": 2.0599, + "theoretical_loss": 4.641320493796349, + "tokens_seen": 128319488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004854062186559679, + "loss": 1.9108, + "theoretical_loss": 4.641005424422213, + "tokens_seen": 128385024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853961885656971, + "loss": 1.9658, + "theoretical_loss": 4.640690560845474, + "tokens_seen": 128450560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853861584754263, + "loss": 2.0277, + "theoretical_loss": 4.640375902826815, + "tokens_seen": 128516096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853761283851555, + "loss": 2.1388, + "theoretical_loss": 4.640061450127316, + "tokens_seen": 128581632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048536609829488464, + "loss": 1.9091, + "theoretical_loss": 4.639747202508459, + "tokens_seen": 128647168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853560682046139, + "loss": 2.0019, + "theoretical_loss": 4.639433159732121, + "tokens_seen": 128712704 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485346038114343, + "loss": 2.0716, + "theoretical_loss": 4.639119321560581, + "tokens_seen": 128778240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048533600802407224, + "loss": 2.0006, + "theoretical_loss": 4.638805687756511, + "tokens_seen": 128843776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853259779338014, + "loss": 1.9323, + "theoretical_loss": 4.63849225808298, + "tokens_seen": 128909312 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853159478435306, + "loss": 2.0347, + "theoretical_loss": 4.6381790323034515, + "tokens_seen": 128974848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004853059177532598, + "loss": 2.1317, + "theoretical_loss": 4.637866010181784, + "tokens_seen": 129040384 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048529588766298897, + "loss": 2.007, + "theoretical_loss": 4.63755319148223, + "tokens_seen": 129105920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048528585757271815, + "loss": 2.0142, + "theoretical_loss": 4.637240575969432, + "tokens_seen": 129171456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852758274824474, + "loss": 1.938, + "theoretical_loss": 4.636928163408426, + "tokens_seen": 129236992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852657973921765, + "loss": 2.0538, + "theoretical_loss": 4.636615953564637, + "tokens_seen": 129302528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048525576730190574, + "loss": 1.8404, + "theoretical_loss": 4.636303946203881, + "tokens_seen": 129368064 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048524573721163487, + "loss": 2.03, + "theoretical_loss": 4.635992141092364, + "tokens_seen": 129433600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852357071213641, + "loss": 1.9173, + "theoretical_loss": 4.635680537996679, + "tokens_seen": 129499136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004852256770310933, + "loss": 2.0093, + "theoretical_loss": 4.635369136683805, + "tokens_seen": 129564672 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048521564694082247, + "loss": 1.9382, + "theoretical_loss": 4.6350579369211085, + "tokens_seen": 129630208 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048520561685055165, + "loss": 2.0094, + "theoretical_loss": 4.634746938476344, + "tokens_seen": 129695744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851955867602809, + "loss": 1.9031, + "theoretical_loss": 4.634436141117645, + "tokens_seen": 129761280 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048518555667001, + "loss": 2.0656, + "theoretical_loss": 4.634125544613536, + "tokens_seen": 129826816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048517552657973925, + "loss": 1.9652, + "theoretical_loss": 4.63381514873292, + "tokens_seen": 129892352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851654964894684, + "loss": 2.1337, + "theoretical_loss": 4.633504953245085, + "tokens_seen": 129957888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851554663991976, + "loss": 2.0488, + "theoretical_loss": 4.633194957919697, + "tokens_seen": 130023424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851454363089268, + "loss": 2.0926, + "theoretical_loss": 4.632885162526806, + "tokens_seen": 130088960 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048513540621865597, + "loss": 1.9962, + "theoretical_loss": 4.63257556683684, + "tokens_seen": 130154496 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048512537612838515, + "loss": 2.0067, + "theoretical_loss": 4.632266170620607, + "tokens_seen": 130220032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048511534603811433, + "loss": 2.0433, + "theoretical_loss": 4.631956973649293, + "tokens_seen": 130285568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004851053159478435, + "loss": 2.1109, + "theoretical_loss": 4.631647975694461, + "tokens_seen": 130351104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048509528585757275, + "loss": 2.0797, + "theoretical_loss": 4.631339176528051, + "tokens_seen": 130416640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850852557673019, + "loss": 1.9578, + "theoretical_loss": 4.631030575922379, + "tokens_seen": 130482176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850752256770311, + "loss": 1.8957, + "theoretical_loss": 4.630722173650135, + "tokens_seen": 130547712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048506519558676024, + "loss": 1.9079, + "theoretical_loss": 4.630413969484383, + "tokens_seen": 130613248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850551654964895, + "loss": 1.8259, + "theoretical_loss": 4.630105963198563, + "tokens_seen": 130678784 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048504513540621866, + "loss": 1.9699, + "theoretical_loss": 4.629798154566485, + "tokens_seen": 130744320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048503510531594784, + "loss": 1.9397, + "theoretical_loss": 4.62949054336233, + "tokens_seen": 130809856 + }, + { + "epoch": 0.04, + "learning_rate": 0.000485025075225677, + "loss": 2.0771, + "theoretical_loss": 4.629183129360653, + "tokens_seen": 130875392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048501504513540625, + "loss": 1.9704, + "theoretical_loss": 4.628875912336378, + "tokens_seen": 130940928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004850050150451354, + "loss": 1.9575, + "theoretical_loss": 4.628568892064797, + "tokens_seen": 131006464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 54897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0116264820098877, + "objective/train/theoretical_loss": 4.62826206832157, + "objective/train/tokens_used": 151532000, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849949849548646, + "loss": 1.925, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849849548645938, + "loss": 2.1905, + "theoretical_loss": 4.6279554408827295, + "tokens_seen": 131137536 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484974924774323, + "loss": 1.9063, + "theoretical_loss": 4.62764900952467, + "tokens_seen": 131203072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849648946840522, + "loss": 1.8823, + "theoretical_loss": 4.627342774024155, + "tokens_seen": 131268608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048495486459378134, + "loss": 2.0375, + "theoretical_loss": 4.6270367341583105, + "tokens_seen": 131334144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849448345035106, + "loss": 1.9291, + "theoretical_loss": 4.626730889704632, + "tokens_seen": 131399680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849348044132397, + "loss": 1.8603, + "theoretical_loss": 4.626425240440973, + "tokens_seen": 131465216 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048492477432296894, + "loss": 2.0999, + "theoretical_loss": 4.626119786145557, + "tokens_seen": 131530752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849147442326981, + "loss": 1.9413, + "theoretical_loss": 4.625814526596962, + "tokens_seen": 131596288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004849047141424273, + "loss": 1.9511, + "theoretical_loss": 4.625509461574135, + "tokens_seen": 131661824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848946840521565, + "loss": 2.1461, + "theoretical_loss": 4.625204590856377, + "tokens_seen": 131727360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848846539618857, + "loss": 2.0044, + "theoretical_loss": 4.624899914223355, + "tokens_seen": 131792896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048487462387161484, + "loss": 1.9739, + "theoretical_loss": 4.624595431455091, + "tokens_seen": 131858432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848645937813441, + "loss": 1.9901, + "theoretical_loss": 4.62429114233197, + "tokens_seen": 131923968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848545636910732, + "loss": 2.0468, + "theoretical_loss": 4.623987046634729, + "tokens_seen": 131989504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048484453360080244, + "loss": 1.9365, + "theoretical_loss": 4.623683144144465, + "tokens_seen": 132055040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848345035105316, + "loss": 1.9548, + "theoretical_loss": 4.623379434642632, + "tokens_seen": 132120576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004848244734202608, + "loss": 1.9606, + "theoretical_loss": 4.623075917911039, + "tokens_seen": 132186112 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048481444332999, + "loss": 2.0245, + "theoretical_loss": 4.622772593731848, + "tokens_seen": 132251648 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048480441323971917, + "loss": 2.0334, + "theoretical_loss": 4.622469461887577, + "tokens_seen": 132317184 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048479438314944835, + "loss": 2.0516, + "theoretical_loss": 4.622166522161096, + "tokens_seen": 132382720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847843530591776, + "loss": 1.9488, + "theoretical_loss": 4.621863774335628, + "tokens_seen": 132448256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847743229689067, + "loss": 2.0951, + "theoretical_loss": 4.621561218194747, + "tokens_seen": 132513792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048476429287863594, + "loss": 2.083, + "theoretical_loss": 4.62125885352238, + "tokens_seen": 132579328 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048475426278836507, + "loss": 2.0202, + "theoretical_loss": 4.620956680102801, + "tokens_seen": 132644864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847442326980943, + "loss": 2.0469, + "theoretical_loss": 4.620654697720637, + "tokens_seen": 132710400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847342026078235, + "loss": 2.1724, + "theoretical_loss": 4.620352906160861, + "tokens_seen": 132775936 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048472417251755267, + "loss": 2.15, + "theoretical_loss": 4.6200513052087935, + "tokens_seen": 132841472 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048471414242728185, + "loss": 2.0599, + "theoretical_loss": 4.6197498946501065, + "tokens_seen": 132907008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004847041123370111, + "loss": 2.0428, + "theoretical_loss": 4.6194486742708145, + "tokens_seen": 132972544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846940822467402, + "loss": 2.0128, + "theoretical_loss": 4.6191476438572785, + "tokens_seen": 133038080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048468405215646945, + "loss": 1.9712, + "theoretical_loss": 4.618846803196206, + "tokens_seen": 133103616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846740220661986, + "loss": 1.9696, + "theoretical_loss": 4.618546152074648, + "tokens_seen": 133169152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846639919759278, + "loss": 1.9895, + "theoretical_loss": 4.61824569028, + "tokens_seen": 133234688 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484653961885657, + "loss": 2.0042, + "theoretical_loss": 4.617945417599998, + "tokens_seen": 133300224 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048464393179538617, + "loss": 1.8727, + "theoretical_loss": 4.617645333822722, + "tokens_seen": 133365760 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048463390170511535, + "loss": 1.8333, + "theoretical_loss": 4.617345438736594, + "tokens_seen": 133431296 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048462387161484453, + "loss": 1.9441, + "theoretical_loss": 4.617045732130375, + "tokens_seen": 133496832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004846138415245737, + "loss": 1.8933, + "theoretical_loss": 4.616746213793167, + "tokens_seen": 133562368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048460381143430295, + "loss": 2.0485, + "theoretical_loss": 4.616446883514412, + "tokens_seen": 133627904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845937813440321, + "loss": 1.9272, + "theoretical_loss": 4.616147741083891, + "tokens_seen": 133693440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845837512537613, + "loss": 1.9667, + "theoretical_loss": 4.6158487862917195, + "tokens_seen": 133758976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048457372116349044, + "loss": 1.9183, + "theoretical_loss": 4.615550018928353, + "tokens_seen": 133824512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845636910732197, + "loss": 2.0627, + "theoretical_loss": 4.615251438784584, + "tokens_seen": 133890048 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048455366098294886, + "loss": 2.056, + "theoretical_loss": 4.61495304565154, + "tokens_seen": 133955584 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048454363089267804, + "loss": 2.0513, + "theoretical_loss": 4.614654839320682, + "tokens_seen": 134021120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845336008024072, + "loss": 1.966, + "theoretical_loss": 4.614356819583808, + "tokens_seen": 134086656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048452357071213645, + "loss": 2.062, + "theoretical_loss": 4.614058986233047, + "tokens_seen": 134152192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845135406218656, + "loss": 2.0407, + "theoretical_loss": 4.613761339060863, + "tokens_seen": 134217728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004845035105315948, + "loss": 1.9284, + "theoretical_loss": 4.613463877860051, + "tokens_seen": 134283264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 56110, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9489567279815674, + "objective/train/theoretical_loss": 4.61316660242374, + "objective/train/tokens_used": 154808800, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048449348044132394, + "loss": 2.0989, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844834503510532, + "loss": 2.068, + "theoretical_loss": 4.6128695125453865, + "tokens_seen": 134414336 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048447342026078236, + "loss": 1.9968, + "theoretical_loss": 4.612572608018779, + "tokens_seen": 134479872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048446339017051154, + "loss": 1.9617, + "theoretical_loss": 4.612275888638036, + "tokens_seen": 134545408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844533600802407, + "loss": 1.9385, + "theoretical_loss": 4.611979354197603, + "tokens_seen": 134610944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844433299899699, + "loss": 1.8816, + "theoretical_loss": 4.6116830044922565, + "tokens_seen": 134676480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844332998996991, + "loss": 2.0039, + "theoretical_loss": 4.611386839317097, + "tokens_seen": 134742016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844232698094283, + "loss": 1.9849, + "theoretical_loss": 4.611090858467554, + "tokens_seen": 134807552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048441323971915745, + "loss": 1.826, + "theoretical_loss": 4.610795061739384, + "tokens_seen": 134873088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004844032096288867, + "loss": 1.9837, + "theoretical_loss": 4.610499448928666, + "tokens_seen": 134938624 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048439317953861586, + "loss": 2.1326, + "theoretical_loss": 4.610204019831805, + "tokens_seen": 135004160 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048438314944834504, + "loss": 2.0422, + "theoretical_loss": 4.609908774245532, + "tokens_seen": 135069696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843731193580742, + "loss": 2.0954, + "theoretical_loss": 4.609613711966898, + "tokens_seen": 135135232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843630892678034, + "loss": 1.932, + "theoretical_loss": 4.60931883279328, + "tokens_seen": 135200768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843530591775326, + "loss": 1.8945, + "theoretical_loss": 4.6090241365223745, + "tokens_seen": 135266304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843430290872618, + "loss": 1.921, + "theoretical_loss": 4.6087296229522, + "tokens_seen": 135331840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048433299899699095, + "loss": 1.9309, + "theoretical_loss": 4.608435291881099, + "tokens_seen": 135397376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843229689067202, + "loss": 1.9828, + "theoretical_loss": 4.608141143107728, + "tokens_seen": 135462912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004843129388164493, + "loss": 1.9943, + "theoretical_loss": 4.607847176431067, + "tokens_seen": 135528448 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048430290872617855, + "loss": 1.9008, + "theoretical_loss": 4.6075533916504146, + "tokens_seen": 135593984 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048429287863590773, + "loss": 1.9616, + "theoretical_loss": 4.607259788565385, + "tokens_seen": 135659520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842828485456369, + "loss": 2.0339, + "theoretical_loss": 4.606966366975914, + "tokens_seen": 135725056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842728184553661, + "loss": 2.1954, + "theoretical_loss": 4.606673126682249, + "tokens_seen": 135790592 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048426278836509527, + "loss": 1.9346, + "theoretical_loss": 4.6063800674849595, + "tokens_seen": 135856128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048425275827482445, + "loss": 1.9896, + "theoretical_loss": 4.606087189184924, + "tokens_seen": 135921664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842427281845537, + "loss": 2.008, + "theoretical_loss": 4.6057944915833415, + "tokens_seen": 135987200 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048423269809428287, + "loss": 2.012, + "theoretical_loss": 4.6055019744817205, + "tokens_seen": 136052736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048422266800401205, + "loss": 2.0515, + "theoretical_loss": 4.605209637681886, + "tokens_seen": 136118272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842126379137413, + "loss": 1.8519, + "theoretical_loss": 4.604917480985975, + "tokens_seen": 136183808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004842026078234704, + "loss": 1.8513, + "theoretical_loss": 4.604625504196438, + "tokens_seen": 136249344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048419257773319965, + "loss": 1.9427, + "theoretical_loss": 4.604333707116033, + "tokens_seen": 136314880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841825476429288, + "loss": 2.1054, + "theoretical_loss": 4.604042089547834, + "tokens_seen": 136380416 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484172517552658, + "loss": 2.0194, + "theoretical_loss": 4.603750651295222, + "tokens_seen": 136445952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841624874623872, + "loss": 1.9492, + "theoretical_loss": 4.603459392161888, + "tokens_seen": 136511488 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048415245737211637, + "loss": 2.0089, + "theoretical_loss": 4.603168311951835, + "tokens_seen": 136577024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048414242728184555, + "loss": 2.1746, + "theoretical_loss": 4.60287741046937, + "tokens_seen": 136642560 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048413239719157473, + "loss": 1.9412, + "theoretical_loss": 4.602586687519111, + "tokens_seen": 136708096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841223671013039, + "loss": 2.0198, + "theoretical_loss": 4.602296142905981, + "tokens_seen": 136773632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048411233701103315, + "loss": 2.1171, + "theoretical_loss": 4.602005776435211, + "tokens_seen": 136839168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004841023069207623, + "loss": 1.892, + "theoretical_loss": 4.601715587912338, + "tokens_seen": 136904704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840922768304915, + "loss": 1.9793, + "theoretical_loss": 4.601425577143203, + "tokens_seen": 136970240 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048408224674022064, + "loss": 1.8991, + "theoretical_loss": 4.601135743933952, + "tokens_seen": 137035776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840722166499499, + "loss": 1.9715, + "theoretical_loss": 4.600846088091036, + "tokens_seen": 137101312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048406218655967906, + "loss": 1.9388, + "theoretical_loss": 4.600556609421206, + "tokens_seen": 137166848 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048405215646940824, + "loss": 2.0472, + "theoretical_loss": 4.600267307731521, + "tokens_seen": 137232384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840421263791374, + "loss": 2.0328, + "theoretical_loss": 4.5999781828293385, + "tokens_seen": 137297920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048403209628886665, + "loss": 1.9668, + "theoretical_loss": 4.5996892345223195, + "tokens_seen": 137363456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004840220661985958, + "loss": 1.9831, + "theoretical_loss": 4.599400462618423, + "tokens_seen": 137428992 + }, + { + "epoch": 0.04, + "learning_rate": 0.000484012036108325, + "loss": 1.9153, + "theoretical_loss": 4.59911186692591, + "tokens_seen": 137494528 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048400200601805414, + "loss": 1.8802, + "theoretical_loss": 4.598823447253343, + "tokens_seen": 137560064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 57180, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1683766841888428, + "objective/train/theoretical_loss": 4.598535203409582, + "objective/train/tokens_used": 158085600, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839919759277834, + "loss": 2.1031, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048398194583751256, + "loss": 1.916, + "theoretical_loss": 4.598247135203783, + "tokens_seen": 137691136 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048397191574724174, + "loss": 1.8867, + "theoretical_loss": 4.5979592424454045, + "tokens_seen": 137756672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839618856569709, + "loss": 1.9907, + "theoretical_loss": 4.597671524944198, + "tokens_seen": 137822208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839518555667001, + "loss": 2.0967, + "theoretical_loss": 4.597383982510216, + "tokens_seen": 137887744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839418254764293, + "loss": 1.9099, + "theoretical_loss": 4.597096614953802, + "tokens_seen": 137953280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839317953861585, + "loss": 1.92, + "theoretical_loss": 4.596809422085599, + "tokens_seen": 138018816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048392176529588765, + "loss": 2.0018, + "theoretical_loss": 4.596522403716542, + "tokens_seen": 138084352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004839117352056169, + "loss": 2.1222, + "theoretical_loss": 4.596235559657863, + "tokens_seen": 138149888 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048390170511534606, + "loss": 1.8306, + "theoretical_loss": 4.595948889721086, + "tokens_seen": 138215424 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048389167502507524, + "loss": 2.0664, + "theoretical_loss": 4.595662393718026, + "tokens_seen": 138280960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838816449348044, + "loss": 2.0821, + "theoretical_loss": 4.595376071460796, + "tokens_seen": 138346496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838716148445336, + "loss": 1.9072, + "theoretical_loss": 4.595089922761796, + "tokens_seen": 138412032 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838615847542628, + "loss": 1.9548, + "theoretical_loss": 4.594803947433717, + "tokens_seen": 138477568 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483851554663992, + "loss": 1.9806, + "theoretical_loss": 4.594518145289545, + "tokens_seen": 138543104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048384152457372115, + "loss": 2.0681, + "theoretical_loss": 4.594232516142553, + "tokens_seen": 138608640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838314944834504, + "loss": 2.1069, + "theoretical_loss": 4.593947059806302, + "tokens_seen": 138674176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004838214643931795, + "loss": 1.9247, + "theoretical_loss": 4.593661776094647, + "tokens_seen": 138739712 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048381143430290875, + "loss": 1.8671, + "theoretical_loss": 4.593376664821726, + "tokens_seen": 138805248 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048380140421263793, + "loss": 1.8299, + "theoretical_loss": 4.593091725801967, + "tokens_seen": 138870784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837913741223671, + "loss": 1.9935, + "theoretical_loss": 4.592806958850087, + "tokens_seen": 138936320 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837813440320963, + "loss": 1.9699, + "theoretical_loss": 4.592522363781086, + "tokens_seen": 139001856 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048377131394182547, + "loss": 1.9641, + "theoretical_loss": 4.592237940410252, + "tokens_seen": 139067392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048376128385155465, + "loss": 1.9024, + "theoretical_loss": 4.591953688553159, + "tokens_seen": 139132928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837512537612839, + "loss": 1.9399, + "theoretical_loss": 4.591669608025663, + "tokens_seen": 139198464 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483741223671013, + "loss": 2.0614, + "theoretical_loss": 4.591385698643908, + "tokens_seen": 139264000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048373119358074225, + "loss": 2.0291, + "theoretical_loss": 4.591101960224321, + "tokens_seen": 139329536 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048372116349047143, + "loss": 1.9599, + "theoretical_loss": 4.590818392583608, + "tokens_seen": 139395072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837111334002006, + "loss": 1.8323, + "theoretical_loss": 4.590534995538762, + "tokens_seen": 139460608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004837011033099298, + "loss": 2.0178, + "theoretical_loss": 4.590251768907057, + "tokens_seen": 139526144 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483691073219659, + "loss": 1.8883, + "theoretical_loss": 4.5899687125060495, + "tokens_seen": 139591680 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048368104312938816, + "loss": 2.0552, + "theoretical_loss": 4.589685826153573, + "tokens_seen": 139657216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836710130391174, + "loss": 1.9134, + "theoretical_loss": 4.589403109667746, + "tokens_seen": 139722752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836609829488465, + "loss": 2.0406, + "theoretical_loss": 4.589120562866961, + "tokens_seen": 139788288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048365095285857575, + "loss": 2.0315, + "theoretical_loss": 4.588838185569897, + "tokens_seen": 139853824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836409227683049, + "loss": 1.9082, + "theoretical_loss": 4.588555977595506, + "tokens_seen": 139919360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836308926780341, + "loss": 1.9824, + "theoretical_loss": 4.58827393876302, + "tokens_seen": 139984896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836208625877633, + "loss": 1.9564, + "theoretical_loss": 4.587992068891948, + "tokens_seen": 140050432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004836108324974925, + "loss": 1.831, + "theoretical_loss": 4.587710367802076, + "tokens_seen": 140115968 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048360080240722166, + "loss": 1.9553, + "theoretical_loss": 4.587428835313469, + "tokens_seen": 140181504 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048359077231695084, + "loss": 1.9849, + "theoretical_loss": 4.587147471246463, + "tokens_seen": 140247040 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048358074222668, + "loss": 1.9275, + "theoretical_loss": 4.586866275421673, + "tokens_seen": 140312576 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048357071213640926, + "loss": 2.0091, + "theoretical_loss": 4.586585247659986, + "tokens_seen": 140378112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835606820461384, + "loss": 1.8949, + "theoretical_loss": 4.586304387782567, + "tokens_seen": 140443648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835506519558676, + "loss": 1.9336, + "theoretical_loss": 4.5860236956108515, + "tokens_seen": 140509184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835406218655968, + "loss": 1.9492, + "theoretical_loss": 4.585743170966548, + "tokens_seen": 140574720 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483530591775326, + "loss": 1.9159, + "theoretical_loss": 4.58546281367164, + "tokens_seen": 140640256 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048352056168505516, + "loss": 2.1478, + "theoretical_loss": 4.585182623548381, + "tokens_seen": 140705792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048351053159478434, + "loss": 1.9322, + "theoretical_loss": 4.584902600419295, + "tokens_seen": 140771328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004835005015045135, + "loss": 1.9691, + "theoretical_loss": 4.584622744107181, + "tokens_seen": 140836864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 58312, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8346216678619385, + "objective/train/theoretical_loss": 4.584343054435103, + "objective/train/tokens_used": 161362400, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048349047141424276, + "loss": 1.8229, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048348044132397194, + "loss": 1.9603, + "theoretical_loss": 4.5840635312264, + "tokens_seen": 140967936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834704112337011, + "loss": 1.9111, + "theoretical_loss": 4.583784174304677, + "tokens_seen": 141033472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834603811434303, + "loss": 1.8075, + "theoretical_loss": 4.583504983493809, + "tokens_seen": 141099008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834503510531595, + "loss": 2.0464, + "theoretical_loss": 4.583225958617938, + "tokens_seen": 141164544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834403209628887, + "loss": 1.9808, + "theoretical_loss": 4.582947099501475, + "tokens_seen": 141230080 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048343029087261785, + "loss": 1.9696, + "theoretical_loss": 4.5826684059691, + "tokens_seen": 141295616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004834202607823471, + "loss": 2.0072, + "theoretical_loss": 4.582389877845754, + "tokens_seen": 141361152 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048341023069207626, + "loss": 1.9796, + "theoretical_loss": 4.58211151495665, + "tokens_seen": 141426688 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048340020060180544, + "loss": 1.7457, + "theoretical_loss": 4.581833317127263, + "tokens_seen": 141492224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833901705115346, + "loss": 1.9425, + "theoretical_loss": 4.581555284183335, + "tokens_seen": 141557760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833801404212638, + "loss": 2.0016, + "theoretical_loss": 4.581277415950872, + "tokens_seen": 141623296 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483370110330993, + "loss": 1.9705, + "theoretical_loss": 4.580999712256142, + "tokens_seen": 141688832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833600802407222, + "loss": 1.9305, + "theoretical_loss": 4.5807221729256815, + "tokens_seen": 141754368 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048335005015045135, + "loss": 1.9716, + "theoretical_loss": 4.580444797786283, + "tokens_seen": 141819904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833400200601806, + "loss": 1.9686, + "theoretical_loss": 4.5801675866650085, + "tokens_seen": 141885440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004833299899699097, + "loss": 2.0312, + "theoretical_loss": 4.579890539389177, + "tokens_seen": 141950976 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048331995987963895, + "loss": 1.9849, + "theoretical_loss": 4.579613655786369, + "tokens_seen": 142016512 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048330992978936813, + "loss": 1.9195, + "theoretical_loss": 4.579336935684431, + "tokens_seen": 142082048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832998996990973, + "loss": 1.9682, + "theoretical_loss": 4.579060378911464, + "tokens_seen": 142147584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832898696088265, + "loss": 1.8982, + "theoretical_loss": 4.578783985295832, + "tokens_seen": 142213120 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048327983951855567, + "loss": 2.0073, + "theoretical_loss": 4.578507754666157, + "tokens_seen": 142278656 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048326980942828485, + "loss": 2.076, + "theoretical_loss": 4.57823168685132, + "tokens_seen": 142344192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832597793380141, + "loss": 2.0302, + "theoretical_loss": 4.577955781680462, + "tokens_seen": 142409728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832497492477432, + "loss": 1.8881, + "theoretical_loss": 4.57768003898298, + "tokens_seen": 142475264 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048323971915747245, + "loss": 1.9153, + "theoretical_loss": 4.577404458588529, + "tokens_seen": 142540800 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048322968906720163, + "loss": 1.9798, + "theoretical_loss": 4.577129040327021, + "tokens_seen": 142606336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004832196589769308, + "loss": 1.9351, + "theoretical_loss": 4.576853784028622, + "tokens_seen": 142671872 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048320962888666, + "loss": 1.8715, + "theoretical_loss": 4.576578689523759, + "tokens_seen": 142737408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831995987963892, + "loss": 1.9599, + "theoretical_loss": 4.576303756643108, + "tokens_seen": 142802944 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048318956870611836, + "loss": 1.9164, + "theoretical_loss": 4.576028985217606, + "tokens_seen": 142868480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831795386158476, + "loss": 1.9976, + "theoretical_loss": 4.5757543750784375, + "tokens_seen": 142934016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831695085255767, + "loss": 1.99, + "theoretical_loss": 4.5754799260570485, + "tokens_seen": 142999552 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048315947843530595, + "loss": 2.0604, + "theoretical_loss": 4.575205637985132, + "tokens_seen": 143065088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831494483450351, + "loss": 2.0393, + "theoretical_loss": 4.574931510694635, + "tokens_seen": 143130624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831394182547643, + "loss": 1.8658, + "theoretical_loss": 4.574657544017761, + "tokens_seen": 143196160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831293881644935, + "loss": 2.0266, + "theoretical_loss": 4.57438373778696, + "tokens_seen": 143261696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004831193580742227, + "loss": 1.9764, + "theoretical_loss": 4.574110091834936, + "tokens_seen": 143327232 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048310932798395186, + "loss": 1.9439, + "theoretical_loss": 4.573836605994645, + "tokens_seen": 143392768 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048309929789368104, + "loss": 1.9752, + "theoretical_loss": 4.573563280099288, + "tokens_seen": 143458304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830892678034102, + "loss": 2.0371, + "theoretical_loss": 4.573290113982324, + "tokens_seen": 143523840 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048307923771313946, + "loss": 1.7958, + "theoretical_loss": 4.573017107477455, + "tokens_seen": 143589376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830692076228686, + "loss": 1.9634, + "theoretical_loss": 4.572744260418633, + "tokens_seen": 143654912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830591775325978, + "loss": 2.0066, + "theoretical_loss": 4.572471572640059, + "tokens_seen": 143720448 + }, + { + "epoch": 0.04, + "learning_rate": 0.000483049147442327, + "loss": 2.0151, + "theoretical_loss": 4.572199043976185, + "tokens_seen": 143785984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830391173520562, + "loss": 1.875, + "theoretical_loss": 4.571926674261705, + "tokens_seen": 143851520 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048302908726178536, + "loss": 1.8238, + "theoretical_loss": 4.571654463331563, + "tokens_seen": 143917056 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048301905717151454, + "loss": 1.9462, + "theoretical_loss": 4.571382411020949, + "tokens_seen": 143982592 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004830090270812437, + "loss": 1.9643, + "theoretical_loss": 4.5711105171653, + "tokens_seen": 144048128 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048299899699097296, + "loss": 1.9603, + "theoretical_loss": 4.570838781600296, + "tokens_seen": 144113664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 59583, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.86801278591156, + "objective/train/theoretical_loss": 4.570567204161865, + "objective/train/tokens_used": 164639200, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829889669007021, + "loss": 1.9536, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829789368104313, + "loss": 1.8541, + "theoretical_loss": 4.570295784686176, + "tokens_seen": 144244736 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048296890672016045, + "loss": 1.9839, + "theoretical_loss": 4.570024523009646, + "tokens_seen": 144310272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829588766298897, + "loss": 1.9506, + "theoretical_loss": 4.569753418968933, + "tokens_seen": 144375808 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048294884653961887, + "loss": 1.9107, + "theoretical_loss": 4.569482472400939, + "tokens_seen": 144441344 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048293881644934805, + "loss": 1.9674, + "theoretical_loss": 4.5692116831428065, + "tokens_seen": 144506880 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048292878635907723, + "loss": 1.9186, + "theoretical_loss": 4.568941051031925, + "tokens_seen": 144572416 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048291875626880646, + "loss": 1.9502, + "theoretical_loss": 4.568670575905921, + "tokens_seen": 144637952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004829087261785356, + "loss": 2.015, + "theoretical_loss": 4.568400257602666, + "tokens_seen": 144703488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828986960882648, + "loss": 2.0003, + "theoretical_loss": 4.568130095960268, + "tokens_seen": 144769024 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048288866599799395, + "loss": 2.0017, + "theoretical_loss": 4.56786009081708, + "tokens_seen": 144834560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828786359077232, + "loss": 1.9382, + "theoretical_loss": 4.567590242011692, + "tokens_seen": 144900096 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048286860581745237, + "loss": 1.9533, + "theoretical_loss": 4.567320549382931, + "tokens_seen": 144965632 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048285857572718155, + "loss": 1.9779, + "theoretical_loss": 4.56705101276987, + "tokens_seen": 145031168 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048284854563691073, + "loss": 1.8978, + "theoretical_loss": 4.566781632011814, + "tokens_seen": 145096704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828385155466399, + "loss": 1.8425, + "theoretical_loss": 4.566512406948309, + "tokens_seen": 145162240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004828284854563691, + "loss": 1.937, + "theoretical_loss": 4.566243337419137, + "tokens_seen": 145227776 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048281845536609833, + "loss": 1.8886, + "theoretical_loss": 4.5659744232643185, + "tokens_seen": 145293312 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048280842527582746, + "loss": 2.1346, + "theoretical_loss": 4.565705664324111, + "tokens_seen": 145358848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827983951855567, + "loss": 1.8264, + "theoretical_loss": 4.565437060439006, + "tokens_seen": 145424384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827883650952858, + "loss": 1.9887, + "theoretical_loss": 4.565168611449732, + "tokens_seen": 145489920 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048277833500501505, + "loss": 1.9618, + "theoretical_loss": 4.564900317197251, + "tokens_seen": 145555456 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048276830491474423, + "loss": 2.0835, + "theoretical_loss": 4.564632177522766, + "tokens_seen": 145620992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827582748244734, + "loss": 1.9279, + "theoretical_loss": 4.564364192267705, + "tokens_seen": 145686528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827482447342026, + "loss": 1.9615, + "theoretical_loss": 4.564096361273736, + "tokens_seen": 145752064 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048273821464393183, + "loss": 1.9952, + "theoretical_loss": 4.563828684382761, + "tokens_seen": 145817600 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482728184553661, + "loss": 2.0234, + "theoretical_loss": 4.563561161436912, + "tokens_seen": 145883136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827181544633902, + "loss": 1.7758, + "theoretical_loss": 4.563293792278554, + "tokens_seen": 145948672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004827081243731194, + "loss": 1.8866, + "theoretical_loss": 4.563026576750286, + "tokens_seen": 146014208 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048269809428284856, + "loss": 1.9297, + "theoretical_loss": 4.562759514694935, + "tokens_seen": 146079744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826880641925778, + "loss": 1.9605, + "theoretical_loss": 4.5624926059555655, + "tokens_seen": 146145280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826780341023069, + "loss": 1.9273, + "theoretical_loss": 4.562225850375466, + "tokens_seen": 146210816 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048266800401203615, + "loss": 1.9298, + "theoretical_loss": 4.561959247798159, + "tokens_seen": 146276352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826579739217653, + "loss": 1.9302, + "theoretical_loss": 4.561692798067396, + "tokens_seen": 146341888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826479438314945, + "loss": 1.8468, + "theoretical_loss": 4.561426501027158, + "tokens_seen": 146407424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826379137412237, + "loss": 1.9553, + "theoretical_loss": 4.561160356521656, + "tokens_seen": 146472960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004826278836509529, + "loss": 1.8782, + "theoretical_loss": 4.5608943643953275, + "tokens_seen": 146538496 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048261785356068206, + "loss": 2.0229, + "theoretical_loss": 4.56062852449284, + "tokens_seen": 146604032 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048260782347041124, + "loss": 1.8713, + "theoretical_loss": 4.5603628366590865, + "tokens_seen": 146669568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825977933801404, + "loss": 2.0677, + "theoretical_loss": 4.560097300739191, + "tokens_seen": 146735104 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048258776328986966, + "loss": 1.9425, + "theoretical_loss": 4.5598319165785, + "tokens_seen": 146800640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825777331995988, + "loss": 1.9439, + "theoretical_loss": 4.55956668402259, + "tokens_seen": 146866176 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482567703109328, + "loss": 1.7906, + "theoretical_loss": 4.559301602917261, + "tokens_seen": 146931712 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825576730190572, + "loss": 1.9162, + "theoretical_loss": 4.55903667310854, + "tokens_seen": 146997248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825476429287864, + "loss": 1.8751, + "theoretical_loss": 4.558771894442677, + "tokens_seen": 147062784 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048253761283851556, + "loss": 1.9619, + "theoretical_loss": 4.558507266766151, + "tokens_seen": 147128320 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048252758274824474, + "loss": 1.9459, + "theoretical_loss": 4.558242789925659, + "tokens_seen": 147193856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004825175526579739, + "loss": 1.8468, + "theoretical_loss": 4.557978463768127, + "tokens_seen": 147259392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048250752256770316, + "loss": 1.9444, + "theoretical_loss": 4.557714288140703, + "tokens_seen": 147324928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824974924774323, + "loss": 1.8815, + "theoretical_loss": 4.557450262890757, + "tokens_seen": 147390464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 60747, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8011186122894287, + "objective/train/theoretical_loss": 4.557186387865883, + "objective/train/tokens_used": 167916000, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824874623871615, + "loss": 1.8386, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048247743229689065, + "loss": 1.8969, + "theoretical_loss": 4.556922662913895, + "tokens_seen": 147521536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824674022066199, + "loss": 1.9275, + "theoretical_loss": 4.55665908788283, + "tokens_seen": 147587072 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048245737211634907, + "loss": 2.0323, + "theoretical_loss": 4.556395662620945, + "tokens_seen": 147652608 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048244734202607825, + "loss": 1.8551, + "theoretical_loss": 4.5561323869767225, + "tokens_seen": 147718144 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048243731193580743, + "loss": 1.9199, + "theoretical_loss": 4.555869260798859, + "tokens_seen": 147783680 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048242728184553666, + "loss": 1.9528, + "theoretical_loss": 4.5556062839362745, + "tokens_seen": 147849216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004824172517552658, + "loss": 1.9735, + "theoretical_loss": 4.555343456238109, + "tokens_seen": 147914752 + }, + { + "epoch": 0.04, + "learning_rate": 0.000482407221664995, + "loss": 2.0058, + "theoretical_loss": 4.55508077755372, + "tokens_seen": 147980288 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048239719157472415, + "loss": 1.9315, + "theoretical_loss": 4.554818247732685, + "tokens_seen": 148045824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823871614844534, + "loss": 1.9411, + "theoretical_loss": 4.554555866624798, + "tokens_seen": 148111360 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048237713139418257, + "loss": 1.9583, + "theoretical_loss": 4.554293634080074, + "tokens_seen": 148176896 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048236710130391175, + "loss": 1.8504, + "theoretical_loss": 4.554031549948743, + "tokens_seen": 148242432 + }, + { + "epoch": 0.04, + "learning_rate": 0.00048235707121364093, + "loss": 1.8867, + "theoretical_loss": 4.5537696140812525, + "tokens_seen": 148307968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823470411233701, + "loss": 1.986, + "theoretical_loss": 4.553507826328268, + "tokens_seen": 148373504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0004823370110330993, + "loss": 1.8366, + "theoretical_loss": 4.553246186540671, + "tokens_seen": 148439040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048232698094282853, + "loss": 1.8335, + "theoretical_loss": 4.552984694569555, + "tokens_seen": 148504576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048231695085255766, + "loss": 1.9898, + "theoretical_loss": 4.552723350266236, + "tokens_seen": 148570112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004823069207622869, + "loss": 1.9258, + "theoretical_loss": 4.552462153482239, + "tokens_seen": 148635648 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482296890672016, + "loss": 1.9236, + "theoretical_loss": 4.552201104069307, + "tokens_seen": 148701184 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048228686058174525, + "loss": 1.9716, + "theoretical_loss": 4.551940201879396, + "tokens_seen": 148766720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048227683049147443, + "loss": 1.7993, + "theoretical_loss": 4.551679446764677, + "tokens_seen": 148832256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822668004012036, + "loss": 2.0361, + "theoretical_loss": 4.551418838577531, + "tokens_seen": 148897792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822567703109328, + "loss": 1.8651, + "theoretical_loss": 4.551158377170557, + "tokens_seen": 148963328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048224674022066203, + "loss": 1.7318, + "theoretical_loss": 4.5508980623965645, + "tokens_seen": 149028864 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048223671013039116, + "loss": 2.0264, + "theoretical_loss": 4.550637894108572, + "tokens_seen": 149094400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822266800401204, + "loss": 1.8048, + "theoretical_loss": 4.5503778721598165, + "tokens_seen": 149159936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004822166499498495, + "loss": 1.7884, + "theoretical_loss": 4.55011799640374, + "tokens_seen": 149225472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048220661985957876, + "loss": 1.9347, + "theoretical_loss": 4.549858266694001, + "tokens_seen": 149291008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048219658976930794, + "loss": 1.8797, + "theoretical_loss": 4.549598682884463, + "tokens_seen": 149356544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821865596790371, + "loss": 1.9403, + "theoretical_loss": 4.549339244829204, + "tokens_seen": 149422080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821765295887663, + "loss": 1.8398, + "theoretical_loss": 4.549079952382511, + "tokens_seen": 149487616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821664994984955, + "loss": 1.9258, + "theoretical_loss": 4.548820805398881, + "tokens_seen": 149553152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048215646940822466, + "loss": 1.7724, + "theoretical_loss": 4.548561803733017, + "tokens_seen": 149618688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821464393179539, + "loss": 2.1288, + "theoretical_loss": 4.548302947239834, + "tokens_seen": 149684224 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482136409227683, + "loss": 1.9011, + "theoretical_loss": 4.548044235774454, + "tokens_seen": 149749760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048212637913741226, + "loss": 1.9445, + "theoretical_loss": 4.547785669192207, + "tokens_seen": 149815296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821163490471414, + "loss": 1.9472, + "theoretical_loss": 4.547527247348631, + "tokens_seen": 149880832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004821063189568706, + "loss": 1.9353, + "theoretical_loss": 4.547268970099471, + "tokens_seen": 149946368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820962888665998, + "loss": 1.9063, + "theoretical_loss": 4.547010837300677, + "tokens_seen": 150011904 + }, + { + "epoch": 0.05, + "learning_rate": 0.000482086258776329, + "loss": 1.9251, + "theoretical_loss": 4.546752848808408, + "tokens_seen": 150077440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048207622868605816, + "loss": 1.8986, + "theoretical_loss": 4.546495004479027, + "tokens_seen": 150142976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820661985957874, + "loss": 1.8271, + "theoretical_loss": 4.546237304169104, + "tokens_seen": 150208512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048205616850551653, + "loss": 1.9658, + "theoretical_loss": 4.545979747735413, + "tokens_seen": 150274048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048204613841524576, + "loss": 1.7625, + "theoretical_loss": 4.5457223350349345, + "tokens_seen": 150339584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820361083249749, + "loss": 1.8313, + "theoretical_loss": 4.545465065924851, + "tokens_seen": 150405120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820260782347041, + "loss": 1.8024, + "theoretical_loss": 4.54520794026255, + "tokens_seen": 150470656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820160481444333, + "loss": 1.8198, + "theoretical_loss": 4.544950957905625, + "tokens_seen": 150536192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004820060180541625, + "loss": 1.8642, + "theoretical_loss": 4.54469411871187, + "tokens_seen": 150601728 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048199598796389167, + "loss": 1.8348, + "theoretical_loss": 4.544437422539282, + "tokens_seen": 150667264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 61752, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.926950216293335, + "objective/train/theoretical_loss": 4.544180869246063, + "objective/train/tokens_used": 171192800, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048198595787362085, + "loss": 1.9285, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004819759277833501, + "loss": 2.0443, + "theoretical_loss": 4.543924458690612, + "tokens_seen": 150798336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048196589769307927, + "loss": 1.9853, + "theoretical_loss": 4.5436681907315375, + "tokens_seen": 150863872 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048195586760280845, + "loss": 2.0087, + "theoretical_loss": 4.543412065227642, + "tokens_seen": 150929408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048194583751253763, + "loss": 1.8587, + "theoretical_loss": 4.5431560820379335, + "tokens_seen": 150994944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048193580742226686, + "loss": 1.8614, + "theoretical_loss": 4.54290024102162, + "tokens_seen": 151060480 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481925777331996, + "loss": 1.9264, + "theoretical_loss": 4.542644542038108, + "tokens_seen": 151126016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004819157472417252, + "loss": 2.0153, + "theoretical_loss": 4.542388984947005, + "tokens_seen": 151191552 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048190571715145435, + "loss": 1.8648, + "theoretical_loss": 4.5421335696081195, + "tokens_seen": 151257088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818956870611836, + "loss": 1.8412, + "theoretical_loss": 4.541878295881456, + "tokens_seen": 151322624 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048188565697091277, + "loss": 1.8328, + "theoretical_loss": 4.541623163627221, + "tokens_seen": 151388160 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048187562688064195, + "loss": 1.8004, + "theoretical_loss": 4.541368172705817, + "tokens_seen": 151453696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048186559679037113, + "loss": 1.8551, + "theoretical_loss": 4.541113322977846, + "tokens_seen": 151519232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818555667001003, + "loss": 1.908, + "theoretical_loss": 4.540858614304108, + "tokens_seen": 151584768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818455366098295, + "loss": 1.8386, + "theoretical_loss": 4.5406040465455995, + "tokens_seen": 151650304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048183550651955873, + "loss": 1.7662, + "theoretical_loss": 4.5403496195635125, + "tokens_seen": 151715840 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048182547642928786, + "loss": 1.9597, + "theoretical_loss": 4.54009533321924, + "tokens_seen": 151781376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818154463390171, + "loss": 1.7955, + "theoretical_loss": 4.539841187374366, + "tokens_seen": 151846912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004818054162487462, + "loss": 1.8802, + "theoretical_loss": 4.539587181890674, + "tokens_seen": 151912448 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048179538615847545, + "loss": 1.8947, + "theoretical_loss": 4.539333316630143, + "tokens_seen": 151977984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048178535606820463, + "loss": 1.7664, + "theoretical_loss": 4.539079591454943, + "tokens_seen": 152043520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817753259779338, + "loss": 1.9939, + "theoretical_loss": 4.5388260062274455, + "tokens_seen": 152109056 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481765295887663, + "loss": 1.8528, + "theoretical_loss": 4.538572560810211, + "tokens_seen": 152174592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048175526579739223, + "loss": 1.8393, + "theoretical_loss": 4.538319255065995, + "tokens_seen": 152240128 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048174523570712136, + "loss": 1.8312, + "theoretical_loss": 4.53806608885775, + "tokens_seen": 152305664 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817352056168506, + "loss": 2.0654, + "theoretical_loss": 4.537813062048618, + "tokens_seen": 152371200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004817251755265797, + "loss": 1.812, + "theoretical_loss": 4.537560174501936, + "tokens_seen": 152436736 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048171514543630896, + "loss": 1.8973, + "theoretical_loss": 4.537307426081232, + "tokens_seen": 152502272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048170511534603814, + "loss": 1.9753, + "theoretical_loss": 4.53705481665023, + "tokens_seen": 152567808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816950852557673, + "loss": 1.9537, + "theoretical_loss": 4.536802346072841, + "tokens_seen": 152633344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816850551654965, + "loss": 1.8685, + "theoretical_loss": 4.536550014213171, + "tokens_seen": 152698880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816750250752257, + "loss": 1.9247, + "theoretical_loss": 4.536297820935517, + "tokens_seen": 152764416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048166499498495486, + "loss": 1.8407, + "theoretical_loss": 4.536045766104366, + "tokens_seen": 152829952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816549648946841, + "loss": 2.0014, + "theoretical_loss": 4.535793849584396, + "tokens_seen": 152895488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816449348044132, + "loss": 2.0279, + "theoretical_loss": 4.535542071240473, + "tokens_seen": 152961024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048163490471414246, + "loss": 1.9047, + "theoretical_loss": 4.535290430937657, + "tokens_seen": 153026560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816248746238716, + "loss": 1.8811, + "theoretical_loss": 4.535038928541194, + "tokens_seen": 153092096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004816148445336008, + "loss": 1.8721, + "theoretical_loss": 4.534787563916522, + "tokens_seen": 153157632 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048160481444333, + "loss": 1.7715, + "theoretical_loss": 4.534536336929264, + "tokens_seen": 153223168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815947843530592, + "loss": 1.9716, + "theoretical_loss": 4.534285247445236, + "tokens_seen": 153288704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048158475426278837, + "loss": 1.9086, + "theoretical_loss": 4.534034295330439, + "tokens_seen": 153354240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815747241725176, + "loss": 1.8321, + "theoretical_loss": 4.5337834804510635, + "tokens_seen": 153419776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048156469408224673, + "loss": 2.0433, + "theoretical_loss": 4.533532802673484, + "tokens_seen": 153485312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048155466399197596, + "loss": 1.8367, + "theoretical_loss": 4.533282261864267, + "tokens_seen": 153550848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815446339017051, + "loss": 2.0128, + "theoretical_loss": 4.533031857890164, + "tokens_seen": 153616384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815346038114343, + "loss": 1.8866, + "theoretical_loss": 4.532781590618109, + "tokens_seen": 153681920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815245737211635, + "loss": 2.0096, + "theoretical_loss": 4.532531459915229, + "tokens_seen": 153747456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004815145436308927, + "loss": 1.939, + "theoretical_loss": 4.532281465648831, + "tokens_seen": 153812992 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048150451354062187, + "loss": 1.9188, + "theoretical_loss": 4.532031607686411, + "tokens_seen": 153878528 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048149448345035105, + "loss": 1.888, + "theoretical_loss": 4.531781885895647, + "tokens_seen": 153944064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 63027, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1068944931030273, + "objective/train/theoretical_loss": 4.531532300144404, + "objective/train/tokens_used": 174469600, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048148445336008023, + "loss": 1.9998, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048147442326980947, + "loss": 1.9468, + "theoretical_loss": 4.531282850300731, + "tokens_seen": 154075136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004814643931795386, + "loss": 2.107, + "theoretical_loss": 4.53103353623286, + "tokens_seen": 154140672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048145436308926783, + "loss": 1.9191, + "theoretical_loss": 4.530784357809209, + "tokens_seen": 154206208 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048144433299899696, + "loss": 1.8021, + "theoretical_loss": 4.5305353148983745, + "tokens_seen": 154271744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004814343029087262, + "loss": 1.9151, + "theoretical_loss": 4.530286407369143, + "tokens_seen": 154337280 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048142427281845537, + "loss": 2.0267, + "theoretical_loss": 4.530037635090476, + "tokens_seen": 154402816 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048141424272818455, + "loss": 1.9752, + "theoretical_loss": 4.529788997931526, + "tokens_seen": 154468352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048140421263791373, + "loss": 1.8725, + "theoretical_loss": 4.52954049576162, + "tokens_seen": 154533888 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048139418254764297, + "loss": 1.9645, + "theoretical_loss": 4.529292128450271, + "tokens_seen": 154599424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813841524573721, + "loss": 1.9127, + "theoretical_loss": 4.52904389586717, + "tokens_seen": 154664960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048137412236710133, + "loss": 1.8989, + "theoretical_loss": 4.528795797882193, + "tokens_seen": 154730496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048136409227683046, + "loss": 1.8918, + "theoretical_loss": 4.528547834365394, + "tokens_seen": 154796032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813540621865597, + "loss": 1.8171, + "theoretical_loss": 4.52830000518701, + "tokens_seen": 154861568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813440320962889, + "loss": 1.8832, + "theoretical_loss": 4.528052310217453, + "tokens_seen": 154927104 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048133400200601806, + "loss": 1.8208, + "theoretical_loss": 4.527804749327321, + "tokens_seen": 154992640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048132397191574724, + "loss": 2.0211, + "theoretical_loss": 4.527557322387387, + "tokens_seen": 155058176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813139418254764, + "loss": 1.9534, + "theoretical_loss": 4.527310029268604, + "tokens_seen": 155123712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004813039117352056, + "loss": 1.893, + "theoretical_loss": 4.527062869842107, + "tokens_seen": 155189248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048129388164493483, + "loss": 1.9665, + "theoretical_loss": 4.526815843979204, + "tokens_seen": 155254784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048128385155466396, + "loss": 2.0227, + "theoretical_loss": 4.5265689515513845, + "tokens_seen": 155320320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812738214643932, + "loss": 1.9324, + "theoretical_loss": 4.526322192430316, + "tokens_seen": 155385856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812637913741223, + "loss": 1.8624, + "theoretical_loss": 4.526075566487842, + "tokens_seen": 155451392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048125376128385156, + "loss": 1.799, + "theoretical_loss": 4.525829073595984, + "tokens_seen": 155516928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048124373119358074, + "loss": 1.7412, + "theoretical_loss": 4.5255827136269415, + "tokens_seen": 155582464 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812337011033099, + "loss": 1.9715, + "theoretical_loss": 4.525336486453088, + "tokens_seen": 155648000 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048122367101303916, + "loss": 1.994, + "theoretical_loss": 4.525090391946974, + "tokens_seen": 155713536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048121364092276834, + "loss": 1.8303, + "theoretical_loss": 4.524844429981327, + "tokens_seen": 155779072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004812036108324975, + "loss": 1.8771, + "theoretical_loss": 4.52459860042905, + "tokens_seen": 155844608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811935807422267, + "loss": 2.0739, + "theoretical_loss": 4.524352903163219, + "tokens_seen": 155910144 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811835506519559, + "loss": 1.7648, + "theoretical_loss": 4.524107338057089, + "tokens_seen": 155975680 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048117352056168506, + "loss": 1.9343, + "theoretical_loss": 4.523861904984086, + "tokens_seen": 156041216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811634904714143, + "loss": 1.7645, + "theoretical_loss": 4.5236166038178105, + "tokens_seen": 156106752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811534603811434, + "loss": 1.858, + "theoretical_loss": 4.52337143443204, + "tokens_seen": 156172288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048114343029087266, + "loss": 1.9444, + "theoretical_loss": 4.523126396700722, + "tokens_seen": 156237824 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811334002006018, + "loss": 2.0111, + "theoretical_loss": 4.522881490497981, + "tokens_seen": 156303360 + }, + { + "epoch": 0.05, + "learning_rate": 0.000481123370110331, + "loss": 1.8464, + "theoretical_loss": 4.522636715698111, + "tokens_seen": 156368896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811133400200602, + "loss": 1.9, + "theoretical_loss": 4.522392072175582, + "tokens_seen": 156434432 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004811033099297894, + "loss": 1.9183, + "theoretical_loss": 4.522147559805034, + "tokens_seen": 156499968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048109327983951857, + "loss": 1.8835, + "theoretical_loss": 4.5219031784612795, + "tokens_seen": 156565504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810832497492478, + "loss": 1.7547, + "theoretical_loss": 4.521658928019305, + "tokens_seen": 156631040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048107321965897693, + "loss": 1.9012, + "theoretical_loss": 4.521414808354266, + "tokens_seen": 156696576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048106318956870616, + "loss": 1.7969, + "theoretical_loss": 4.52117081934149, + "tokens_seen": 156762112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810531594784353, + "loss": 1.7592, + "theoretical_loss": 4.520926960856476, + "tokens_seen": 156827648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810431293881645, + "loss": 1.9087, + "theoretical_loss": 4.520683232774893, + "tokens_seen": 156893184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810330992978937, + "loss": 1.8933, + "theoretical_loss": 4.520439634972579, + "tokens_seen": 156958720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004810230692076229, + "loss": 1.797, + "theoretical_loss": 4.520196167325546, + "tokens_seen": 157024256 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048101303911735207, + "loss": 1.9547, + "theoretical_loss": 4.519952829709972, + "tokens_seen": 157089792 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048100300902708125, + "loss": 1.8236, + "theoretical_loss": 4.519709622002205, + "tokens_seen": 157155328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048099297893681043, + "loss": 1.8369, + "theoretical_loss": 4.519466544078762, + "tokens_seen": 157220864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 64097, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0807628631591797, + "objective/train/theoretical_loss": 4.519223595816332, + "objective/train/tokens_used": 177746400, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048098294884653967, + "loss": 2.0891, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004809729187562688, + "loss": 1.7834, + "theoretical_loss": 4.518980777091768, + "tokens_seen": 157351936 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048096288866599803, + "loss": 1.9701, + "theoretical_loss": 4.518738087782093, + "tokens_seen": 157417472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048095285857572716, + "loss": 1.8863, + "theoretical_loss": 4.5184955277644985, + "tokens_seen": 157483008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004809428284854564, + "loss": 1.9597, + "theoretical_loss": 4.518253096916341, + "tokens_seen": 157548544 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048093279839518557, + "loss": 1.9003, + "theoretical_loss": 4.518010795115151, + "tokens_seen": 157614080 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048092276830491475, + "loss": 1.8889, + "theoretical_loss": 4.517768622238616, + "tokens_seen": 157679616 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048091273821464393, + "loss": 1.8274, + "theoretical_loss": 4.5175265781646, + "tokens_seen": 157745152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048090270812437317, + "loss": 1.7707, + "theoretical_loss": 4.517284662771125, + "tokens_seen": 157810688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808926780341023, + "loss": 1.9333, + "theoretical_loss": 4.5170428759363865, + "tokens_seen": 157876224 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048088264794383153, + "loss": 1.9162, + "theoretical_loss": 4.516801217538739, + "tokens_seen": 157941760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048087261785356066, + "loss": 1.8411, + "theoretical_loss": 4.516559687456708, + "tokens_seen": 158007296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808625877632899, + "loss": 1.9015, + "theoretical_loss": 4.516318285568982, + "tokens_seen": 158072832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808525576730191, + "loss": 1.8838, + "theoretical_loss": 4.516077011754414, + "tokens_seen": 158138368 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048084252758274826, + "loss": 1.9097, + "theoretical_loss": 4.51583586589202, + "tokens_seen": 158203904 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048083249749247744, + "loss": 1.9953, + "theoretical_loss": 4.515594847860985, + "tokens_seen": 158269440 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808224674022066, + "loss": 1.9107, + "theoretical_loss": 4.515353957540656, + "tokens_seen": 158334976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004808124373119358, + "loss": 1.9545, + "theoretical_loss": 4.515113194810541, + "tokens_seen": 158400512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048080240722166503, + "loss": 1.8666, + "theoretical_loss": 4.514872559550316, + "tokens_seen": 158466048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048079237713139416, + "loss": 1.8511, + "theoretical_loss": 4.514632051639815, + "tokens_seen": 158531584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807823470411234, + "loss": 1.877, + "theoretical_loss": 4.51439167095904, + "tokens_seen": 158597120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807723169508525, + "loss": 1.8384, + "theoretical_loss": 4.5141514173881525, + "tokens_seen": 158662656 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048076228686058176, + "loss": 1.8626, + "theoretical_loss": 4.513911290807478, + "tokens_seen": 158728192 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048075225677031094, + "loss": 1.9049, + "theoretical_loss": 4.513671291097502, + "tokens_seen": 158793728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807422266800401, + "loss": 1.9293, + "theoretical_loss": 4.513431418138874, + "tokens_seen": 158859264 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807321965897693, + "loss": 1.8546, + "theoretical_loss": 4.513191671812404, + "tokens_seen": 158924800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048072216649949854, + "loss": 1.859, + "theoretical_loss": 4.512952051999061, + "tokens_seen": 158990336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048071213640922766, + "loss": 1.7494, + "theoretical_loss": 4.512712558579979, + "tokens_seen": 159055872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004807021063189569, + "loss": 1.8677, + "theoretical_loss": 4.512473191436451, + "tokens_seen": 159121408 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480692076228686, + "loss": 1.8424, + "theoretical_loss": 4.512233950449927, + "tokens_seen": 159186944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048068204613841526, + "loss": 1.9319, + "theoretical_loss": 4.511994835502023, + "tokens_seen": 159252480 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048067201604814444, + "loss": 1.8491, + "theoretical_loss": 4.511755846474511, + "tokens_seen": 159318016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806619859578736, + "loss": 1.9275, + "theoretical_loss": 4.511516983249322, + "tokens_seen": 159383552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806519558676028, + "loss": 1.8146, + "theoretical_loss": 4.511278245708548, + "tokens_seen": 159449088 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480641925777332, + "loss": 1.8401, + "theoretical_loss": 4.511039633734439, + "tokens_seen": 159514624 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048063189568706117, + "loss": 1.9337, + "theoretical_loss": 4.5108011472094045, + "tokens_seen": 159580160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004806218655967904, + "loss": 1.8922, + "theoretical_loss": 4.510562786016012, + "tokens_seen": 159645696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048061183550651953, + "loss": 1.9104, + "theoretical_loss": 4.5103245500369855, + "tokens_seen": 159711232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048060180541624877, + "loss": 1.9396, + "theoretical_loss": 4.510086439155209, + "tokens_seen": 159776768 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048059177532597795, + "loss": 1.7978, + "theoretical_loss": 4.509848453253722, + "tokens_seen": 159842304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048058174523570713, + "loss": 1.8909, + "theoretical_loss": 4.509610592215726, + "tokens_seen": 159907840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805717151454363, + "loss": 1.8422, + "theoretical_loss": 4.509372855924571, + "tokens_seen": 159973376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805616850551655, + "loss": 1.729, + "theoretical_loss": 4.5091352442637715, + "tokens_seen": 160038912 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048055165496489467, + "loss": 1.8846, + "theoretical_loss": 4.508897757116995, + "tokens_seen": 160104448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805416248746239, + "loss": 1.8558, + "theoretical_loss": 4.508660394368066, + "tokens_seen": 160169984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048053159478435303, + "loss": 1.8359, + "theoretical_loss": 4.508423155900964, + "tokens_seen": 160235520 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048052156469408227, + "loss": 1.95, + "theoretical_loss": 4.508186041599823, + "tokens_seen": 160301056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004805115346038114, + "loss": 1.8616, + "theoretical_loss": 4.5079490513489375, + "tokens_seen": 160366592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048050150451354063, + "loss": 2.0668, + "theoretical_loss": 4.50771218503275, + "tokens_seen": 160432128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804914744232698, + "loss": 2.0328, + "theoretical_loss": 4.507475442535863, + "tokens_seen": 160497664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 65316, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9318039417266846, + "objective/train/theoretical_loss": 4.507238823743033, + "objective/train/tokens_used": 181023200, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480481444332999, + "loss": 1.9246, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048047141424272823, + "loss": 1.8001, + "theoretical_loss": 4.507002328539167, + "tokens_seen": 160628736 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048046138415245736, + "loss": 1.783, + "theoretical_loss": 4.506765956809331, + "tokens_seen": 160694272 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804513540621866, + "loss": 2.0128, + "theoretical_loss": 4.50652970843874, + "tokens_seen": 160759808 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048044132397191577, + "loss": 1.839, + "theoretical_loss": 4.506293583312766, + "tokens_seen": 160825344 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048043129388164495, + "loss": 1.9112, + "theoretical_loss": 4.506057581316933, + "tokens_seen": 160890880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048042126379137413, + "loss": 1.8204, + "theoretical_loss": 4.505821702336917, + "tokens_seen": 160956416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048041123370110337, + "loss": 1.9135, + "theoretical_loss": 4.505585946258547, + "tokens_seen": 161021952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004804012036108325, + "loss": 1.8858, + "theoretical_loss": 4.505350312967805, + "tokens_seen": 161087488 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048039117352056173, + "loss": 1.8375, + "theoretical_loss": 4.5051148023508265, + "tokens_seen": 161153024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048038114343029086, + "loss": 1.7301, + "theoretical_loss": 4.504879414293896, + "tokens_seen": 161218560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803711133400201, + "loss": 1.9533, + "theoretical_loss": 4.50464414868345, + "tokens_seen": 161284096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803610832497493, + "loss": 1.9314, + "theoretical_loss": 4.504409005406078, + "tokens_seen": 161349632 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048035105315947846, + "loss": 1.8819, + "theoretical_loss": 4.504173984348521, + "tokens_seen": 161415168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048034102306920764, + "loss": 1.9089, + "theoretical_loss": 4.503939085397667, + "tokens_seen": 161480704 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004803309929789368, + "loss": 1.8901, + "theoretical_loss": 4.5037043084405575, + "tokens_seen": 161546240 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480320962888666, + "loss": 1.9392, + "theoretical_loss": 4.503469653364386, + "tokens_seen": 161611776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048031093279839523, + "loss": 1.8697, + "theoretical_loss": 4.503235120056491, + "tokens_seen": 161677312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048030090270812436, + "loss": 1.9263, + "theoretical_loss": 4.503000708404365, + "tokens_seen": 161742848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802908726178536, + "loss": 1.8463, + "theoretical_loss": 4.502766418295647, + "tokens_seen": 161808384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802808425275827, + "loss": 2.0265, + "theoretical_loss": 4.502532249618129, + "tokens_seen": 161873920 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048027081243731196, + "loss": 1.9913, + "theoretical_loss": 4.502298202259748, + "tokens_seen": 161939456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048026078234704114, + "loss": 1.8076, + "theoretical_loss": 4.502064276108593, + "tokens_seen": 162004992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802507522567703, + "loss": 1.8695, + "theoretical_loss": 4.501830471052898, + "tokens_seen": 162070528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802407221664995, + "loss": 1.8321, + "theoretical_loss": 4.501596786981047, + "tokens_seen": 162136064 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048023069207622874, + "loss": 1.9024, + "theoretical_loss": 4.501363223781574, + "tokens_seen": 162201600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048022066198595786, + "loss": 1.9289, + "theoretical_loss": 4.501129781343157, + "tokens_seen": 162267136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004802106318956871, + "loss": 1.9206, + "theoretical_loss": 4.500896459554624, + "tokens_seen": 162332672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048020060180541623, + "loss": 1.8564, + "theoretical_loss": 4.500663258304948, + "tokens_seen": 162398208 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048019057171514546, + "loss": 1.8076, + "theoretical_loss": 4.500430177483253, + "tokens_seen": 162463744 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048018054162487464, + "loss": 1.9249, + "theoretical_loss": 4.500197216978804, + "tokens_seen": 162529280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801705115346038, + "loss": 1.9855, + "theoretical_loss": 4.499964376681016, + "tokens_seen": 162594816 + }, + { + "epoch": 0.05, + "learning_rate": 0.000480160481444333, + "loss": 1.9436, + "theoretical_loss": 4.49973165647945, + "tokens_seen": 162660352 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801504513540622, + "loss": 2.0195, + "theoretical_loss": 4.4994990562638115, + "tokens_seen": 162725888 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048014042126379137, + "loss": 1.776, + "theoretical_loss": 4.499266575923954, + "tokens_seen": 162791424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004801303911735206, + "loss": 2.1088, + "theoretical_loss": 4.499034215349875, + "tokens_seen": 162856960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048012036108324973, + "loss": 1.9247, + "theoretical_loss": 4.498801974431714, + "tokens_seen": 162922496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048011033099297897, + "loss": 1.8725, + "theoretical_loss": 4.498569853059762, + "tokens_seen": 162988032 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048010030090270815, + "loss": 1.872, + "theoretical_loss": 4.49833785112445, + "tokens_seen": 163053568 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048009027081243733, + "loss": 1.9337, + "theoretical_loss": 4.498105968516354, + "tokens_seen": 163119104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800802407221665, + "loss": 1.8489, + "theoretical_loss": 4.497874205126195, + "tokens_seen": 163184640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800702106318957, + "loss": 1.8684, + "theoretical_loss": 4.497642560844838, + "tokens_seen": 163250176 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048006018054162487, + "loss": 1.9655, + "theoretical_loss": 4.497411035563291, + "tokens_seen": 163315712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800501504513541, + "loss": 1.7391, + "theoretical_loss": 4.497179629172704, + "tokens_seen": 163381248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048004012036108323, + "loss": 1.9934, + "theoretical_loss": 4.4969483415643765, + "tokens_seen": 163446784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048003009027081247, + "loss": 1.8055, + "theoretical_loss": 4.496717172629742, + "tokens_seen": 163512320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004800200601805416, + "loss": 1.9352, + "theoretical_loss": 4.496486122260382, + "tokens_seen": 163577856 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048001003009027083, + "loss": 2.0043, + "theoretical_loss": 4.4962551903480215, + "tokens_seen": 163643392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00048, + "loss": 1.9224, + "theoretical_loss": 4.496024376784522, + "tokens_seen": 163708928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799899699097292, + "loss": 1.8458, + "theoretical_loss": 4.495793681461894, + "tokens_seen": 163774464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 66569, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7990031242370605, + "objective/train/theoretical_loss": 4.495563104272286, + "objective/train/tokens_used": 184300000, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799799398194584, + "loss": 1.6923, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047996990972918756, + "loss": 1.8159, + "theoretical_loss": 4.495332645107985, + "tokens_seen": 163905536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047995987963891674, + "loss": 1.9482, + "theoretical_loss": 4.495102303861426, + "tokens_seen": 163971072 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047994984954864597, + "loss": 1.9019, + "theoretical_loss": 4.4948720804251785, + "tokens_seen": 164036608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799398194583751, + "loss": 1.9404, + "theoretical_loss": 4.494641974691959, + "tokens_seen": 164102144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047992978936810433, + "loss": 1.9808, + "theoretical_loss": 4.494411986554619, + "tokens_seen": 164167680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799197592778335, + "loss": 1.9204, + "theoretical_loss": 4.494182115906153, + "tokens_seen": 164233216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004799097291875627, + "loss": 1.8584, + "theoretical_loss": 4.493952362639694, + "tokens_seen": 164298752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798996990972919, + "loss": 1.9255, + "theoretical_loss": 4.493722726648519, + "tokens_seen": 164364288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047988966900702106, + "loss": 1.8375, + "theoretical_loss": 4.493493207826037, + "tokens_seen": 164429824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047987963891675024, + "loss": 1.8695, + "theoretical_loss": 4.493263806065803, + "tokens_seen": 164495360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798696088264795, + "loss": 1.8788, + "theoretical_loss": 4.493034521261509, + "tokens_seen": 164560896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798595787362086, + "loss": 1.8567, + "theoretical_loss": 4.492805353306984, + "tokens_seen": 164626432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047984954864593784, + "loss": 1.8214, + "theoretical_loss": 4.492576302096199, + "tokens_seen": 164691968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047983951855566696, + "loss": 1.744, + "theoretical_loss": 4.49234736752326, + "tokens_seen": 164757504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798294884653962, + "loss": 1.9869, + "theoretical_loss": 4.492118549482413, + "tokens_seen": 164823040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004798194583751254, + "loss": 1.8374, + "theoretical_loss": 4.491889847868041, + "tokens_seen": 164888576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047980942828485456, + "loss": 1.7223, + "theoretical_loss": 4.491661262574665, + "tokens_seen": 164954112 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047979939819458374, + "loss": 1.8432, + "theoretical_loss": 4.491432793496945, + "tokens_seen": 165019648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797893681043129, + "loss": 1.7675, + "theoretical_loss": 4.4912044405296765, + "tokens_seen": 165085184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797793380140421, + "loss": 1.8435, + "theoretical_loss": 4.49097620356779, + "tokens_seen": 165150720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047976930792377134, + "loss": 1.8205, + "theoretical_loss": 4.490748082506356, + "tokens_seen": 165216256 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047975927783350047, + "loss": 1.8189, + "theoretical_loss": 4.490520077240582, + "tokens_seen": 165281792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797492477432297, + "loss": 1.7573, + "theoretical_loss": 4.490292187665808, + "tokens_seen": 165347328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797392176529589, + "loss": 1.8759, + "theoretical_loss": 4.490064413677514, + "tokens_seen": 165412864 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047972918756268807, + "loss": 1.8214, + "theoretical_loss": 4.4898367551713125, + "tokens_seen": 165478400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004797191574724173, + "loss": 1.7156, + "theoretical_loss": 4.489609212042954, + "tokens_seen": 165543936 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047970912738214643, + "loss": 1.9101, + "theoretical_loss": 4.489381784188322, + "tokens_seen": 165609472 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047969909729187566, + "loss": 1.8659, + "theoretical_loss": 4.489154471503438, + "tokens_seen": 165675008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047968906720160484, + "loss": 1.7125, + "theoretical_loss": 4.488927273884456, + "tokens_seen": 165740544 + }, + { + "epoch": 0.05, + "learning_rate": 0.000479679037111334, + "loss": 1.6428, + "theoretical_loss": 4.488700191227665, + "tokens_seen": 165806080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796690070210632, + "loss": 1.9008, + "theoretical_loss": 4.488473223429491, + "tokens_seen": 165871616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796589769307924, + "loss": 1.8297, + "theoretical_loss": 4.48824637038649, + "tokens_seen": 165937152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047964894684052157, + "loss": 1.8137, + "theoretical_loss": 4.488019631995354, + "tokens_seen": 166002688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004796389167502508, + "loss": 1.814, + "theoretical_loss": 4.48779300815291, + "tokens_seen": 166068224 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047962888665997993, + "loss": 1.7679, + "theoretical_loss": 4.4875664987561175, + "tokens_seen": 166133760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047961885656970917, + "loss": 1.8054, + "theoretical_loss": 4.487340103702068, + "tokens_seen": 166199296 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047960882647943835, + "loss": 1.9058, + "theoretical_loss": 4.487113822887988, + "tokens_seen": 166264832 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047959879638916753, + "loss": 1.8797, + "theoretical_loss": 4.486887656211234, + "tokens_seen": 166330368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795887662988967, + "loss": 1.7523, + "theoretical_loss": 4.486661603569301, + "tokens_seen": 166395904 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795787362086259, + "loss": 1.8053, + "theoretical_loss": 4.486435664859809, + "tokens_seen": 166461440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047956870611835507, + "loss": 1.9175, + "theoretical_loss": 4.486209839980516, + "tokens_seen": 166526976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795586760280843, + "loss": 1.8592, + "theoretical_loss": 4.485984128829307, + "tokens_seen": 166592512 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047954864593781343, + "loss": 1.8273, + "theoretical_loss": 4.485758531304205, + "tokens_seen": 166658048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047953861584754267, + "loss": 1.8967, + "theoretical_loss": 4.485533047303358, + "tokens_seen": 166723584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795285857572718, + "loss": 1.8113, + "theoretical_loss": 4.485307676725048, + "tokens_seen": 166789120 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047951855566700103, + "loss": 1.763, + "theoretical_loss": 4.485082419467691, + "tokens_seen": 166854656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004795085255767302, + "loss": 1.786, + "theoretical_loss": 4.484857275429828, + "tokens_seen": 166920192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794984954864594, + "loss": 1.6971, + "theoretical_loss": 4.484632244510136, + "tokens_seen": 166985728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794884653961886, + "loss": 1.8839, + "theoretical_loss": 4.4844073266074185, + "tokens_seen": 167051264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 67770, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9591978788375854, + "objective/train/theoretical_loss": 4.484182521620611, + "objective/train/tokens_used": 187576800, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047947843530591776, + "loss": 1.9131, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047946840521564694, + "loss": 1.8666, + "theoretical_loss": 4.48395782944878, + "tokens_seen": 167182336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047945837512537617, + "loss": 1.9626, + "theoretical_loss": 4.483733249991118, + "tokens_seen": 167247872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794483450351053, + "loss": 1.7689, + "theoretical_loss": 4.483508783146952, + "tokens_seen": 167313408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047943831494483453, + "loss": 1.7973, + "theoretical_loss": 4.483284428815735, + "tokens_seen": 167378944 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794282848545637, + "loss": 1.9353, + "theoretical_loss": 4.483060186897049, + "tokens_seen": 167444480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794182547642929, + "loss": 1.8312, + "theoretical_loss": 4.482836057290607, + "tokens_seen": 167510016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004794082246740221, + "loss": 1.8742, + "theoretical_loss": 4.4826120398962495, + "tokens_seen": 167575552 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047939819458375126, + "loss": 1.9145, + "theoretical_loss": 4.482388134613943, + "tokens_seen": 167641088 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047938816449348044, + "loss": 1.8143, + "theoretical_loss": 4.482164341343788, + "tokens_seen": 167706624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793781344032097, + "loss": 1.7316, + "theoretical_loss": 4.481940659986008, + "tokens_seen": 167772160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793681043129388, + "loss": 1.952, + "theoretical_loss": 4.481717090440956, + "tokens_seen": 167837696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047935807422266804, + "loss": 1.8433, + "theoretical_loss": 4.481493632609112, + "tokens_seen": 167903232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047934804413239716, + "loss": 1.6269, + "theoretical_loss": 4.481270286391084, + "tokens_seen": 167968768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793380140421264, + "loss": 1.8253, + "theoretical_loss": 4.481047051687609, + "tokens_seen": 168034304 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004793279839518556, + "loss": 1.8253, + "theoretical_loss": 4.480823928399547, + "tokens_seen": 168099840 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047931795386158476, + "loss": 1.8078, + "theoretical_loss": 4.480600916427887, + "tokens_seen": 168165376 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047930792377131394, + "loss": 1.7254, + "theoretical_loss": 4.480378015673745, + "tokens_seen": 168230912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792978936810431, + "loss": 1.9475, + "theoretical_loss": 4.480155226038362, + "tokens_seen": 168296448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792878635907723, + "loss": 1.8596, + "theoretical_loss": 4.479932547423107, + "tokens_seen": 168361984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047927783350050154, + "loss": 1.9109, + "theoretical_loss": 4.479709979729472, + "tokens_seen": 168427520 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047926780341023067, + "loss": 1.81, + "theoretical_loss": 4.479487522859078, + "tokens_seen": 168493056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792577733199599, + "loss": 1.931, + "theoretical_loss": 4.4792651767136675, + "tokens_seen": 168558592 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792477432296891, + "loss": 1.9458, + "theoretical_loss": 4.479042941195113, + "tokens_seen": 168624128 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047923771313941827, + "loss": 1.7493, + "theoretical_loss": 4.478820816205408, + "tokens_seen": 168689664 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047922768304914745, + "loss": 1.8716, + "theoretical_loss": 4.4785988016466725, + "tokens_seen": 168755200 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047921765295887663, + "loss": 1.8754, + "theoretical_loss": 4.4783768974211515, + "tokens_seen": 168820736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004792076228686058, + "loss": 1.6764, + "theoretical_loss": 4.478155103431215, + "tokens_seen": 168886272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047919759277833504, + "loss": 1.9117, + "theoretical_loss": 4.477933419579356, + "tokens_seen": 168951808 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047918756268806417, + "loss": 1.6752, + "theoretical_loss": 4.477711845768191, + "tokens_seen": 169017344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791775325977934, + "loss": 1.7426, + "theoretical_loss": 4.477490381900461, + "tokens_seen": 169082880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047916750250752253, + "loss": 2.0202, + "theoretical_loss": 4.477269027879032, + "tokens_seen": 169148416 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047915747241725177, + "loss": 1.7561, + "theoretical_loss": 4.477047783606892, + "tokens_seen": 169213952 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047914744232698095, + "loss": 1.8904, + "theoretical_loss": 4.476826648987151, + "tokens_seen": 169279488 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047913741223671013, + "loss": 1.8832, + "theoretical_loss": 4.476605623923044, + "tokens_seen": 169345024 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791273821464393, + "loss": 1.8444, + "theoretical_loss": 4.47638470831793, + "tokens_seen": 169410560 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047911735205616855, + "loss": 1.8899, + "theoretical_loss": 4.476163902075286, + "tokens_seen": 169476096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004791073219658977, + "loss": 1.8852, + "theoretical_loss": 4.475943205098716, + "tokens_seen": 169541632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790972918756269, + "loss": 1.9797, + "theoretical_loss": 4.475722617291944, + "tokens_seen": 169607168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047908726178535604, + "loss": 1.8633, + "theoretical_loss": 4.4755021385588165, + "tokens_seen": 169672704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047907723169508527, + "loss": 1.8087, + "theoretical_loss": 4.475281768803301, + "tokens_seen": 169738240 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047906720160481445, + "loss": 1.9166, + "theoretical_loss": 4.475061507929487, + "tokens_seen": 169803776 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047905717151454363, + "loss": 1.6505, + "theoretical_loss": 4.474841355841587, + "tokens_seen": 169869312 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790471414242728, + "loss": 1.9628, + "theoretical_loss": 4.474621312443932, + "tokens_seen": 169934848 + }, + { + "epoch": 0.05, + "learning_rate": 0.000479037111334002, + "loss": 1.8365, + "theoretical_loss": 4.474401377640977, + "tokens_seen": 170000384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790270812437312, + "loss": 1.9356, + "theoretical_loss": 4.474181551337294, + "tokens_seen": 170065920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004790170511534604, + "loss": 1.7402, + "theoretical_loss": 4.473961833437579, + "tokens_seen": 170131456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047900702106318954, + "loss": 1.8606, + "theoretical_loss": 4.473742223846646, + "tokens_seen": 170196992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789969909729188, + "loss": 1.9252, + "theoretical_loss": 4.473522722469431, + "tokens_seen": 170262528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789869608826479, + "loss": 2.01, + "theoretical_loss": 4.47330332921099, + "tokens_seen": 170328064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 69022, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.040296792984009, + "objective/train/theoretical_loss": 4.473084043976495, + "objective/train/tokens_used": 190853600, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047897693079237714, + "loss": 1.9396, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047896690070210637, + "loss": 2.0069, + "theoretical_loss": 4.472864866671245, + "tokens_seen": 170459136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789568706118355, + "loss": 1.9378, + "theoretical_loss": 4.472645797200651, + "tokens_seen": 170524672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047894684052156473, + "loss": 1.7642, + "theoretical_loss": 4.472426835470247, + "tokens_seen": 170590208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789368104312939, + "loss": 1.7892, + "theoretical_loss": 4.472207981385686, + "tokens_seen": 170655744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789267803410231, + "loss": 1.9218, + "theoretical_loss": 4.471989234852737, + "tokens_seen": 170721280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004789167502507523, + "loss": 1.8436, + "theoretical_loss": 4.471770595777292, + "tokens_seen": 170786816 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047890672016048146, + "loss": 1.8199, + "theoretical_loss": 4.4715520640653565, + "tokens_seen": 170852352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047889669007021064, + "loss": 1.8332, + "theoretical_loss": 4.471333639623059, + "tokens_seen": 170917888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788866599799399, + "loss": 1.8739, + "theoretical_loss": 4.471115322356643, + "tokens_seen": 170983424 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478876629889669, + "loss": 1.6784, + "theoretical_loss": 4.470897112172471, + "tokens_seen": 171048960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047886659979939824, + "loss": 2.04, + "theoretical_loss": 4.470679008977022, + "tokens_seen": 171114496 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047885656970912736, + "loss": 1.8184, + "theoretical_loss": 4.470461012676894, + "tokens_seen": 171180032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788465396188566, + "loss": 1.8021, + "theoretical_loss": 4.470243123178802, + "tokens_seen": 171245568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788365095285858, + "loss": 1.8595, + "theoretical_loss": 4.4700253403895776, + "tokens_seen": 171311104 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047882647943831496, + "loss": 1.9179, + "theoretical_loss": 4.469807664216168, + "tokens_seen": 171376640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047881644934804414, + "loss": 1.8009, + "theoretical_loss": 4.469590094565638, + "tokens_seen": 171442176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004788064192577733, + "loss": 1.8474, + "theoretical_loss": 4.469372631345172, + "tokens_seen": 171507712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787963891675025, + "loss": 1.7457, + "theoretical_loss": 4.469155274462064, + "tokens_seen": 171573248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047878635907723174, + "loss": 1.8383, + "theoretical_loss": 4.468938023823732, + "tokens_seen": 171638784 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047877632898696087, + "loss": 1.8343, + "theoretical_loss": 4.468720879337704, + "tokens_seen": 171704320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787662988966901, + "loss": 1.7628, + "theoretical_loss": 4.468503840911627, + "tokens_seen": 171769856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004787562688064193, + "loss": 1.8492, + "theoretical_loss": 4.46828690845326, + "tokens_seen": 171835392 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047874623871614847, + "loss": 1.8251, + "theoretical_loss": 4.468070081870483, + "tokens_seen": 171900928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047873620862587765, + "loss": 1.7482, + "theoretical_loss": 4.467853361071284, + "tokens_seen": 171966464 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047872617853560683, + "loss": 1.7824, + "theoretical_loss": 4.467636745963772, + "tokens_seen": 172032000 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478716148445336, + "loss": 1.9515, + "theoretical_loss": 4.467420236456169, + "tokens_seen": 172097536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047870611835506524, + "loss": 1.9124, + "theoretical_loss": 4.46720383245681, + "tokens_seen": 172163072 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047869608826479437, + "loss": 1.7907, + "theoretical_loss": 4.466987533874145, + "tokens_seen": 172228608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786860581745236, + "loss": 1.8609, + "theoretical_loss": 4.4667713406167415, + "tokens_seen": 172294144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047867602808425273, + "loss": 1.8033, + "theoretical_loss": 4.466555252593277, + "tokens_seen": 172359680 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047866599799398197, + "loss": 1.9225, + "theoretical_loss": 4.4663392697125435, + "tokens_seen": 172425216 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047865596790371115, + "loss": 1.7801, + "theoretical_loss": 4.466123391883448, + "tokens_seen": 172490752 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047864593781344033, + "loss": 1.9662, + "theoretical_loss": 4.4659076190150095, + "tokens_seen": 172556288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786359077231695, + "loss": 1.7843, + "theoretical_loss": 4.465691951016364, + "tokens_seen": 172621824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047862587763289875, + "loss": 1.9884, + "theoretical_loss": 4.465476387796755, + "tokens_seen": 172687360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786158475426279, + "loss": 1.8375, + "theoretical_loss": 4.465260929265543, + "tokens_seen": 172752896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004786058174523571, + "loss": 1.8607, + "theoretical_loss": 4.4650455753322005, + "tokens_seen": 172818432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047859578736208624, + "loss": 1.8359, + "theoretical_loss": 4.464830325906312, + "tokens_seen": 172883968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047858575727181547, + "loss": 1.8227, + "theoretical_loss": 4.464615180897574, + "tokens_seen": 172949504 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047857572718154465, + "loss": 1.777, + "theoretical_loss": 4.464400140215796, + "tokens_seen": 173015040 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047856569709127383, + "loss": 1.8729, + "theoretical_loss": 4.4641852037708984, + "tokens_seen": 173080576 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478555667001003, + "loss": 1.8612, + "theoretical_loss": 4.463970371472916, + "tokens_seen": 173146112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785456369107322, + "loss": 2.0096, + "theoretical_loss": 4.463755643231992, + "tokens_seen": 173211648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785356068204614, + "loss": 1.7839, + "theoretical_loss": 4.463541018958384, + "tokens_seen": 173277184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004785255767301906, + "loss": 1.9613, + "theoretical_loss": 4.463326498562459, + "tokens_seen": 173342720 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047851554663991974, + "loss": 1.7698, + "theoretical_loss": 4.463112081954696, + "tokens_seen": 173408256 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478505516549649, + "loss": 1.9606, + "theoretical_loss": 4.462897769045684, + "tokens_seen": 173473792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784954864593781, + "loss": 1.8781, + "theoretical_loss": 4.462683559746123, + "tokens_seen": 173539328 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047848545636910734, + "loss": 1.8281, + "theoretical_loss": 4.4624694539668255, + "tokens_seen": 173604864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 70257, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.975022792816162, + "objective/train/theoretical_loss": 4.462255451618713, + "objective/train/tokens_used": 194130400, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784754262788365, + "loss": 1.9307, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784653961885657, + "loss": 1.9126, + "theoretical_loss": 4.462041552612815, + "tokens_seen": 173735936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784553660982949, + "loss": 1.8725, + "theoretical_loss": 4.461827756860275, + "tokens_seen": 173801472 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784453360080241, + "loss": 1.7578, + "theoretical_loss": 4.461614064272345, + "tokens_seen": 173867008 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047843530591775324, + "loss": 1.7776, + "theoretical_loss": 4.461400474760383, + "tokens_seen": 173932544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784252758274825, + "loss": 1.8534, + "theoretical_loss": 4.461186988235864, + "tokens_seen": 173998080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004784152457372116, + "loss": 1.6522, + "theoretical_loss": 4.460973604610366, + "tokens_seen": 174063616 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047840521564694084, + "loss": 1.8021, + "theoretical_loss": 4.4607603237955775, + "tokens_seen": 174129152 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047839518555667, + "loss": 1.7859, + "theoretical_loss": 4.460547145703298, + "tokens_seen": 174194688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783851554663992, + "loss": 1.8501, + "theoretical_loss": 4.460334070245436, + "tokens_seen": 174260224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783751253761284, + "loss": 1.8837, + "theoretical_loss": 4.460121097334004, + "tokens_seen": 174325760 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047836509528585756, + "loss": 1.9276, + "theoretical_loss": 4.4599082268811285, + "tokens_seen": 174391296 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047835506519558675, + "loss": 1.7181, + "theoretical_loss": 4.459695458799041, + "tokens_seen": 174456832 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478345035105316, + "loss": 1.6872, + "theoretical_loss": 4.459482793000083, + "tokens_seen": 174522368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783350050150451, + "loss": 1.7859, + "theoretical_loss": 4.459270229396701, + "tokens_seen": 174587904 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047832497492477434, + "loss": 1.8433, + "theoretical_loss": 4.459057767901454, + "tokens_seen": 174653440 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047831494483450347, + "loss": 1.7102, + "theoretical_loss": 4.458845408427004, + "tokens_seen": 174718976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004783049147442327, + "loss": 1.7892, + "theoretical_loss": 4.458633150886124, + "tokens_seen": 174784512 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782948846539619, + "loss": 1.8798, + "theoretical_loss": 4.458420995191691, + "tokens_seen": 174850048 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047828485456369107, + "loss": 2.0131, + "theoretical_loss": 4.458208941256691, + "tokens_seen": 174915584 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047827482447342025, + "loss": 2.0061, + "theoretical_loss": 4.457996988994218, + "tokens_seen": 174981120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782647943831495, + "loss": 1.8304, + "theoretical_loss": 4.457785138317469, + "tokens_seen": 175046656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782547642928786, + "loss": 1.8332, + "theoretical_loss": 4.4575733891397515, + "tokens_seen": 175112192 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047824473420260785, + "loss": 1.7492, + "theoretical_loss": 4.457361741374477, + "tokens_seen": 175177728 + }, + { + "epoch": 0.05, + "learning_rate": 0.000478234704112337, + "loss": 1.8447, + "theoretical_loss": 4.457150194935165, + "tokens_seen": 175243264 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004782246740220662, + "loss": 1.8824, + "theoretical_loss": 4.456938749735439, + "tokens_seen": 175308800 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047821464393179544, + "loss": 1.9271, + "theoretical_loss": 4.45672740568903, + "tokens_seen": 175374336 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047820461384152457, + "loss": 1.8163, + "theoretical_loss": 4.456516162709774, + "tokens_seen": 175439872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781945837512538, + "loss": 1.8567, + "theoretical_loss": 4.456305020711613, + "tokens_seen": 175505408 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047818455366098293, + "loss": 1.8198, + "theoretical_loss": 4.456093979608594, + "tokens_seen": 175570944 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047817452357071217, + "loss": 1.8195, + "theoretical_loss": 4.455883039314869, + "tokens_seen": 175636480 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047816449348044135, + "loss": 1.9102, + "theoretical_loss": 4.455672199744695, + "tokens_seen": 175702016 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047815446339017053, + "loss": 1.7859, + "theoretical_loss": 4.455461460812435, + "tokens_seen": 175767552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781444332998997, + "loss": 1.9248, + "theoretical_loss": 4.455250822432556, + "tokens_seen": 175833088 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047813440320962895, + "loss": 1.9421, + "theoretical_loss": 4.4550402845196295, + "tokens_seen": 175898624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781243731193581, + "loss": 1.8866, + "theoretical_loss": 4.45482984698833, + "tokens_seen": 175964160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004781143430290873, + "loss": 1.8111, + "theoretical_loss": 4.454619509753439, + "tokens_seen": 176029696 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047810431293881644, + "loss": 1.7741, + "theoretical_loss": 4.45440927272984, + "tokens_seen": 176095232 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047809428284854567, + "loss": 1.8439, + "theoretical_loss": 4.454199135832521, + "tokens_seen": 176160768 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047808425275827485, + "loss": 1.8988, + "theoretical_loss": 4.4539890989765745, + "tokens_seen": 176226304 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047807422266800403, + "loss": 1.9486, + "theoretical_loss": 4.453779162077193, + "tokens_seen": 176291840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780641925777332, + "loss": 1.8072, + "theoretical_loss": 4.453569325049679, + "tokens_seen": 176357376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780541624874624, + "loss": 1.8568, + "theoretical_loss": 4.453359587809431, + "tokens_seen": 176422912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780441323971916, + "loss": 1.809, + "theoretical_loss": 4.453149950271956, + "tokens_seen": 176488448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780341023069208, + "loss": 2.0089, + "theoretical_loss": 4.452940412352861, + "tokens_seen": 176553984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047802407221664994, + "loss": 2.0084, + "theoretical_loss": 4.452730973967856, + "tokens_seen": 176619520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780140421263792, + "loss": 1.7459, + "theoretical_loss": 4.452521635032754, + "tokens_seen": 176685056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004780040120361083, + "loss": 1.9538, + "theoretical_loss": 4.452312395463473, + "tokens_seen": 176750592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047799398194583754, + "loss": 1.8654, + "theoretical_loss": 4.452103255176027, + "tokens_seen": 176816128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779839518555667, + "loss": 1.9499, + "theoretical_loss": 4.451894214086537, + "tokens_seen": 176881664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 71338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6787827014923096, + "objective/train/theoretical_loss": 4.4516852721112254, + "objective/train/tokens_used": 197407200, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779739217652959, + "loss": 1.778, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779638916750251, + "loss": 1.7961, + "theoretical_loss": 4.451476429166416, + "tokens_seen": 177012736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779538615847543, + "loss": 1.8952, + "theoretical_loss": 4.451267685168532, + "tokens_seen": 177078272 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047794383149448344, + "loss": 1.9757, + "theoretical_loss": 4.451059040034102, + "tokens_seen": 177143808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779338014042127, + "loss": 1.8228, + "theoretical_loss": 4.450850493679752, + "tokens_seen": 177209344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779237713139418, + "loss": 1.9571, + "theoretical_loss": 4.450642046022211, + "tokens_seen": 177274880 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047791374122367104, + "loss": 1.8028, + "theoretical_loss": 4.45043369697831, + "tokens_seen": 177340416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004779037111334002, + "loss": 1.7989, + "theoretical_loss": 4.450225446464978, + "tokens_seen": 177405952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778936810431294, + "loss": 1.9079, + "theoretical_loss": 4.450017294399247, + "tokens_seen": 177471488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778836509528586, + "loss": 1.8468, + "theoretical_loss": 4.449809240698249, + "tokens_seen": 177537024 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047787362086258776, + "loss": 1.7733, + "theoretical_loss": 4.449601285279215, + "tokens_seen": 177602560 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047786359077231695, + "loss": 1.7538, + "theoretical_loss": 4.449393428059478, + "tokens_seen": 177668096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778535606820462, + "loss": 1.9144, + "theoretical_loss": 4.449185668956469, + "tokens_seen": 177733632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778435305917753, + "loss": 1.8595, + "theoretical_loss": 4.448978007887721, + "tokens_seen": 177799168 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047783350050150454, + "loss": 1.699, + "theoretical_loss": 4.448770444770865, + "tokens_seen": 177864704 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047782347041123367, + "loss": 1.807, + "theoretical_loss": 4.448562979523633, + "tokens_seen": 177930240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778134403209629, + "loss": 1.8265, + "theoretical_loss": 4.448355612063855, + "tokens_seen": 177995776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004778034102306921, + "loss": 1.8642, + "theoretical_loss": 4.4481483423094605, + "tokens_seen": 178061312 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047779338014042127, + "loss": 1.8599, + "theoretical_loss": 4.447941170178478, + "tokens_seen": 178126848 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047778335005015045, + "loss": 1.8471, + "theoretical_loss": 4.447734095589036, + "tokens_seen": 178192384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777733199598797, + "loss": 1.8217, + "theoretical_loss": 4.4475271184593606, + "tokens_seen": 178257920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777632898696088, + "loss": 1.733, + "theoretical_loss": 4.447320238707777, + "tokens_seen": 178323456 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047775325977933805, + "loss": 1.9466, + "theoretical_loss": 4.447113456252707, + "tokens_seen": 178388992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777432296890672, + "loss": 1.7422, + "theoretical_loss": 4.446906771012674, + "tokens_seen": 178454528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777331995987964, + "loss": 1.6674, + "theoretical_loss": 4.446700182906298, + "tokens_seen": 178520064 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004777231695085256, + "loss": 1.9177, + "theoretical_loss": 4.446493691852297, + "tokens_seen": 178585600 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047771313941825477, + "loss": 1.7879, + "theoretical_loss": 4.4462872977694845, + "tokens_seen": 178651136 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047770310932798395, + "loss": 1.908, + "theoretical_loss": 4.4460810005767755, + "tokens_seen": 178716672 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047769307923771313, + "loss": 1.8022, + "theoretical_loss": 4.445874800193181, + "tokens_seen": 178782208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776830491474423, + "loss": 1.7249, + "theoretical_loss": 4.445668696537808, + "tokens_seen": 178847744 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047767301905717155, + "loss": 1.7876, + "theoretical_loss": 4.4454626895298635, + "tokens_seen": 178913280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776629889669007, + "loss": 1.6413, + "theoretical_loss": 4.4452567790886475, + "tokens_seen": 178978816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776529588766299, + "loss": 1.9218, + "theoretical_loss": 4.445050965133561, + "tokens_seen": 179044352 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047764292878635904, + "loss": 1.7814, + "theoretical_loss": 4.4448452475841, + "tokens_seen": 179109888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776328986960883, + "loss": 1.949, + "theoretical_loss": 4.444639626359856, + "tokens_seen": 179175424 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047762286860581746, + "loss": 1.9286, + "theoretical_loss": 4.444434101380519, + "tokens_seen": 179240960 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047761283851554664, + "loss": 1.987, + "theoretical_loss": 4.444228672565874, + "tokens_seen": 179306496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004776028084252758, + "loss": 1.9695, + "theoretical_loss": 4.444023339835803, + "tokens_seen": 179372032 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047759277833500505, + "loss": 2.0253, + "theoretical_loss": 4.443818103110283, + "tokens_seen": 179437568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775827482447342, + "loss": 1.7306, + "theoretical_loss": 4.443612962309387, + "tokens_seen": 179503104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775727181544634, + "loss": 1.8692, + "theoretical_loss": 4.4434079173532846, + "tokens_seen": 179568640 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047756268806419254, + "loss": 1.8708, + "theoretical_loss": 4.4432029681622405, + "tokens_seen": 179634176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775526579739218, + "loss": 1.8362, + "theoretical_loss": 4.442998114656614, + "tokens_seen": 179699712 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047754262788365096, + "loss": 1.9233, + "theoretical_loss": 4.442793356756861, + "tokens_seen": 179765248 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047753259779338014, + "loss": 1.8271, + "theoretical_loss": 4.442588694383532, + "tokens_seen": 179830784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775225677031093, + "loss": 1.7673, + "theoretical_loss": 4.44238412745727, + "tokens_seen": 179896320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775125376128385, + "loss": 1.8997, + "theoretical_loss": 4.442179655898819, + "tokens_seen": 179961856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004775025075225677, + "loss": 1.8708, + "theoretical_loss": 4.44197527962901, + "tokens_seen": 180027392 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774924774322969, + "loss": 1.9149, + "theoretical_loss": 4.441770998568773, + "tokens_seen": 180092928 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047748244734202605, + "loss": 1.9494, + "theoretical_loss": 4.441566812639133, + "tokens_seen": 180158464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 72550, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7831600904464722, + "objective/train/theoretical_loss": 4.441362721761207, + "objective/train/tokens_used": 200684000, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774724172517553, + "loss": 1.8553, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774623871614845, + "loss": 2.0006, + "theoretical_loss": 4.441158725856207, + "tokens_seen": 180289536 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047745235707121364, + "loss": 1.9894, + "theoretical_loss": 4.440954824845438, + "tokens_seen": 180355072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774423269809429, + "loss": 1.9667, + "theoretical_loss": 4.4407510186503, + "tokens_seen": 180420608 + }, + { + "epoch": 0.05, + "learning_rate": 0.000477432296890672, + "loss": 1.8333, + "theoretical_loss": 4.440547307192286, + "tokens_seen": 180486144 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047742226680040124, + "loss": 1.7217, + "theoretical_loss": 4.440343690392984, + "tokens_seen": 180551680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774122367101304, + "loss": 1.8191, + "theoretical_loss": 4.4401401681740715, + "tokens_seen": 180617216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004774022066198596, + "loss": 1.8367, + "theoretical_loss": 4.439936740457323, + "tokens_seen": 180682752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773921765295888, + "loss": 1.934, + "theoretical_loss": 4.439733407164605, + "tokens_seen": 180748288 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047738214643931797, + "loss": 1.8859, + "theoretical_loss": 4.439530168217875, + "tokens_seen": 180813824 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047737211634904715, + "loss": 1.7101, + "theoretical_loss": 4.439327023539187, + "tokens_seen": 180879360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773620862587764, + "loss": 1.8606, + "theoretical_loss": 4.439123973050684, + "tokens_seen": 180944896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773520561685055, + "loss": 1.9118, + "theoretical_loss": 4.438921016674603, + "tokens_seen": 181010432 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047734202607823474, + "loss": 1.7674, + "theoretical_loss": 4.438718154333274, + "tokens_seen": 181075968 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047733199598796387, + "loss": 1.8812, + "theoretical_loss": 4.438515385949118, + "tokens_seen": 181141504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773219658976931, + "loss": 1.8164, + "theoretical_loss": 4.438312711444647, + "tokens_seen": 181207040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004773119358074223, + "loss": 1.9285, + "theoretical_loss": 4.438110130742469, + "tokens_seen": 181272576 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047730190571715147, + "loss": 1.8633, + "theoretical_loss": 4.437907643765281, + "tokens_seen": 181338112 + }, + { + "epoch": 0.05, + "learning_rate": 0.00047729187562688065, + "loss": 1.8459, + "theoretical_loss": 4.43770525043587, + "tokens_seen": 181403648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0004772818455366099, + "loss": 2.0334, + "theoretical_loss": 4.437502950677116, + "tokens_seen": 181469184 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477271815446339, + "loss": 1.8676, + "theoretical_loss": 4.437300744411992, + "tokens_seen": 181534720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047726178535606825, + "loss": 1.8772, + "theoretical_loss": 4.43709863156356, + "tokens_seen": 181600256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772517552657974, + "loss": 1.8032, + "theoretical_loss": 4.436896612054975, + "tokens_seen": 181665792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772417251755266, + "loss": 1.813, + "theoretical_loss": 4.43669468580948, + "tokens_seen": 181731328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004772316950852558, + "loss": 1.9436, + "theoretical_loss": 4.43649285275041, + "tokens_seen": 181796864 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047722166499498497, + "loss": 1.7658, + "theoretical_loss": 4.436291112801193, + "tokens_seen": 181862400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047721163490471415, + "loss": 1.9259, + "theoretical_loss": 4.436089465885345, + "tokens_seen": 181927936 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047720160481444333, + "loss": 1.8376, + "theoretical_loss": 4.435887911926472, + "tokens_seen": 181993472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771915747241725, + "loss": 1.8543, + "theoretical_loss": 4.435686450848273, + "tokens_seen": 182059008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047718154463390175, + "loss": 1.8671, + "theoretical_loss": 4.435485082574533, + "tokens_seen": 182124544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771715145436309, + "loss": 1.8932, + "theoretical_loss": 4.43528380702913, + "tokens_seen": 182190080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771614844533601, + "loss": 1.8102, + "theoretical_loss": 4.435082624136031, + "tokens_seen": 182255616 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047715145436308924, + "loss": 1.9554, + "theoretical_loss": 4.4348815338192935, + "tokens_seen": 182321152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004771414242728185, + "loss": 1.7864, + "theoretical_loss": 4.434680536003062, + "tokens_seen": 182386688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047713139418254766, + "loss": 1.8149, + "theoretical_loss": 4.434479630611573, + "tokens_seen": 182452224 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047712136409227684, + "loss": 1.9381, + "theoretical_loss": 4.43427881756915, + "tokens_seen": 182517760 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477111334002006, + "loss": 1.8684, + "theoretical_loss": 4.434078096800209, + "tokens_seen": 182583296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047710130391173525, + "loss": 1.8959, + "theoretical_loss": 4.433877468229251, + "tokens_seen": 182648832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770912738214644, + "loss": 1.8518, + "theoretical_loss": 4.433676931780868, + "tokens_seen": 182714368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770812437311936, + "loss": 1.98, + "theoretical_loss": 4.43347648737974, + "tokens_seen": 182779904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047707121364092274, + "loss": 1.9876, + "theoretical_loss": 4.433276134950637, + "tokens_seen": 182845440 + }, + { + "epoch": 0.06, + "learning_rate": 0.000477061183550652, + "loss": 1.7756, + "theoretical_loss": 4.4330758744184156, + "tokens_seen": 182910976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047705115346038116, + "loss": 1.7251, + "theoretical_loss": 4.432875705708021, + "tokens_seen": 182976512 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047704112337011034, + "loss": 1.8671, + "theoretical_loss": 4.432675628744488, + "tokens_seen": 183042048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770310932798395, + "loss": 1.8938, + "theoretical_loss": 4.432475643452937, + "tokens_seen": 183107584 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770210631895687, + "loss": 2.0295, + "theoretical_loss": 4.432275749758579, + "tokens_seen": 183173120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770110330992979, + "loss": 1.8994, + "theoretical_loss": 4.432075947586711, + "tokens_seen": 183238656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004770010030090271, + "loss": 1.7694, + "theoretical_loss": 4.431876236862717, + "tokens_seen": 183304192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047699097291875625, + "loss": 1.7542, + "theoretical_loss": 4.43167661751207, + "tokens_seen": 183369728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769809428284855, + "loss": 1.7995, + "theoretical_loss": 4.431477089460331, + "tokens_seen": 183435264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 73723, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.927101492881775, + "objective/train/theoretical_loss": 4.431277652633145, + "objective/train/tokens_used": 203960800, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047697091273821466, + "loss": 1.8518, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047696088264794384, + "loss": 1.9039, + "theoretical_loss": 4.431078306956248, + "tokens_seen": 183566336 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476950852557673, + "loss": 1.8704, + "theoretical_loss": 4.43087905235546, + "tokens_seen": 183631872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769408224674022, + "loss": 1.8034, + "theoretical_loss": 4.43067988875669, + "tokens_seen": 183697408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769307923771314, + "loss": 1.837, + "theoretical_loss": 4.430480816085931, + "tokens_seen": 183762944 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004769207622868606, + "loss": 1.8937, + "theoretical_loss": 4.430281834269265, + "tokens_seen": 183828480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047691073219658975, + "loss": 1.7199, + "theoretical_loss": 4.4300829432328594, + "tokens_seen": 183894016 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476900702106319, + "loss": 1.8711, + "theoretical_loss": 4.429884142902968, + "tokens_seen": 183959552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768906720160481, + "loss": 1.8456, + "theoretical_loss": 4.429685433205931, + "tokens_seen": 184025088 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047688064192577735, + "loss": 1.8767, + "theoretical_loss": 4.429486814068174, + "tokens_seen": 184090624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047687061183550653, + "loss": 1.8218, + "theoretical_loss": 4.4292882854162094, + "tokens_seen": 184156160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768605817452357, + "loss": 1.8582, + "theoretical_loss": 4.429089847176634, + "tokens_seen": 184221696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768505516549649, + "loss": 1.899, + "theoretical_loss": 4.428891499276132, + "tokens_seen": 184287232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047684052156469407, + "loss": 1.9655, + "theoretical_loss": 4.42869324164147, + "tokens_seen": 184352768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047683049147442325, + "loss": 1.8517, + "theoretical_loss": 4.428495074199505, + "tokens_seen": 184418304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768204613841525, + "loss": 1.8624, + "theoretical_loss": 4.428296996877176, + "tokens_seen": 184483840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004768104312938816, + "loss": 1.8707, + "theoretical_loss": 4.428099009601506, + "tokens_seen": 184549376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047680040120361085, + "loss": 1.8307, + "theoretical_loss": 4.4279011122996055, + "tokens_seen": 184614912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047679037111334003, + "loss": 1.9586, + "theoretical_loss": 4.427703304898668, + "tokens_seen": 184680448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767803410230692, + "loss": 1.7943, + "theoretical_loss": 4.427505587325974, + "tokens_seen": 184745984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767703109327984, + "loss": 1.8007, + "theoretical_loss": 4.427307959508885, + "tokens_seen": 184811520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767602808425276, + "loss": 1.8016, + "theoretical_loss": 4.4271104213748504, + "tokens_seen": 184877056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047675025075225676, + "loss": 1.7987, + "theoretical_loss": 4.426912972851404, + "tokens_seen": 184942592 + }, + { + "epoch": 0.06, + "learning_rate": 0.000476740220661986, + "loss": 1.8022, + "theoretical_loss": 4.426715613866159, + "tokens_seen": 185008128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767301905717151, + "loss": 1.9601, + "theoretical_loss": 4.426518344346819, + "tokens_seen": 185073664 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047672016048144435, + "loss": 1.987, + "theoretical_loss": 4.426321164221168, + "tokens_seen": 185139200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047671013039117353, + "loss": 1.9213, + "theoretical_loss": 4.426124073417074, + "tokens_seen": 185204736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004767001003009027, + "loss": 1.7978, + "theoretical_loss": 4.42592707186249, + "tokens_seen": 185270272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047669007021063195, + "loss": 1.8172, + "theoretical_loss": 4.42573015948545, + "tokens_seen": 185335808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766800401203611, + "loss": 1.9843, + "theoretical_loss": 4.425533336214076, + "tokens_seen": 185401344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766700100300903, + "loss": 1.8588, + "theoretical_loss": 4.425336601976568, + "tokens_seen": 185466880 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047665997993981944, + "loss": 1.9467, + "theoretical_loss": 4.425139956701214, + "tokens_seen": 185532416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766499498495487, + "loss": 1.8152, + "theoretical_loss": 4.42494340031638, + "tokens_seen": 185597952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047663991975927786, + "loss": 1.943, + "theoretical_loss": 4.424746932750521, + "tokens_seen": 185663488 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047662988966900704, + "loss": 1.851, + "theoretical_loss": 4.4245505539321694, + "tokens_seen": 185729024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004766198595787362, + "loss": 1.8413, + "theoretical_loss": 4.424354263789943, + "tokens_seen": 185794560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047660982948846545, + "loss": 1.8398, + "theoretical_loss": 4.424158062252541, + "tokens_seen": 185860096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765997993981946, + "loss": 1.8699, + "theoretical_loss": 4.423961949248748, + "tokens_seen": 185925632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765897693079238, + "loss": 1.7918, + "theoretical_loss": 4.423765924707427, + "tokens_seen": 185991168 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047657973921765294, + "loss": 1.8338, + "theoretical_loss": 4.4235699885575235, + "tokens_seen": 186056704 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765697091273822, + "loss": 1.8126, + "theoretical_loss": 4.423374140728068, + "tokens_seen": 186122240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047655967903711136, + "loss": 1.7518, + "theoretical_loss": 4.423178381148172, + "tokens_seen": 186187776 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047654964894684054, + "loss": 1.8235, + "theoretical_loss": 4.422982709747026, + "tokens_seen": 186253312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765396188565697, + "loss": 1.7932, + "theoretical_loss": 4.4227871264539065, + "tokens_seen": 186318848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765295887662989, + "loss": 1.7827, + "theoretical_loss": 4.422591631198167, + "tokens_seen": 186384384 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765195586760281, + "loss": 1.7481, + "theoretical_loss": 4.422396223909248, + "tokens_seen": 186449920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004765095285857573, + "loss": 1.7649, + "theoretical_loss": 4.422200904516666, + "tokens_seen": 186515456 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047649949849548645, + "loss": 1.7817, + "theoretical_loss": 4.4220056729500214, + "tokens_seen": 186580992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764894684052157, + "loss": 1.8613, + "theoretical_loss": 4.4218105291389955, + "tokens_seen": 186646528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047647943831494486, + "loss": 1.8993, + "theoretical_loss": 4.4216154730133495, + "tokens_seen": 186712064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 74904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8248214721679688, + "objective/train/theoretical_loss": 4.421420504502928, + "objective/train/tokens_used": 207237600, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047646940822467404, + "loss": 1.802, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764593781344032, + "loss": 1.9432, + "theoretical_loss": 4.421225623537652, + "tokens_seen": 186843136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764493480441324, + "loss": 1.9359, + "theoretical_loss": 4.421030830047528, + "tokens_seen": 186908672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764393179538616, + "loss": 1.978, + "theoretical_loss": 4.420836123962639, + "tokens_seen": 186974208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764292878635908, + "loss": 1.7742, + "theoretical_loss": 4.420641505213151, + "tokens_seen": 187039744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047641925777331995, + "loss": 1.7903, + "theoretical_loss": 4.42044697372931, + "tokens_seen": 187105280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004764092276830492, + "loss": 1.8156, + "theoretical_loss": 4.420252529441439, + "tokens_seen": 187170816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763991975927783, + "loss": 1.9389, + "theoretical_loss": 4.420058172279945, + "tokens_seen": 187236352 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047638916750250755, + "loss": 1.8771, + "theoretical_loss": 4.419863902175313, + "tokens_seen": 187301888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047637913741223673, + "loss": 1.8732, + "theoretical_loss": 4.419669719058108, + "tokens_seen": 187367424 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763691073219659, + "loss": 1.9389, + "theoretical_loss": 4.419475622858974, + "tokens_seen": 187432960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763590772316951, + "loss": 1.8891, + "theoretical_loss": 4.419281613508637, + "tokens_seen": 187498496 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047634904714142427, + "loss": 1.9643, + "theoretical_loss": 4.419087690937898, + "tokens_seen": 187564032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047633901705115345, + "loss": 1.9345, + "theoretical_loss": 4.418893855077641, + "tokens_seen": 187629568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763289869608827, + "loss": 1.8425, + "theoretical_loss": 4.4187001058588296, + "tokens_seen": 187695104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004763189568706118, + "loss": 1.8999, + "theoretical_loss": 4.418506443212503, + "tokens_seen": 187760640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047630892678034105, + "loss": 1.8876, + "theoretical_loss": 4.4183128670697815, + "tokens_seen": 187826176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047629889669007023, + "loss": 1.9531, + "theoretical_loss": 4.418119377361863, + "tokens_seen": 187891712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762888665997994, + "loss": 1.8538, + "theoretical_loss": 4.417925974020028, + "tokens_seen": 187957248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762788365095286, + "loss": 1.8619, + "theoretical_loss": 4.4177326569756294, + "tokens_seen": 188022784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762688064192578, + "loss": 1.8976, + "theoretical_loss": 4.417539426160104, + "tokens_seen": 188088320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047625877632898696, + "loss": 1.8385, + "theoretical_loss": 4.417346281504963, + "tokens_seen": 188153856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762487462387162, + "loss": 1.8781, + "theoretical_loss": 4.417153222941798, + "tokens_seen": 188219392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762387161484453, + "loss": 1.9133, + "theoretical_loss": 4.416960250402279, + "tokens_seen": 188284928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047622868605817455, + "loss": 1.8918, + "theoretical_loss": 4.416767363818153, + "tokens_seen": 188350464 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762186559679037, + "loss": 1.7543, + "theoretical_loss": 4.4165745631212445, + "tokens_seen": 188416000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004762086258776329, + "loss": 1.8484, + "theoretical_loss": 4.416381848243456, + "tokens_seen": 188481536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761985957873621, + "loss": 1.7877, + "theoretical_loss": 4.41618921911677, + "tokens_seen": 188547072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761885656970913, + "loss": 1.9759, + "theoretical_loss": 4.415996675673242, + "tokens_seen": 188612608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047617853560682046, + "loss": 1.9731, + "theoretical_loss": 4.415804217845009, + "tokens_seen": 188678144 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047616850551654964, + "loss": 2.0283, + "theoretical_loss": 4.415611845564284, + "tokens_seen": 188743680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761584754262788, + "loss": 1.8343, + "theoretical_loss": 4.415419558763356, + "tokens_seen": 188809216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047614844533600806, + "loss": 1.9442, + "theoretical_loss": 4.415227357374592, + "tokens_seen": 188874752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761384152457372, + "loss": 1.8583, + "theoretical_loss": 4.415035241330436, + "tokens_seen": 188940288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761283851554664, + "loss": 1.868, + "theoretical_loss": 4.41484321056341, + "tokens_seen": 189005824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761183550651956, + "loss": 1.8842, + "theoretical_loss": 4.4146512650061105, + "tokens_seen": 189071360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004761083249749248, + "loss": 1.8975, + "theoretical_loss": 4.414459404591211, + "tokens_seen": 189136896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047609829488465396, + "loss": 1.7294, + "theoretical_loss": 4.414267629251463, + "tokens_seen": 189202432 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047608826479438314, + "loss": 1.7811, + "theoretical_loss": 4.414075938919694, + "tokens_seen": 189267968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760782347041123, + "loss": 1.8996, + "theoretical_loss": 4.413884333528807, + "tokens_seen": 189333504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047606820461384156, + "loss": 1.8248, + "theoretical_loss": 4.413692813011781, + "tokens_seen": 189399040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760581745235707, + "loss": 1.8051, + "theoretical_loss": 4.413501377301671, + "tokens_seen": 189464576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760481444332999, + "loss": 1.7678, + "theoretical_loss": 4.41331002633161, + "tokens_seen": 189530112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047603811434302905, + "loss": 1.9277, + "theoretical_loss": 4.413118760034804, + "tokens_seen": 189595648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004760280842527583, + "loss": 1.8622, + "theoretical_loss": 4.4129275783445365, + "tokens_seen": 189661184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047601805416248746, + "loss": 1.8344, + "theoretical_loss": 4.412736481194166, + "tokens_seen": 189726720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047600802407221665, + "loss": 1.8603, + "theoretical_loss": 4.412545468517127, + "tokens_seen": 189792256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759979939819458, + "loss": 1.8766, + "theoretical_loss": 4.412354540246929, + "tokens_seen": 189857792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047598796389167506, + "loss": 1.7326, + "theoretical_loss": 4.4121636963171555, + "tokens_seen": 189923328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759779338014042, + "loss": 1.9364, + "theoretical_loss": 4.411972936661467, + "tokens_seen": 189988864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 76077, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8357908725738525, + "objective/train/theoretical_loss": 4.411782261213599, + "objective/train/tokens_used": 210514400, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759679037111334, + "loss": 1.8987, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759578736208626, + "loss": 1.8517, + "theoretical_loss": 4.41159166990736, + "tokens_seen": 190119936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759478435305918, + "loss": 1.9031, + "theoretical_loss": 4.411401162676636, + "tokens_seen": 190185472 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475937813440321, + "loss": 2.0048, + "theoretical_loss": 4.411210739455386, + "tokens_seen": 190251008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047592778335005015, + "loss": 1.7991, + "theoretical_loss": 4.4110204001776445, + "tokens_seen": 190316544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759177532597794, + "loss": 1.6843, + "theoretical_loss": 4.410830144777519, + "tokens_seen": 190382080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004759077231695085, + "loss": 1.7438, + "theoretical_loss": 4.410639973189192, + "tokens_seen": 190447616 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047589769307923775, + "loss": 1.8337, + "theoretical_loss": 4.410449885346922, + "tokens_seen": 190513152 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047588766298896693, + "loss": 1.9578, + "theoretical_loss": 4.4102598811850395, + "tokens_seen": 190578688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758776328986961, + "loss": 1.9119, + "theoretical_loss": 4.410069960637949, + "tokens_seen": 190644224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758676028084253, + "loss": 1.9047, + "theoretical_loss": 4.409880123640131, + "tokens_seen": 190709760 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047585757271815447, + "loss": 1.8414, + "theoretical_loss": 4.409690370126139, + "tokens_seen": 190775296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047584754262788365, + "loss": 1.9169, + "theoretical_loss": 4.409500700030598, + "tokens_seen": 190840832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004758375125376129, + "loss": 1.9468, + "theoretical_loss": 4.409311113288211, + "tokens_seen": 190906368 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475827482447342, + "loss": 1.8748, + "theoretical_loss": 4.409121609833749, + "tokens_seen": 190971904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047581745235707125, + "loss": 1.7513, + "theoretical_loss": 4.408932189602061, + "tokens_seen": 191037440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047580742226680043, + "loss": 1.9248, + "theoretical_loss": 4.408742852528068, + "tokens_seen": 191102976 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757973921765296, + "loss": 1.7699, + "theoretical_loss": 4.408553598546764, + "tokens_seen": 191168512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757873620862588, + "loss": 1.8203, + "theoretical_loss": 4.408364427593215, + "tokens_seen": 191234048 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475777331995988, + "loss": 1.9246, + "theoretical_loss": 4.40817533960256, + "tokens_seen": 191299584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047576730190571716, + "loss": 1.783, + "theoretical_loss": 4.407986334510014, + "tokens_seen": 191365120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757572718154464, + "loss": 2.0247, + "theoretical_loss": 4.407797412250861, + "tokens_seen": 191430656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757472417251755, + "loss": 1.8656, + "theoretical_loss": 4.40760857276046, + "tokens_seen": 191496192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047573721163490475, + "loss": 1.7372, + "theoretical_loss": 4.407419815974242, + "tokens_seen": 191561728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757271815446339, + "loss": 1.8682, + "theoretical_loss": 4.407231141827708, + "tokens_seen": 191627264 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757171514543631, + "loss": 1.9003, + "theoretical_loss": 4.407042550256436, + "tokens_seen": 191692800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004757071213640923, + "loss": 1.8885, + "theoretical_loss": 4.406854041196072, + "tokens_seen": 191758336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756970912738215, + "loss": 1.7701, + "theoretical_loss": 4.406665614582338, + "tokens_seen": 191823872 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047568706118355066, + "loss": 1.8405, + "theoretical_loss": 4.406477270351024, + "tokens_seen": 191889408 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047567703109327984, + "loss": 1.832, + "theoretical_loss": 4.4062890084379935, + "tokens_seen": 191954944 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475667001003009, + "loss": 1.7279, + "theoretical_loss": 4.406100828779183, + "tokens_seen": 192020480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047565697091273826, + "loss": 1.7257, + "theoretical_loss": 4.4059127313106, + "tokens_seen": 192086016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756469408224674, + "loss": 1.8353, + "theoretical_loss": 4.405724715968323, + "tokens_seen": 192151552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756369107321966, + "loss": 1.8569, + "theoretical_loss": 4.405536782688502, + "tokens_seen": 192217088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004756268806419258, + "loss": 1.9061, + "theoretical_loss": 4.405348931407359, + "tokens_seen": 192282624 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475616850551655, + "loss": 1.8508, + "theoretical_loss": 4.405161162061187, + "tokens_seen": 192348160 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047560682046138416, + "loss": 1.7857, + "theoretical_loss": 4.404973474586351, + "tokens_seen": 192413696 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047559679037111334, + "loss": 1.9346, + "theoretical_loss": 4.404785868919285, + "tokens_seen": 192479232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755867602808425, + "loss": 1.7048, + "theoretical_loss": 4.404598344996497, + "tokens_seen": 192544768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047557673019057176, + "loss": 1.7385, + "theoretical_loss": 4.404410902754562, + "tokens_seen": 192610304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755667001003009, + "loss": 1.8177, + "theoretical_loss": 4.404223542130129, + "tokens_seen": 192675840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755566700100301, + "loss": 1.8056, + "theoretical_loss": 4.404036263059917, + "tokens_seen": 192741376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047554663991975925, + "loss": 1.8134, + "theoretical_loss": 4.403849065480714, + "tokens_seen": 192806912 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004755366098294885, + "loss": 1.9289, + "theoretical_loss": 4.403661949329381, + "tokens_seen": 192872448 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047552657973921766, + "loss": 1.9265, + "theoretical_loss": 4.403474914542848, + "tokens_seen": 192937984 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047551654964894685, + "loss": 1.8414, + "theoretical_loss": 4.403287961058112, + "tokens_seen": 193003520 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047550651955867603, + "loss": 1.7267, + "theoretical_loss": 4.403101088812248, + "tokens_seen": 193069056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047549648946840526, + "loss": 1.8826, + "theoretical_loss": 4.4029142977423925, + "tokens_seen": 193134592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754864593781344, + "loss": 1.7844, + "theoretical_loss": 4.402727587785758, + "tokens_seen": 193200128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754764292878636, + "loss": 1.8465, + "theoretical_loss": 4.402540958879625, + "tokens_seen": 193265664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 77266, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.906221866607666, + "objective/train/theoretical_loss": 4.402354410961341, + "objective/train/tokens_used": 213791200, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047546639919759275, + "loss": 1.8333, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475456369107322, + "loss": 1.762, + "theoretical_loss": 4.402167943968328, + "tokens_seen": 193396736 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047544633901705117, + "loss": 1.75, + "theoretical_loss": 4.4019815578380745, + "tokens_seen": 193462272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047543630892678035, + "loss": 1.7785, + "theoretical_loss": 4.401795252508139, + "tokens_seen": 193527808 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047542627883650953, + "loss": 1.7384, + "theoretical_loss": 4.401609027916149, + "tokens_seen": 193593344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754162487462387, + "loss": 1.6754, + "theoretical_loss": 4.4014228839998015, + "tokens_seen": 193658880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004754062186559679, + "loss": 1.8201, + "theoretical_loss": 4.401236820696863, + "tokens_seen": 193724416 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047539618856569713, + "loss": 1.778, + "theoretical_loss": 4.40105083794517, + "tokens_seen": 193789952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047538615847542625, + "loss": 1.7556, + "theoretical_loss": 4.400864935682625, + "tokens_seen": 193855488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753761283851555, + "loss": 1.79, + "theoretical_loss": 4.400679113847202, + "tokens_seen": 193921024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753660982948846, + "loss": 1.9875, + "theoretical_loss": 4.400493372376943, + "tokens_seen": 193986560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047535606820461385, + "loss": 1.7805, + "theoretical_loss": 4.400307711209957, + "tokens_seen": 194052096 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047534603811434303, + "loss": 1.848, + "theoretical_loss": 4.400122130284426, + "tokens_seen": 194117632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753360080240722, + "loss": 1.7652, + "theoretical_loss": 4.399936629538595, + "tokens_seen": 194183168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004753259779338014, + "loss": 1.648, + "theoretical_loss": 4.3997512089107795, + "tokens_seen": 194248704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047531594784353063, + "loss": 1.7608, + "theoretical_loss": 4.399565868339366, + "tokens_seen": 194314240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047530591775325976, + "loss": 1.8503, + "theoretical_loss": 4.399380607762804, + "tokens_seen": 194379776 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475295887662989, + "loss": 1.7991, + "theoretical_loss": 4.399195427119616, + "tokens_seen": 194445312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752858575727181, + "loss": 1.8693, + "theoretical_loss": 4.399010326348388, + "tokens_seen": 194510848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047527582748244736, + "loss": 1.8385, + "theoretical_loss": 4.398825305387778, + "tokens_seen": 194576384 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047526579739217654, + "loss": 1.8962, + "theoretical_loss": 4.39864036417651, + "tokens_seen": 194641920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752557673019057, + "loss": 1.7977, + "theoretical_loss": 4.398455502653373, + "tokens_seen": 194707456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752457372116349, + "loss": 1.911, + "theoretical_loss": 4.398270720757228, + "tokens_seen": 194772992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752357071213641, + "loss": 1.9495, + "theoretical_loss": 4.398086018427001, + "tokens_seen": 194838528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047522567703109326, + "loss": 1.8285, + "theoretical_loss": 4.397901395601687, + "tokens_seen": 194904064 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752156469408225, + "loss": 1.913, + "theoretical_loss": 4.397716852220345, + "tokens_seen": 194969600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004752056168505517, + "loss": 1.9054, + "theoretical_loss": 4.397532388222105, + "tokens_seen": 195035136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047519558676028086, + "loss": 1.8772, + "theoretical_loss": 4.397348003546161, + "tokens_seen": 195100672 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047518555667001004, + "loss": 1.7698, + "theoretical_loss": 4.397163698131776, + "tokens_seen": 195166208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751755265797392, + "loss": 1.7344, + "theoretical_loss": 4.39697947191828, + "tokens_seen": 195231744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047516549648946846, + "loss": 1.8607, + "theoretical_loss": 4.396795324845069, + "tokens_seen": 195297280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751554663991976, + "loss": 1.7986, + "theoretical_loss": 4.396611256851604, + "tokens_seen": 195362816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751454363089268, + "loss": 1.7518, + "theoretical_loss": 4.3964272678774154, + "tokens_seen": 195428352 + }, + { + "epoch": 0.06, + "learning_rate": 0.000475135406218656, + "loss": 1.7518, + "theoretical_loss": 4.396243357862099, + "tokens_seen": 195493888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004751253761283852, + "loss": 1.833, + "theoretical_loss": 4.396059526745318, + "tokens_seen": 195559424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047511534603811436, + "loss": 1.801, + "theoretical_loss": 4.395875774466799, + "tokens_seen": 195624960 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047510531594784354, + "loss": 1.8172, + "theoretical_loss": 4.395692100966338, + "tokens_seen": 195690496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750952858575727, + "loss": 1.7004, + "theoretical_loss": 4.395508506183796, + "tokens_seen": 195756032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047508525576730196, + "loss": 1.9864, + "theoretical_loss": 4.395324990059098, + "tokens_seen": 195821568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750752256770311, + "loss": 1.7878, + "theoretical_loss": 4.395141552532239, + "tokens_seen": 195887104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750651955867603, + "loss": 1.9065, + "theoretical_loss": 4.394958193543277, + "tokens_seen": 195952640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047505516549648945, + "loss": 1.8361, + "theoretical_loss": 4.394774913032337, + "tokens_seen": 196018176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004750451354062187, + "loss": 1.7508, + "theoretical_loss": 4.394591710939608, + "tokens_seen": 196083712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047503510531594787, + "loss": 1.73, + "theoretical_loss": 4.394408587205344, + "tokens_seen": 196149248 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047502507522567705, + "loss": 1.8403, + "theoretical_loss": 4.39422554176987, + "tokens_seen": 196214784 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047501504513540623, + "loss": 1.7099, + "theoretical_loss": 4.39404257457357, + "tokens_seen": 196280320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047500501504513546, + "loss": 1.6901, + "theoretical_loss": 4.393859685556895, + "tokens_seen": 196345856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749949849548646, + "loss": 1.8942, + "theoretical_loss": 4.393676874660364, + "tokens_seen": 196411392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749849548645938, + "loss": 1.7573, + "theoretical_loss": 4.393494141824557, + "tokens_seen": 196476928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047497492477432295, + "loss": 1.8021, + "theoretical_loss": 4.393311486990122, + "tokens_seen": 196542464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 78313, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6117005348205566, + "objective/train/theoretical_loss": 4.39312891009777, + "objective/train/tokens_used": 217068000, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749648946840522, + "loss": 1.7135, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047495486459378137, + "loss": 1.8009, + "theoretical_loss": 4.392946411088279, + "tokens_seen": 196673536 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047494483450351055, + "loss": 1.86, + "theoretical_loss": 4.392763989902489, + "tokens_seen": 196739072 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047493480441323973, + "loss": 1.7829, + "theoretical_loss": 4.3925816464813074, + "tokens_seen": 196804608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749247743229689, + "loss": 1.8078, + "theoretical_loss": 4.392399380765704, + "tokens_seen": 196870144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004749147442326981, + "loss": 1.8044, + "theoretical_loss": 4.392217192696712, + "tokens_seen": 196935680 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047490471414242733, + "loss": 1.9192, + "theoretical_loss": 4.392035082215434, + "tokens_seen": 197001216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047489468405215646, + "loss": 1.7652, + "theoretical_loss": 4.391853049263032, + "tokens_seen": 197066752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748846539618857, + "loss": 1.7871, + "theoretical_loss": 4.391671093780733, + "tokens_seen": 197132288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748746238716148, + "loss": 1.6972, + "theoretical_loss": 4.39148921570983, + "tokens_seen": 197197824 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047486459378134405, + "loss": 1.6219, + "theoretical_loss": 4.391307414991678, + "tokens_seen": 197263360 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047485456369107323, + "loss": 1.8762, + "theoretical_loss": 4.391125691567697, + "tokens_seen": 197328896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748445336008024, + "loss": 1.8685, + "theoretical_loss": 4.390944045379371, + "tokens_seen": 197394432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748345035105316, + "loss": 1.8267, + "theoretical_loss": 4.390762476368245, + "tokens_seen": 197459968 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047482447342026083, + "loss": 1.8007, + "theoretical_loss": 4.390580984475932, + "tokens_seen": 197525504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047481444332998996, + "loss": 1.87, + "theoretical_loss": 4.3903995696441065, + "tokens_seen": 197591040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004748044132397192, + "loss": 1.9231, + "theoretical_loss": 4.390218231814504, + "tokens_seen": 197656576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747943831494483, + "loss": 1.7233, + "theoretical_loss": 4.390036970928929, + "tokens_seen": 197722112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047478435305917756, + "loss": 1.7265, + "theoretical_loss": 4.3898557869292425, + "tokens_seen": 197787648 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047477432296890674, + "loss": 1.7902, + "theoretical_loss": 4.389674679757374, + "tokens_seen": 197853184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747642928786359, + "loss": 1.7361, + "theoretical_loss": 4.389493649355314, + "tokens_seen": 197918720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747542627883651, + "loss": 1.746, + "theoretical_loss": 4.389312695665115, + "tokens_seen": 197984256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747442326980943, + "loss": 1.8561, + "theoretical_loss": 4.389131818628895, + "tokens_seen": 198049792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047473420260782346, + "loss": 1.8381, + "theoretical_loss": 4.388951018188832, + "tokens_seen": 198115328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747241725175527, + "loss": 1.8203, + "theoretical_loss": 4.388770294287169, + "tokens_seen": 198180864 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004747141424272818, + "loss": 1.871, + "theoretical_loss": 4.38858964686621, + "tokens_seen": 198246400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047470411233701106, + "loss": 1.8664, + "theoretical_loss": 4.388409075868323, + "tokens_seen": 198311936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746940822467402, + "loss": 1.9059, + "theoretical_loss": 4.388228581235937, + "tokens_seen": 198377472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746840521564694, + "loss": 1.8484, + "theoretical_loss": 4.388048162911545, + "tokens_seen": 198443008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746740220661986, + "loss": 1.768, + "theoretical_loss": 4.3878678208377, + "tokens_seen": 198508544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746639919759278, + "loss": 1.8619, + "theoretical_loss": 4.38768755495702, + "tokens_seen": 198574080 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047465396188565696, + "loss": 1.8075, + "theoretical_loss": 4.387507365212182, + "tokens_seen": 198639616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746439317953862, + "loss": 1.7656, + "theoretical_loss": 4.387327251545927, + "tokens_seen": 198705152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746339017051153, + "loss": 1.7677, + "theoretical_loss": 4.387147213901059, + "tokens_seen": 198770688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047462387161484456, + "loss": 1.7415, + "theoretical_loss": 4.386967252220441, + "tokens_seen": 198836224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746138415245737, + "loss": 1.8826, + "theoretical_loss": 4.386787366447, + "tokens_seen": 198901760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004746038114343029, + "loss": 1.8953, + "theoretical_loss": 4.386607556523723, + "tokens_seen": 198967296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745937813440321, + "loss": 1.8814, + "theoretical_loss": 4.386427822393661, + "tokens_seen": 199032832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745837512537613, + "loss": 1.8198, + "theoretical_loss": 4.386248163999923, + "tokens_seen": 199098368 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047457372116349047, + "loss": 1.8839, + "theoretical_loss": 4.3860685812856826, + "tokens_seen": 199163904 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047456369107321965, + "loss": 1.8966, + "theoretical_loss": 4.385889074194173, + "tokens_seen": 199229440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047455366098294883, + "loss": 1.7641, + "theoretical_loss": 4.385709642668688, + "tokens_seen": 199294976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047454363089267807, + "loss": 1.7095, + "theoretical_loss": 4.385530286652586, + "tokens_seen": 199360512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745336008024072, + "loss": 1.7495, + "theoretical_loss": 4.385351006089283, + "tokens_seen": 199426048 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047452357071213643, + "loss": 1.7678, + "theoretical_loss": 4.385171800922257, + "tokens_seen": 199491584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047451354062186555, + "loss": 1.8091, + "theoretical_loss": 4.384992671095047, + "tokens_seen": 199557120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004745035105315948, + "loss": 1.7582, + "theoretical_loss": 4.384813616551252, + "tokens_seen": 199622656 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047449348044132397, + "loss": 1.7214, + "theoretical_loss": 4.384634637234534, + "tokens_seen": 199688192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047448345035105315, + "loss": 1.8077, + "theoretical_loss": 4.384455733088612, + "tokens_seen": 199753728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744734202607824, + "loss": 1.82, + "theoretical_loss": 4.38427690405727, + "tokens_seen": 199819264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 79488, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6834254264831543, + "objective/train/theoretical_loss": 4.384098150084349, + "objective/train/tokens_used": 220344800, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047446339017051157, + "loss": 1.7548, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047445336008024075, + "loss": 1.8479, + "theoretical_loss": 4.383919471113751, + "tokens_seen": 199950336 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047444332998996993, + "loss": 1.8975, + "theoretical_loss": 4.383740867089441, + "tokens_seen": 200015872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744332998996991, + "loss": 1.801, + "theoretical_loss": 4.383562337955439, + "tokens_seen": 200081408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004744232698094283, + "loss": 1.6689, + "theoretical_loss": 4.3833838836558305, + "tokens_seen": 200146944 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047441323971915753, + "loss": 1.8239, + "theoretical_loss": 4.383205504134758, + "tokens_seen": 200212480 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047440320962888666, + "loss": 1.8739, + "theoretical_loss": 4.383027199336424, + "tokens_seen": 200278016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743931795386159, + "loss": 1.8294, + "theoretical_loss": 4.382848969205092, + "tokens_seen": 200343552 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474383149448345, + "loss": 1.9202, + "theoretical_loss": 4.3826708136850865, + "tokens_seen": 200409088 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047437311935807425, + "loss": 1.8644, + "theoretical_loss": 4.382492732720786, + "tokens_seen": 200474624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047436308926780343, + "loss": 1.9199, + "theoretical_loss": 4.382314726256637, + "tokens_seen": 200540160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743530591775326, + "loss": 1.7535, + "theoretical_loss": 4.382136794237138, + "tokens_seen": 200605696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743430290872618, + "loss": 1.8425, + "theoretical_loss": 4.3819589366068525, + "tokens_seen": 200671232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047433299899699103, + "loss": 1.7503, + "theoretical_loss": 4.381781153310399, + "tokens_seen": 200736768 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047432296890672016, + "loss": 1.7913, + "theoretical_loss": 4.381603444292459, + "tokens_seen": 200802304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743129388164494, + "loss": 1.7183, + "theoretical_loss": 4.38142580949777, + "tokens_seen": 200867840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004743029087261785, + "loss": 1.7801, + "theoretical_loss": 4.381248248871131, + "tokens_seen": 200933376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047429287863590776, + "loss": 1.7758, + "theoretical_loss": 4.3810707623574, + "tokens_seen": 200998912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047428284854563694, + "loss": 1.8005, + "theoretical_loss": 4.380893349901491, + "tokens_seen": 201064448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742728184553661, + "loss": 1.7153, + "theoretical_loss": 4.380716011448381, + "tokens_seen": 201129984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742627883650953, + "loss": 1.8249, + "theoretical_loss": 4.380538746943103, + "tokens_seen": 201195520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742527582748245, + "loss": 1.6826, + "theoretical_loss": 4.380361556330749, + "tokens_seen": 201261056 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047424272818455366, + "loss": 1.6561, + "theoretical_loss": 4.380184439556472, + "tokens_seen": 201326592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742326980942829, + "loss": 1.7845, + "theoretical_loss": 4.38000739656548, + "tokens_seen": 201392128 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474222668004012, + "loss": 1.7979, + "theoretical_loss": 4.379830427303041, + "tokens_seen": 201457664 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047421263791374126, + "loss": 1.7656, + "theoretical_loss": 4.379653531714483, + "tokens_seen": 201523200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004742026078234704, + "loss": 1.9076, + "theoretical_loss": 4.37947670974519, + "tokens_seen": 201588736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741925777331996, + "loss": 1.7101, + "theoretical_loss": 4.379299961340605, + "tokens_seen": 201654272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741825476429288, + "loss": 1.8579, + "theoretical_loss": 4.37912328644623, + "tokens_seen": 201719808 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474172517552658, + "loss": 1.8391, + "theoretical_loss": 4.378946685007623, + "tokens_seen": 201785344 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047416248746238716, + "loss": 1.879, + "theoretical_loss": 4.378770156970403, + "tokens_seen": 201850880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741524573721164, + "loss": 1.7863, + "theoretical_loss": 4.378593702280243, + "tokens_seen": 201916416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741424272818455, + "loss": 1.7658, + "theoretical_loss": 4.378417320882879, + "tokens_seen": 201981952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047413239719157476, + "loss": 1.6862, + "theoretical_loss": 4.378241012724099, + "tokens_seen": 202047488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741223671013039, + "loss": 1.8474, + "theoretical_loss": 4.378064777749753, + "tokens_seen": 202113024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741123370110331, + "loss": 1.7418, + "theoretical_loss": 4.377888615905746, + "tokens_seen": 202178560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004741023069207623, + "loss": 1.8115, + "theoretical_loss": 4.377712527138042, + "tokens_seen": 202244096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004740922768304915, + "loss": 1.8503, + "theoretical_loss": 4.3775365113926625, + "tokens_seen": 202309632 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047408224674022067, + "loss": 1.7036, + "theoretical_loss": 4.377360568615685, + "tokens_seen": 202375168 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047407221664994985, + "loss": 1.7851, + "theoretical_loss": 4.377184698753244, + "tokens_seen": 202440704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047406218655967903, + "loss": 1.6966, + "theoretical_loss": 4.377008901751536, + "tokens_seen": 202506240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047405215646940827, + "loss": 1.8088, + "theoretical_loss": 4.376833177556806, + "tokens_seen": 202571776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004740421263791374, + "loss": 1.8431, + "theoretical_loss": 4.3766575261153635, + "tokens_seen": 202637312 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047403209628886663, + "loss": 1.8585, + "theoretical_loss": 4.3764819473735725, + "tokens_seen": 202702848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047402206619859575, + "loss": 1.7544, + "theoretical_loss": 4.376306441277852, + "tokens_seen": 202768384 + }, + { + "epoch": 0.06, + "learning_rate": 0.000474012036108325, + "loss": 1.7614, + "theoretical_loss": 4.376131007774681, + "tokens_seen": 202833920 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047400200601805417, + "loss": 1.8107, + "theoretical_loss": 4.375955646810594, + "tokens_seen": 202899456 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047399197592778335, + "loss": 1.7123, + "theoretical_loss": 4.37578035833218, + "tokens_seen": 202964992 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047398194583751253, + "loss": 1.7367, + "theoretical_loss": 4.375605142286087, + "tokens_seen": 203030528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047397191574724177, + "loss": 1.7996, + "theoretical_loss": 4.37542999861902, + "tokens_seen": 203096064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 80605, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8830738067626953, + "objective/train/theoretical_loss": 4.375254927277737, + "objective/train/tokens_used": 223621600, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739618856569709, + "loss": 1.8886, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047395185556670013, + "loss": 1.8956, + "theoretical_loss": 4.375079928209058, + "tokens_seen": 203227136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047394182547642926, + "loss": 1.7047, + "theoretical_loss": 4.374905001359851, + "tokens_seen": 203292672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739317953861585, + "loss": 1.8006, + "theoretical_loss": 4.374730146677049, + "tokens_seen": 203358208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004739217652958877, + "loss": 1.8381, + "theoretical_loss": 4.374555364107635, + "tokens_seen": 203423744 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047391173520561686, + "loss": 1.7831, + "theoretical_loss": 4.37438065359865, + "tokens_seen": 203489280 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047390170511534604, + "loss": 1.8453, + "theoretical_loss": 4.374206015097193, + "tokens_seen": 203554816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738916750250752, + "loss": 1.6967, + "theoretical_loss": 4.374031448550414, + "tokens_seen": 203620352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738816449348044, + "loss": 1.7847, + "theoretical_loss": 4.373856953905523, + "tokens_seen": 203685888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047387161484453363, + "loss": 1.8968, + "theoretical_loss": 4.373682531109785, + "tokens_seen": 203751424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047386158475426276, + "loss": 1.7937, + "theoretical_loss": 4.37350818011052, + "tokens_seen": 203816960 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473851554663992, + "loss": 1.8662, + "theoretical_loss": 4.3733339008551, + "tokens_seen": 203882496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738415245737211, + "loss": 1.8164, + "theoretical_loss": 4.373159693290961, + "tokens_seen": 203948032 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047383149448345036, + "loss": 1.7793, + "theoretical_loss": 4.372985557365586, + "tokens_seen": 204013568 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047382146439317954, + "loss": 1.806, + "theoretical_loss": 4.372811493026518, + "tokens_seen": 204079104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738114343029087, + "loss": 1.7607, + "theoretical_loss": 4.3726375002213524, + "tokens_seen": 204144640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004738014042126379, + "loss": 1.7009, + "theoretical_loss": 4.372463578897743, + "tokens_seen": 204210176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047379137412236714, + "loss": 1.8284, + "theoretical_loss": 4.372289729003397, + "tokens_seen": 204275712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047378134403209626, + "loss": 1.7306, + "theoretical_loss": 4.372115950486075, + "tokens_seen": 204341248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737713139418255, + "loss": 1.8965, + "theoretical_loss": 4.371942243293595, + "tokens_seen": 204406784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737612838515546, + "loss": 1.8325, + "theoretical_loss": 4.371768607373829, + "tokens_seen": 204472320 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047375125376128386, + "loss": 1.8471, + "theoretical_loss": 4.371595042674704, + "tokens_seen": 204537856 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047374122367101304, + "loss": 1.6713, + "theoretical_loss": 4.371421549144202, + "tokens_seen": 204603392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737311935807422, + "loss": 1.8123, + "theoretical_loss": 4.371248126730357, + "tokens_seen": 204668928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047372116349047146, + "loss": 1.9643, + "theoretical_loss": 4.3710747753812615, + "tokens_seen": 204734464 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737111334002006, + "loss": 1.7152, + "theoretical_loss": 4.370901495045061, + "tokens_seen": 204800000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004737011033099298, + "loss": 1.9091, + "theoretical_loss": 4.370728285669954, + "tokens_seen": 204865536 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473691073219659, + "loss": 1.8764, + "theoretical_loss": 4.370555147204195, + "tokens_seen": 204931072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736810431293882, + "loss": 1.7944, + "theoretical_loss": 4.370382079596092, + "tokens_seen": 204996608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047367101303911736, + "loss": 1.9121, + "theoretical_loss": 4.370209082794006, + "tokens_seen": 205062144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736609829488466, + "loss": 1.7955, + "theoretical_loss": 4.370036156746357, + "tokens_seen": 205127680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736509528585757, + "loss": 1.6985, + "theoretical_loss": 4.369863301401612, + "tokens_seen": 205193216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047364092276830496, + "loss": 1.9803, + "theoretical_loss": 4.369690516708298, + "tokens_seen": 205258752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736308926780341, + "loss": 1.7962, + "theoretical_loss": 4.3695178026149915, + "tokens_seen": 205324288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736208625877633, + "loss": 1.7637, + "theoretical_loss": 4.369345159070327, + "tokens_seen": 205389824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736108324974925, + "loss": 1.6925, + "theoretical_loss": 4.369172586022989, + "tokens_seen": 205455360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004736008024072217, + "loss": 1.8397, + "theoretical_loss": 4.369000083421716, + "tokens_seen": 205520896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047359077231695087, + "loss": 1.7593, + "theoretical_loss": 4.368827651215304, + "tokens_seen": 205586432 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047358074222668005, + "loss": 1.7446, + "theoretical_loss": 4.368655289352598, + "tokens_seen": 205651968 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047357071213640923, + "loss": 1.7966, + "theoretical_loss": 4.3684829977824995, + "tokens_seen": 205717504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047356068204613847, + "loss": 1.7009, + "theoretical_loss": 4.368310776453961, + "tokens_seen": 205783040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004735506519558676, + "loss": 1.7392, + "theoretical_loss": 4.368138625315989, + "tokens_seen": 205848576 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047354062186559683, + "loss": 1.7146, + "theoretical_loss": 4.367966544317646, + "tokens_seen": 205914112 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047353059177532595, + "loss": 1.8135, + "theoretical_loss": 4.367794533408043, + "tokens_seen": 205979648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004735205616850552, + "loss": 1.6708, + "theoretical_loss": 4.367622592536348, + "tokens_seen": 206045184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047351053159478437, + "loss": 1.7702, + "theoretical_loss": 4.36745072165178, + "tokens_seen": 206110720 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047350050150451355, + "loss": 1.7999, + "theoretical_loss": 4.36727892070361, + "tokens_seen": 206176256 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047349047141424273, + "loss": 1.7723, + "theoretical_loss": 4.367107189641167, + "tokens_seen": 206241792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047348044132397197, + "loss": 1.7797, + "theoretical_loss": 4.3669355284138245, + "tokens_seen": 206307328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734704112337011, + "loss": 1.8309, + "theoretical_loss": 4.366763936971017, + "tokens_seen": 206372864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 81724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7647249698638916, + "objective/train/theoretical_loss": 4.366592415262226, + "objective/train/tokens_used": 226898400, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047346038114343033, + "loss": 1.7656, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047345035105315946, + "loss": 1.8441, + "theoretical_loss": 4.366420963236989, + "tokens_seen": 206503936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734403209628887, + "loss": 1.9181, + "theoretical_loss": 4.366249580844894, + "tokens_seen": 206569472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734302908726179, + "loss": 1.8258, + "theoretical_loss": 4.366078268035581, + "tokens_seen": 206635008 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047342026078234706, + "loss": 1.7749, + "theoretical_loss": 4.365907024758744, + "tokens_seen": 206700544 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047341023069207624, + "loss": 1.9016, + "theoretical_loss": 4.365735850964131, + "tokens_seen": 206766080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004734002006018054, + "loss": 1.7342, + "theoretical_loss": 4.365564746601537, + "tokens_seen": 206831616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733901705115346, + "loss": 1.9224, + "theoretical_loss": 4.365393711620814, + "tokens_seen": 206897152 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047338014042126383, + "loss": 1.7935, + "theoretical_loss": 4.365222745971863, + "tokens_seen": 206962688 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047337011033099296, + "loss": 1.5462, + "theoretical_loss": 4.365051849604638, + "tokens_seen": 207028224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733600802407222, + "loss": 1.8155, + "theoretical_loss": 4.364881022469148, + "tokens_seen": 207093760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733500501504513, + "loss": 1.8691, + "theoretical_loss": 4.364710264515448, + "tokens_seen": 207159296 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047334002006018056, + "loss": 1.6979, + "theoretical_loss": 4.36453957569365, + "tokens_seen": 207224832 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047332998996990974, + "loss": 1.7634, + "theoretical_loss": 4.364368955953913, + "tokens_seen": 207290368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733199598796389, + "loss": 1.7224, + "theoretical_loss": 4.364198405246453, + "tokens_seen": 207355904 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004733099297893681, + "loss": 1.7985, + "theoretical_loss": 4.364027923521535, + "tokens_seen": 207421440 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047329989969909734, + "loss": 1.716, + "theoretical_loss": 4.363857510729473, + "tokens_seen": 207486976 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047328986960882646, + "loss": 1.8407, + "theoretical_loss": 4.363687166820637, + "tokens_seen": 207552512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732798395185557, + "loss": 1.9034, + "theoretical_loss": 4.363516891745445, + "tokens_seen": 207618048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732698094282848, + "loss": 1.9315, + "theoretical_loss": 4.36334668545437, + "tokens_seen": 207683584 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047325977933801406, + "loss": 1.8464, + "theoretical_loss": 4.363176547897931, + "tokens_seen": 207749120 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047324974924774324, + "loss": 1.8252, + "theoretical_loss": 4.363006479026702, + "tokens_seen": 207814656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732397191574724, + "loss": 1.8257, + "theoretical_loss": 4.3628364787913085, + "tokens_seen": 207880192 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732296890672016, + "loss": 1.7059, + "theoretical_loss": 4.362666547142424, + "tokens_seen": 207945728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004732196589769308, + "loss": 1.6127, + "theoretical_loss": 4.362496684030775, + "tokens_seen": 208011264 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047320962888665997, + "loss": 1.7426, + "theoretical_loss": 4.36232688940714, + "tokens_seen": 208076800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731995987963892, + "loss": 1.9333, + "theoretical_loss": 4.362157163222346, + "tokens_seen": 208142336 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047318956870611833, + "loss": 1.8319, + "theoretical_loss": 4.361987505427272, + "tokens_seen": 208207872 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047317953861584757, + "loss": 1.8763, + "theoretical_loss": 4.361817915972846, + "tokens_seen": 208273408 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047316950852557675, + "loss": 1.8169, + "theoretical_loss": 4.36164839481005, + "tokens_seen": 208338944 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047315947843530593, + "loss": 1.96, + "theoretical_loss": 4.361478941889915, + "tokens_seen": 208404480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731494483450351, + "loss": 1.8294, + "theoretical_loss": 4.36130955716352, + "tokens_seen": 208470016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731394182547643, + "loss": 1.7687, + "theoretical_loss": 4.361140240581998, + "tokens_seen": 208535552 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047312938816449347, + "loss": 1.811, + "theoretical_loss": 4.36097099209653, + "tokens_seen": 208601088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004731193580742227, + "loss": 1.8138, + "theoretical_loss": 4.36080181165835, + "tokens_seen": 208666624 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047310932798395183, + "loss": 1.7259, + "theoretical_loss": 4.360632699218739, + "tokens_seen": 208732160 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047309929789368107, + "loss": 1.7383, + "theoretical_loss": 4.360463654729029, + "tokens_seen": 208797696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730892678034102, + "loss": 1.8036, + "theoretical_loss": 4.360294678140605, + "tokens_seen": 208863232 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047307923771313943, + "loss": 1.8128, + "theoretical_loss": 4.360125769404897, + "tokens_seen": 208928768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730692076228686, + "loss": 1.8951, + "theoretical_loss": 4.35995692847339, + "tokens_seen": 208994304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730591775325978, + "loss": 1.7101, + "theoretical_loss": 4.359788155297615, + "tokens_seen": 209059840 + }, + { + "epoch": 0.06, + "learning_rate": 0.000473049147442327, + "loss": 1.7326, + "theoretical_loss": 4.359619449829156, + "tokens_seen": 209125376 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047303911735205615, + "loss": 1.8619, + "theoretical_loss": 4.359450812019643, + "tokens_seen": 209190912 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047302908726178534, + "loss": 1.8419, + "theoretical_loss": 4.359282241820761, + "tokens_seen": 209256448 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047301905717151457, + "loss": 1.6755, + "theoretical_loss": 4.359113739184238, + "tokens_seen": 209321984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004730090270812437, + "loss": 1.8649, + "theoretical_loss": 4.358945304061858, + "tokens_seen": 209387520 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047299899699097293, + "loss": 1.7864, + "theoretical_loss": 4.35877693640545, + "tokens_seen": 209453056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729889669007021, + "loss": 1.8986, + "theoretical_loss": 4.358608636166894, + "tokens_seen": 209518592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729789368104313, + "loss": 1.8224, + "theoretical_loss": 4.358440403298119, + "tokens_seen": 209584128 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047296890672016053, + "loss": 1.7195, + "theoretical_loss": 4.3582722377511045, + "tokens_seen": 209649664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 82917, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7261275053024292, + "objective/train/theoretical_loss": 4.358104139477878, + "objective/train/tokens_used": 230175200, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047295887662988966, + "loss": 1.7882, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729488465396189, + "loss": 1.8067, + "theoretical_loss": 4.357936108430517, + "tokens_seen": 209780736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729388164493481, + "loss": 1.8048, + "theoretical_loss": 4.357768144561146, + "tokens_seen": 209846272 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047292878635907726, + "loss": 1.8471, + "theoretical_loss": 4.357600247821942, + "tokens_seen": 209911808 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047291875626880644, + "loss": 1.7575, + "theoretical_loss": 4.3574324181651285, + "tokens_seen": 209977344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004729087261785356, + "loss": 1.7582, + "theoretical_loss": 4.357264655542978, + "tokens_seen": 210042880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728986960882648, + "loss": 1.7481, + "theoretical_loss": 4.357096959907813, + "tokens_seen": 210108416 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047288866599799403, + "loss": 1.7118, + "theoretical_loss": 4.356929331212004, + "tokens_seen": 210173952 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047287863590772316, + "loss": 1.724, + "theoretical_loss": 4.356761769407969, + "tokens_seen": 210239488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728686058174524, + "loss": 1.9151, + "theoretical_loss": 4.356594274448178, + "tokens_seen": 210305024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728585757271815, + "loss": 1.9301, + "theoretical_loss": 4.356426846285147, + "tokens_seen": 210370560 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047284854563691076, + "loss": 1.8191, + "theoretical_loss": 4.3562594848714395, + "tokens_seen": 210436096 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047283851554663994, + "loss": 1.8411, + "theoretical_loss": 4.3560921901596705, + "tokens_seen": 210501632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728284854563691, + "loss": 1.8997, + "theoretical_loss": 4.355924962102502, + "tokens_seen": 210567168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004728184553660983, + "loss": 1.8692, + "theoretical_loss": 4.3557578006526425, + "tokens_seen": 210632704 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047280842527582754, + "loss": 1.7409, + "theoretical_loss": 4.355590705762852, + "tokens_seen": 210698240 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047279839518555666, + "loss": 1.7883, + "theoretical_loss": 4.355423677385937, + "tokens_seen": 210763776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727883650952859, + "loss": 1.7599, + "theoretical_loss": 4.355256715474751, + "tokens_seen": 210829312 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472778335005015, + "loss": 1.8356, + "theoretical_loss": 4.355089819982198, + "tokens_seen": 210894848 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047276830491474426, + "loss": 1.8478, + "theoretical_loss": 4.354922990861229, + "tokens_seen": 210960384 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047275827482447344, + "loss": 1.8184, + "theoretical_loss": 4.354756228064842, + "tokens_seen": 211025920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727482447342026, + "loss": 1.8766, + "theoretical_loss": 4.354589531546083, + "tokens_seen": 211091456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727382146439318, + "loss": 1.7866, + "theoretical_loss": 4.354422901258048, + "tokens_seen": 211156992 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472728184553661, + "loss": 1.6652, + "theoretical_loss": 4.354256337153878, + "tokens_seen": 211222528 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047271815446339017, + "loss": 1.7519, + "theoretical_loss": 4.354089839186763, + "tokens_seen": 211288064 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004727081243731194, + "loss": 1.7915, + "theoretical_loss": 4.35392340730994, + "tokens_seen": 211353600 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047269809428284853, + "loss": 1.7559, + "theoretical_loss": 4.353757041476695, + "tokens_seen": 211419136 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047268806419257777, + "loss": 1.8316, + "theoretical_loss": 4.35359074164036, + "tokens_seen": 211484672 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047267803410230695, + "loss": 1.7781, + "theoretical_loss": 4.3534245077543146, + "tokens_seen": 211550208 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047266800401203613, + "loss": 1.8688, + "theoretical_loss": 4.353258339771986, + "tokens_seen": 211615744 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726579739217653, + "loss": 1.7365, + "theoretical_loss": 4.353092237646849, + "tokens_seen": 211681280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726479438314945, + "loss": 1.7689, + "theoretical_loss": 4.3529262013324255, + "tokens_seen": 211746816 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047263791374122367, + "loss": 1.7243, + "theoretical_loss": 4.352760230782284, + "tokens_seen": 211812352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004726278836509529, + "loss": 1.7723, + "theoretical_loss": 4.352594325950041, + "tokens_seen": 211877888 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047261785356068203, + "loss": 1.7904, + "theoretical_loss": 4.35242848678936, + "tokens_seen": 211943424 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047260782347041127, + "loss": 1.8312, + "theoretical_loss": 4.35226271325395, + "tokens_seen": 212008960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725977933801404, + "loss": 1.9041, + "theoretical_loss": 4.352097005297569, + "tokens_seen": 212074496 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047258776328986963, + "loss": 1.7358, + "theoretical_loss": 4.351931362874019, + "tokens_seen": 212140032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725777331995988, + "loss": 1.8199, + "theoretical_loss": 4.351765785937154, + "tokens_seen": 212205568 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472567703109328, + "loss": 1.8399, + "theoretical_loss": 4.3516002744408695, + "tokens_seen": 212271104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725576730190572, + "loss": 1.8085, + "theoretical_loss": 4.351434828339109, + "tokens_seen": 212336640 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047254764292878636, + "loss": 1.851, + "theoretical_loss": 4.351269447585864, + "tokens_seen": 212402176 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047253761283851554, + "loss": 1.8679, + "theoretical_loss": 4.3511041321351716, + "tokens_seen": 212467712 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047252758274824477, + "loss": 1.7947, + "theoretical_loss": 4.350938881941116, + "tokens_seen": 212533248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004725175526579739, + "loss": 1.7812, + "theoretical_loss": 4.350773696957827, + "tokens_seen": 212598784 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047250752256770313, + "loss": 1.8764, + "theoretical_loss": 4.350608577139481, + "tokens_seen": 212664320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724974924774323, + "loss": 1.5963, + "theoretical_loss": 4.350443522440301, + "tokens_seen": 212729856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724874623871615, + "loss": 1.8501, + "theoretical_loss": 4.350278532814556, + "tokens_seen": 212795392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724774322968907, + "loss": 1.7311, + "theoretical_loss": 4.350113608216563, + "tokens_seen": 212860928 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047246740220661986, + "loss": 1.8253, + "theoretical_loss": 4.3499487486006805, + "tokens_seen": 212926464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 84140, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.64862859249115, + "objective/train/theoretical_loss": 4.349783953921318, + "objective/train/tokens_used": 233452000, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047245737211634904, + "loss": 1.6702, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724473420260783, + "loss": 1.7146, + "theoretical_loss": 4.349619224132929, + "tokens_seen": 213057536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004724373119358074, + "loss": 1.7314, + "theoretical_loss": 4.349454559190011, + "tokens_seen": 213123072 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047242728184553664, + "loss": 1.7707, + "theoretical_loss": 4.349289959047112, + "tokens_seen": 213188608 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047241725175526576, + "loss": 1.7188, + "theoretical_loss": 4.349125423658821, + "tokens_seen": 213254144 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472407221664995, + "loss": 1.8368, + "theoretical_loss": 4.348960952979776, + "tokens_seen": 213319680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723971915747242, + "loss": 1.7782, + "theoretical_loss": 4.3487965469646594, + "tokens_seen": 213385216 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047238716148445336, + "loss": 1.7919, + "theoretical_loss": 4.3486322055682, + "tokens_seen": 213450752 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047237713139418254, + "loss": 1.815, + "theoretical_loss": 4.348467928745171, + "tokens_seen": 213516288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723671013039117, + "loss": 1.6945, + "theoretical_loss": 4.348303716450391, + "tokens_seen": 213581824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723570712136409, + "loss": 1.8512, + "theoretical_loss": 4.348139568638727, + "tokens_seen": 213647360 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047234704112337014, + "loss": 1.7422, + "theoretical_loss": 4.347975485265087, + "tokens_seen": 213712896 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047233701103309927, + "loss": 1.8063, + "theoretical_loss": 4.347811466284428, + "tokens_seen": 213778432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723269809428285, + "loss": 1.7432, + "theoretical_loss": 4.34764751165175, + "tokens_seen": 213843968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004723169508525577, + "loss": 1.7311, + "theoretical_loss": 4.347483621322099, + "tokens_seen": 213909504 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047230692076228686, + "loss": 1.7934, + "theoretical_loss": 4.347319795250567, + "tokens_seen": 213975040 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047229689067201605, + "loss": 1.7614, + "theoretical_loss": 4.347156033392291, + "tokens_seen": 214040576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004722868605817452, + "loss": 1.8365, + "theoretical_loss": 4.346992335702451, + "tokens_seen": 214106112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0004722768304914744, + "loss": 1.7812, + "theoretical_loss": 4.346828702136273, + "tokens_seen": 214171648 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047226680040120364, + "loss": 1.7864, + "theoretical_loss": 4.34666513264903, + "tokens_seen": 214237184 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047225677031093277, + "loss": 1.8177, + "theoretical_loss": 4.346501627196037, + "tokens_seen": 214302720 + }, + { + "epoch": 0.06, + "learning_rate": 0.000472246740220662, + "loss": 1.817, + "theoretical_loss": 4.346338185732655, + "tokens_seen": 214368256 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047223671013039113, + "loss": 1.8226, + "theoretical_loss": 4.3461748082142915, + "tokens_seen": 214433792 + }, + { + "epoch": 0.06, + "learning_rate": 0.00047222668004012037, + "loss": 1.9009, + "theoretical_loss": 4.346011494596395, + "tokens_seen": 214499328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004722166499498496, + "loss": 1.6802, + "theoretical_loss": 4.34584824483446, + "tokens_seen": 214564864 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047220661985957873, + "loss": 1.8136, + "theoretical_loss": 4.345685058884029, + "tokens_seen": 214630400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047219658976930797, + "loss": 1.7196, + "theoretical_loss": 4.345521936700684, + "tokens_seen": 214695936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047218655967903715, + "loss": 1.8314, + "theoretical_loss": 4.345358878240053, + "tokens_seen": 214761472 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047217652958876633, + "loss": 1.9206, + "theoretical_loss": 4.345195883457811, + "tokens_seen": 214827008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721664994984955, + "loss": 1.8532, + "theoretical_loss": 4.3450329523096745, + "tokens_seen": 214892544 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721564694082247, + "loss": 1.8934, + "theoretical_loss": 4.344870084751406, + "tokens_seen": 214958080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047214643931795387, + "loss": 1.8966, + "theoretical_loss": 4.34470728073881, + "tokens_seen": 215023616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721364092276831, + "loss": 1.7929, + "theoretical_loss": 4.344544540227737, + "tokens_seen": 215089152 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047212637913741223, + "loss": 1.7767, + "theoretical_loss": 4.344381863174082, + "tokens_seen": 215154688 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047211634904714147, + "loss": 1.8217, + "theoretical_loss": 4.344219249533783, + "tokens_seen": 215220224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004721063189568706, + "loss": 1.9374, + "theoretical_loss": 4.34405669926282, + "tokens_seen": 215285760 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047209628886659983, + "loss": 1.7948, + "theoretical_loss": 4.343894212317224, + "tokens_seen": 215351296 + }, + { + "epoch": 0.07, + "learning_rate": 0.000472086258776329, + "loss": 1.7831, + "theoretical_loss": 4.34373178865306, + "tokens_seen": 215416832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720762286860582, + "loss": 1.6392, + "theoretical_loss": 4.343569428226445, + "tokens_seen": 215482368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720661985957874, + "loss": 1.6728, + "theoretical_loss": 4.343407130993536, + "tokens_seen": 215547904 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047205616850551656, + "loss": 1.8347, + "theoretical_loss": 4.343244896910534, + "tokens_seen": 215613440 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047204613841524574, + "loss": 1.9378, + "theoretical_loss": 4.343082725933684, + "tokens_seen": 215678976 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047203610832497497, + "loss": 1.8383, + "theoretical_loss": 4.342920618019275, + "tokens_seen": 215744512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720260782347041, + "loss": 1.6674, + "theoretical_loss": 4.342758573123639, + "tokens_seen": 215810048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047201604814443333, + "loss": 1.8419, + "theoretical_loss": 4.342596591203153, + "tokens_seen": 215875584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004720060180541625, + "loss": 1.6723, + "theoretical_loss": 4.3424346722142335, + "tokens_seen": 215941120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719959879638917, + "loss": 1.7339, + "theoretical_loss": 4.342272816113344, + "tokens_seen": 216006656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719859578736209, + "loss": 1.8355, + "theoretical_loss": 4.342111022856992, + "tokens_seen": 216072192 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047197592778335006, + "loss": 1.8576, + "theoretical_loss": 4.341949292401725, + "tokens_seen": 216137728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047196589769307924, + "loss": 1.7736, + "theoretical_loss": 4.341787624704135, + "tokens_seen": 216203264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 85191, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8263232707977295, + "objective/train/theoretical_loss": 4.341626019720858, + "objective/train/tokens_used": 236728800, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719558676028085, + "loss": 1.7557, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719458375125376, + "loss": 1.7973, + "theoretical_loss": 4.341464477408572, + "tokens_seen": 216334336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047193580742226684, + "loss": 1.7896, + "theoretical_loss": 4.341302997724, + "tokens_seen": 216399872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047192577733199596, + "loss": 1.9257, + "theoretical_loss": 4.341141580623905, + "tokens_seen": 216465408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719157472417252, + "loss": 1.7876, + "theoretical_loss": 4.340980226065096, + "tokens_seen": 216530944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004719057171514544, + "loss": 1.8388, + "theoretical_loss": 4.340818934004421, + "tokens_seen": 216596480 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047189568706118356, + "loss": 1.7148, + "theoretical_loss": 4.340657704398774, + "tokens_seen": 216662016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047188565697091274, + "loss": 1.8396, + "theoretical_loss": 4.340496537205093, + "tokens_seen": 216727552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718756268806419, + "loss": 1.7572, + "theoretical_loss": 4.340335432380353, + "tokens_seen": 216793088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718655967903711, + "loss": 1.8542, + "theoretical_loss": 4.340174389881579, + "tokens_seen": 216858624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047185556670010034, + "loss": 1.862, + "theoretical_loss": 4.340013409665833, + "tokens_seen": 216924160 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047184553660982947, + "loss": 1.7007, + "theoretical_loss": 4.339852491690221, + "tokens_seen": 216989696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718355065195587, + "loss": 1.8006, + "theoretical_loss": 4.339691635911892, + "tokens_seen": 217055232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004718254764292879, + "loss": 1.7816, + "theoretical_loss": 4.339530842288037, + "tokens_seen": 217120768 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047181544633901706, + "loss": 1.7447, + "theoretical_loss": 4.339370110775892, + "tokens_seen": 217186304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047180541624874625, + "loss": 1.7571, + "theoretical_loss": 4.33920944133273, + "tokens_seen": 217251840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717953861584754, + "loss": 1.8404, + "theoretical_loss": 4.3390488339158715, + "tokens_seen": 217317376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717853560682046, + "loss": 1.8259, + "theoretical_loss": 4.338888288482675, + "tokens_seen": 217382912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047177532597793384, + "loss": 1.7734, + "theoretical_loss": 4.338727804990544, + "tokens_seen": 217448448 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047176529588766297, + "loss": 1.741, + "theoretical_loss": 4.338567383396924, + "tokens_seen": 217513984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717552657973922, + "loss": 1.801, + "theoretical_loss": 4.338407023659302, + "tokens_seen": 217579520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047174523570712133, + "loss": 1.6264, + "theoretical_loss": 4.338246725735205, + "tokens_seen": 217645056 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047173520561685057, + "loss": 1.7792, + "theoretical_loss": 4.338086489582205, + "tokens_seen": 217710592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047172517552657975, + "loss": 1.7282, + "theoretical_loss": 4.3379263151579135, + "tokens_seen": 217776128 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047171514543630893, + "loss": 1.6674, + "theoretical_loss": 4.337766202419986, + "tokens_seen": 217841664 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004717051153460381, + "loss": 1.8294, + "theoretical_loss": 4.337606151326119, + "tokens_seen": 217907200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047169508525576735, + "loss": 1.7957, + "theoretical_loss": 4.337446161834048, + "tokens_seen": 217972736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716850551654965, + "loss": 1.8669, + "theoretical_loss": 4.3372862339015565, + "tokens_seen": 218038272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716750250752257, + "loss": 1.7109, + "theoretical_loss": 4.337126367486462, + "tokens_seen": 218103808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047166499498495484, + "loss": 1.7993, + "theoretical_loss": 4.33696656254663, + "tokens_seen": 218169344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047165496489468407, + "loss": 1.8077, + "theoretical_loss": 4.336806819039962, + "tokens_seen": 218234880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047164493480441325, + "loss": 1.5778, + "theoretical_loss": 4.336647136924407, + "tokens_seen": 218300416 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047163490471414243, + "loss": 1.7134, + "theoretical_loss": 4.33648751615795, + "tokens_seen": 218365952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716248746238716, + "loss": 1.7848, + "theoretical_loss": 4.33632795669862, + "tokens_seen": 218431488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004716148445336008, + "loss": 1.8188, + "theoretical_loss": 4.336168458504486, + "tokens_seen": 218497024 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047160481444333, + "loss": 1.8059, + "theoretical_loss": 4.336009021533661, + "tokens_seen": 218562560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715947843530592, + "loss": 1.9224, + "theoretical_loss": 4.335849645744295, + "tokens_seen": 218628096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047158475426278834, + "loss": 1.7638, + "theoretical_loss": 4.3356903310945825, + "tokens_seen": 218693632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715747241725176, + "loss": 1.7717, + "theoretical_loss": 4.335531077542758, + "tokens_seen": 218759168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715646940822467, + "loss": 1.7801, + "theoretical_loss": 4.335371885047096, + "tokens_seen": 218824704 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047155466399197594, + "loss": 1.7278, + "theoretical_loss": 4.335212753565914, + "tokens_seen": 218890240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715446339017051, + "loss": 1.8328, + "theoretical_loss": 4.33505368305757, + "tokens_seen": 218955776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715346038114343, + "loss": 1.809, + "theoretical_loss": 4.33489467348046, + "tokens_seen": 219021312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715245737211635, + "loss": 1.676, + "theoretical_loss": 4.334735724793026, + "tokens_seen": 219086848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004715145436308927, + "loss": 1.8563, + "theoretical_loss": 4.334576836953744, + "tokens_seen": 219152384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047150451354062184, + "loss": 1.7999, + "theoretical_loss": 4.334418009921139, + "tokens_seen": 219217920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714944834503511, + "loss": 1.7493, + "theoretical_loss": 4.334259243653769, + "tokens_seen": 219283456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714844533600802, + "loss": 1.8808, + "theoretical_loss": 4.334100538110237, + "tokens_seen": 219348992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047147442326980944, + "loss": 1.8473, + "theoretical_loss": 4.333941893249184, + "tokens_seen": 219414528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714643931795387, + "loss": 1.766, + "theoretical_loss": 4.333783309029296, + "tokens_seen": 219480064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 86440, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7075774669647217, + "objective/train/theoretical_loss": 4.3336247854092935, + "objective/train/tokens_used": 240005600, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714543630892678, + "loss": 1.7127, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047144433299899704, + "loss": 1.7006, + "theoretical_loss": 4.333466322347942, + "tokens_seen": 219611136 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047143430290872616, + "loss": 1.8626, + "theoretical_loss": 4.333307919804045, + "tokens_seen": 219676672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714242728184554, + "loss": 1.8908, + "theoretical_loss": 4.333149577736446, + "tokens_seen": 219742208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004714142427281846, + "loss": 1.8558, + "theoretical_loss": 4.332991296104032, + "tokens_seen": 219807744 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047140421263791376, + "loss": 1.809, + "theoretical_loss": 4.332833074865725, + "tokens_seen": 219873280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047139418254764294, + "loss": 1.7972, + "theoretical_loss": 4.332674913980492, + "tokens_seen": 219938816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713841524573721, + "loss": 1.8248, + "theoretical_loss": 4.332516813407337, + "tokens_seen": 220004352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713741223671013, + "loss": 1.7124, + "theoretical_loss": 4.332358773105307, + "tokens_seen": 220069888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047136409227683054, + "loss": 1.7516, + "theoretical_loss": 4.332200793033485, + "tokens_seen": 220135424 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047135406218655967, + "loss": 1.6992, + "theoretical_loss": 4.332042873150997, + "tokens_seen": 220200960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713440320962889, + "loss": 1.807, + "theoretical_loss": 4.3318850134170095, + "tokens_seen": 220266496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004713340020060181, + "loss": 1.8006, + "theoretical_loss": 4.331727213790725, + "tokens_seen": 220332032 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047132397191574726, + "loss": 1.8363, + "theoretical_loss": 4.331569474231388, + "tokens_seen": 220397568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047131394182547645, + "loss": 1.845, + "theoretical_loss": 4.331411794698286, + "tokens_seen": 220463104 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047130391173520563, + "loss": 1.7323, + "theoretical_loss": 4.33125417515074, + "tokens_seen": 220528640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712938816449348, + "loss": 1.7987, + "theoretical_loss": 4.331096615548115, + "tokens_seen": 220594176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047128385155466404, + "loss": 1.832, + "theoretical_loss": 4.330939115849814, + "tokens_seen": 220659712 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047127382146439317, + "loss": 1.7676, + "theoretical_loss": 4.33078167601528, + "tokens_seen": 220725248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712637913741224, + "loss": 1.8565, + "theoretical_loss": 4.330624296003996, + "tokens_seen": 220790784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047125376128385153, + "loss": 1.7741, + "theoretical_loss": 4.330466975775483, + "tokens_seen": 220856320 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047124373119358077, + "loss": 1.867, + "theoretical_loss": 4.330309715289301, + "tokens_seen": 220921856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047123370110330995, + "loss": 1.7426, + "theoretical_loss": 4.330152514505053, + "tokens_seen": 220987392 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047122367101303913, + "loss": 1.6571, + "theoretical_loss": 4.329995373382378, + "tokens_seen": 221052928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004712136409227683, + "loss": 1.7547, + "theoretical_loss": 4.329838291880954, + "tokens_seen": 221118464 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047120361083249755, + "loss": 1.8932, + "theoretical_loss": 4.329681269960499, + "tokens_seen": 221184000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711935807422267, + "loss": 1.7836, + "theoretical_loss": 4.329524307580773, + "tokens_seen": 221249536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711835506519559, + "loss": 1.6076, + "theoretical_loss": 4.3293674047015696, + "tokens_seen": 221315072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047117352056168504, + "loss": 1.7504, + "theoretical_loss": 4.329210561282726, + "tokens_seen": 221380608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047116349047141427, + "loss": 1.6752, + "theoretical_loss": 4.329053777284115, + "tokens_seen": 221446144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047115346038114345, + "loss": 1.7427, + "theoretical_loss": 4.328897052665653, + "tokens_seen": 221511680 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047114343029087263, + "loss": 1.8631, + "theoretical_loss": 4.328740387387289, + "tokens_seen": 221577216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711334002006018, + "loss": 1.7122, + "theoretical_loss": 4.328583781409016, + "tokens_seen": 221642752 + }, + { + "epoch": 0.07, + "learning_rate": 0.000471123370110331, + "loss": 1.7631, + "theoretical_loss": 4.328427234690864, + "tokens_seen": 221708288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711133400200602, + "loss": 1.8294, + "theoretical_loss": 4.328270747192901, + "tokens_seen": 221773824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004711033099297894, + "loss": 1.7912, + "theoretical_loss": 4.328114318875236, + "tokens_seen": 221839360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047109327983951854, + "loss": 1.9138, + "theoretical_loss": 4.327957949698012, + "tokens_seen": 221904896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710832497492478, + "loss": 1.8255, + "theoretical_loss": 4.327801639621417, + "tokens_seen": 221970432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710732196589769, + "loss": 1.8212, + "theoretical_loss": 4.3276453886056725, + "tokens_seen": 222035968 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047106318956870614, + "loss": 1.6815, + "theoretical_loss": 4.32748919661104, + "tokens_seen": 222101504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710531594784353, + "loss": 1.7875, + "theoretical_loss": 4.327333063597821, + "tokens_seen": 222167040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710431293881645, + "loss": 1.7333, + "theoretical_loss": 4.327176989526352, + "tokens_seen": 222232576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710330992978937, + "loss": 1.732, + "theoretical_loss": 4.327020974357012, + "tokens_seen": 222298112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710230692076229, + "loss": 1.8275, + "theoretical_loss": 4.326865018050215, + "tokens_seen": 222363648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047101303911735204, + "loss": 1.718, + "theoretical_loss": 4.326709120566415, + "tokens_seen": 222429184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004710030090270813, + "loss": 1.7163, + "theoretical_loss": 4.326553281866103, + "tokens_seen": 222494720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709929789368104, + "loss": 1.7503, + "theoretical_loss": 4.326397501909811, + "tokens_seen": 222560256 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047098294884653964, + "loss": 1.6666, + "theoretical_loss": 4.326241780658104, + "tokens_seen": 222625792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709729187562688, + "loss": 1.7769, + "theoretical_loss": 4.32608611807159, + "tokens_seen": 222691328 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470962888665998, + "loss": 1.6651, + "theoretical_loss": 4.325930514110912, + "tokens_seen": 222756864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 87601, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5978519916534424, + "objective/train/theoretical_loss": 4.325774968736753, + "objective/train/tokens_used": 243282400, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709528585757272, + "loss": 1.6457, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047094282848545636, + "loss": 1.8468, + "theoretical_loss": 4.325619481909833, + "tokens_seen": 222887936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047093279839518555, + "loss": 1.6522, + "theoretical_loss": 4.325464053590908, + "tokens_seen": 222953472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709227683049148, + "loss": 1.7791, + "theoretical_loss": 4.325308683740776, + "tokens_seen": 223019008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004709127382146439, + "loss": 1.8121, + "theoretical_loss": 4.325153372320269, + "tokens_seen": 223084544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047090270812437314, + "loss": 1.7872, + "theoretical_loss": 4.32499811929026, + "tokens_seen": 223150080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047089267803410227, + "loss": 1.8296, + "theoretical_loss": 4.324842924611655, + "tokens_seen": 223215616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708826479438315, + "loss": 1.7746, + "theoretical_loss": 4.324687788245403, + "tokens_seen": 223281152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708726178535607, + "loss": 1.7535, + "theoretical_loss": 4.3245327101524875, + "tokens_seen": 223346688 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047086258776328987, + "loss": 1.7783, + "theoretical_loss": 4.324377690293929, + "tokens_seen": 223412224 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047085255767301905, + "loss": 1.8082, + "theoretical_loss": 4.324222728630787, + "tokens_seen": 223477760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708425275827483, + "loss": 1.794, + "theoretical_loss": 4.324067825124159, + "tokens_seen": 223543296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708324974924774, + "loss": 1.7458, + "theoretical_loss": 4.323912979735178, + "tokens_seen": 223608832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047082246740220665, + "loss": 1.8419, + "theoretical_loss": 4.323758192425016, + "tokens_seen": 223674368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004708124373119358, + "loss": 1.7358, + "theoretical_loss": 4.323603463154882, + "tokens_seen": 223739904 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470802407221665, + "loss": 1.7843, + "theoretical_loss": 4.323448791886022, + "tokens_seen": 223805440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707923771313942, + "loss": 1.9194, + "theoretical_loss": 4.323294178579717, + "tokens_seen": 223870976 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047078234704112337, + "loss": 1.7196, + "theoretical_loss": 4.323139623197289, + "tokens_seen": 223936512 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047077231695085255, + "loss": 1.8418, + "theoretical_loss": 4.322985125700096, + "tokens_seen": 224002048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047076228686058173, + "loss": 1.7168, + "theoretical_loss": 4.322830686049532, + "tokens_seen": 224067584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707522567703109, + "loss": 1.8595, + "theoretical_loss": 4.322676304207027, + "tokens_seen": 224133120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047074222668004015, + "loss": 1.8272, + "theoretical_loss": 4.32252198013405, + "tokens_seen": 224198656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707321965897693, + "loss": 1.7536, + "theoretical_loss": 4.32236771379211, + "tokens_seen": 224264192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707221664994985, + "loss": 1.7721, + "theoretical_loss": 4.322213505142744, + "tokens_seen": 224329728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047071213640922775, + "loss": 1.8087, + "theoretical_loss": 4.322059354147535, + "tokens_seen": 224395264 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004707021063189569, + "loss": 1.6251, + "theoretical_loss": 4.321905260768097, + "tokens_seen": 224460800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706920762286861, + "loss": 1.7732, + "theoretical_loss": 4.321751224966083, + "tokens_seen": 224526336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047068204613841524, + "loss": 1.6766, + "theoretical_loss": 4.321597246703183, + "tokens_seen": 224591872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047067201604814447, + "loss": 1.809, + "theoretical_loss": 4.321443325941123, + "tokens_seen": 224657408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047066198595787365, + "loss": 1.7688, + "theoretical_loss": 4.3212894626416665, + "tokens_seen": 224722944 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047065195586760283, + "loss": 1.6793, + "theoretical_loss": 4.321135656766611, + "tokens_seen": 224788480 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470641925777332, + "loss": 1.7505, + "theoretical_loss": 4.320981908277793, + "tokens_seen": 224854016 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706318956870612, + "loss": 1.867, + "theoretical_loss": 4.320828217137086, + "tokens_seen": 224919552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706218655967904, + "loss": 1.8542, + "theoretical_loss": 4.320674583306396, + "tokens_seen": 224985088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004706118355065196, + "loss": 1.8967, + "theoretical_loss": 4.32052100674767, + "tokens_seen": 225050624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047060180541624874, + "loss": 1.943, + "theoretical_loss": 4.320367487422891, + "tokens_seen": 225116160 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470591775325978, + "loss": 1.8144, + "theoretical_loss": 4.3202140252940735, + "tokens_seen": 225181696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705817452357071, + "loss": 1.7632, + "theoretical_loss": 4.3200606203232725, + "tokens_seen": 225247232 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047057171514543634, + "loss": 1.7951, + "theoretical_loss": 4.319907272472579, + "tokens_seen": 225312768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705616850551655, + "loss": 1.6841, + "theoretical_loss": 4.319753981704119, + "tokens_seen": 225378304 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705516549648947, + "loss": 1.7361, + "theoretical_loss": 4.319600747980056, + "tokens_seen": 225443840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705416248746239, + "loss": 1.7793, + "theoretical_loss": 4.319447571262586, + "tokens_seen": 225509376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705315947843531, + "loss": 1.9502, + "theoretical_loss": 4.3192944515139455, + "tokens_seen": 225574912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047052156469408224, + "loss": 1.8687, + "theoretical_loss": 4.319141388696404, + "tokens_seen": 225640448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705115346038115, + "loss": 1.7161, + "theoretical_loss": 4.318988382772269, + "tokens_seen": 225705984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004705015045135406, + "loss": 1.8266, + "theoretical_loss": 4.318835433703883, + "tokens_seen": 225771520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047049147442326984, + "loss": 1.8798, + "theoretical_loss": 4.318682541453622, + "tokens_seen": 225837056 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470481444332999, + "loss": 1.6821, + "theoretical_loss": 4.318529705983903, + "tokens_seen": 225902592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704714142427282, + "loss": 1.8017, + "theoretical_loss": 4.318376927257174, + "tokens_seen": 225968128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704613841524574, + "loss": 1.7515, + "theoretical_loss": 4.3182242052359205, + "tokens_seen": 226033664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 88830, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7419260740280151, + "objective/train/theoretical_loss": 4.318071539882665, + "objective/train/tokens_used": 246559200, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047045135406218656, + "loss": 1.7333, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047044132397191575, + "loss": 1.8766, + "theoretical_loss": 4.317918931159963, + "tokens_seen": 226164736 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470431293881645, + "loss": 1.6832, + "theoretical_loss": 4.317766379030407, + "tokens_seen": 226230272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004704212637913741, + "loss": 1.7568, + "theoretical_loss": 4.317613883456625, + "tokens_seen": 226295808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047041123370110334, + "loss": 1.7026, + "theoretical_loss": 4.317461444401281, + "tokens_seen": 226361344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047040120361083247, + "loss": 1.8031, + "theoretical_loss": 4.317309061827073, + "tokens_seen": 226426880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703911735205617, + "loss": 1.6504, + "theoretical_loss": 4.317156735696736, + "tokens_seen": 226492416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703811434302909, + "loss": 1.6467, + "theoretical_loss": 4.3170044659730396, + "tokens_seen": 226557952 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047037111334002007, + "loss": 1.7359, + "theoretical_loss": 4.316852252618788, + "tokens_seen": 226623488 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047036108324974925, + "loss": 1.8151, + "theoretical_loss": 4.316700095596822, + "tokens_seen": 226689024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703510531594785, + "loss": 1.7422, + "theoretical_loss": 4.3165479948700165, + "tokens_seen": 226754560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703410230692076, + "loss": 1.6683, + "theoretical_loss": 4.316395950401283, + "tokens_seen": 226820096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047033099297893685, + "loss": 1.7495, + "theoretical_loss": 4.316243962153566, + "tokens_seen": 226885632 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470320962888666, + "loss": 1.732, + "theoretical_loss": 4.316092030089847, + "tokens_seen": 226951168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703109327983952, + "loss": 1.7675, + "theoretical_loss": 4.315940154173143, + "tokens_seen": 227016704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004703009027081244, + "loss": 1.7144, + "theoretical_loss": 4.315788334366504, + "tokens_seen": 227082240 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047029087261785357, + "loss": 1.7035, + "theoretical_loss": 4.315636570633015, + "tokens_seen": 227147776 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047028084252758275, + "loss": 1.7084, + "theoretical_loss": 4.315484862935799, + "tokens_seen": 227213312 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047027081243731193, + "loss": 1.8185, + "theoretical_loss": 4.315333211238009, + "tokens_seen": 227278848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702607823470411, + "loss": 1.6479, + "theoretical_loss": 4.315181615502838, + "tokens_seen": 227344384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047025075225677035, + "loss": 1.6886, + "theoretical_loss": 4.31503007569351, + "tokens_seen": 227409920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702407221664995, + "loss": 1.7694, + "theoretical_loss": 4.314878591773287, + "tokens_seen": 227475456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702306920762287, + "loss": 1.7861, + "theoretical_loss": 4.31472716370546, + "tokens_seen": 227540992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047022066198595784, + "loss": 1.6569, + "theoretical_loss": 4.314575791453362, + "tokens_seen": 227606528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004702106318956871, + "loss": 1.7279, + "theoretical_loss": 4.314424474980355, + "tokens_seen": 227672064 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047020060180541626, + "loss": 1.7448, + "theoretical_loss": 4.31427321424984, + "tokens_seen": 227737600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047019057171514544, + "loss": 1.7814, + "theoretical_loss": 4.314122009225248, + "tokens_seen": 227803136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701805416248746, + "loss": 1.7762, + "theoretical_loss": 4.313970859870048, + "tokens_seen": 227868672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047017051153460385, + "loss": 1.6962, + "theoretical_loss": 4.313819766147741, + "tokens_seen": 227934208 + }, + { + "epoch": 0.07, + "learning_rate": 0.000470160481444333, + "loss": 1.7776, + "theoretical_loss": 4.313668728021866, + "tokens_seen": 227999744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701504513540622, + "loss": 1.908, + "theoretical_loss": 4.313517745455991, + "tokens_seen": 228065280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047014042126379134, + "loss": 1.7057, + "theoretical_loss": 4.313366818413724, + "tokens_seen": 228130816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701303911735206, + "loss": 1.7148, + "theoretical_loss": 4.313215946858702, + "tokens_seen": 228196352 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047012036108324976, + "loss": 1.8105, + "theoretical_loss": 4.313065130754601, + "tokens_seen": 228261888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047011033099297894, + "loss": 1.8294, + "theoretical_loss": 4.312914370065128, + "tokens_seen": 228327424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004701003009027081, + "loss": 1.7868, + "theoretical_loss": 4.312763664754025, + "tokens_seen": 228392960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700902708124373, + "loss": 1.7432, + "theoretical_loss": 4.312613014785069, + "tokens_seen": 228458496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700802407221665, + "loss": 1.8252, + "theoretical_loss": 4.3124624201220705, + "tokens_seen": 228524032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700702106318957, + "loss": 1.7738, + "theoretical_loss": 4.312311880728873, + "tokens_seen": 228589568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047006018054162484, + "loss": 1.79, + "theoretical_loss": 4.312161396569356, + "tokens_seen": 228655104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700501504513541, + "loss": 1.7989, + "theoretical_loss": 4.312010967607431, + "tokens_seen": 228720640 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047004012036108326, + "loss": 1.8145, + "theoretical_loss": 4.311860593807045, + "tokens_seen": 228786176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047003009027081244, + "loss": 1.7728, + "theoretical_loss": 4.311710275132178, + "tokens_seen": 228851712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700200601805416, + "loss": 1.778, + "theoretical_loss": 4.3115600115468435, + "tokens_seen": 228917248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004700100300902708, + "loss": 1.7344, + "theoretical_loss": 4.311409803015089, + "tokens_seen": 228982784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00047, + "loss": 1.8756, + "theoretical_loss": 4.3112596495009985, + "tokens_seen": 229048320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699899699097292, + "loss": 1.7671, + "theoretical_loss": 4.311109550968684, + "tokens_seen": 229113856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046997993981945835, + "loss": 1.8271, + "theoretical_loss": 4.310959507382296, + "tokens_seen": 229179392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699699097291876, + "loss": 1.7516, + "theoretical_loss": 4.310809518706017, + "tokens_seen": 229244928 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046995987963891676, + "loss": 1.8064, + "theoretical_loss": 4.310659584904062, + "tokens_seen": 229310464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 90061, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8844830989837646, + "objective/train/theoretical_loss": 4.310509705940682, + "objective/train/tokens_used": 249836000, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046994984954864595, + "loss": 1.8273, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699398194583752, + "loss": 1.7117, + "theoretical_loss": 4.310359881780159, + "tokens_seen": 229441536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004699297893681043, + "loss": 1.8713, + "theoretical_loss": 4.310210112386811, + "tokens_seen": 229507072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046991975927783354, + "loss": 1.8788, + "theoretical_loss": 4.310060397724987, + "tokens_seen": 229572608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046990972918756267, + "loss": 1.8225, + "theoretical_loss": 4.30991073775907, + "tokens_seen": 229638144 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698996990972919, + "loss": 1.7319, + "theoretical_loss": 4.3097611324534775, + "tokens_seen": 229703680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698896690070211, + "loss": 1.8443, + "theoretical_loss": 4.309611581772659, + "tokens_seen": 229769216 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046987963891675027, + "loss": 1.8026, + "theoretical_loss": 4.309462085681098, + "tokens_seen": 229834752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046986960882647945, + "loss": 1.8862, + "theoretical_loss": 4.3093126441433105, + "tokens_seen": 229900288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698595787362087, + "loss": 1.7815, + "theoretical_loss": 4.3091632571238465, + "tokens_seen": 229965824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698495486459378, + "loss": 1.8163, + "theoretical_loss": 4.30901392458729, + "tokens_seen": 230031360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046983951855566705, + "loss": 1.7786, + "theoretical_loss": 4.308864646498254, + "tokens_seen": 230096896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698294884653962, + "loss": 1.7199, + "theoretical_loss": 4.30871542282139, + "tokens_seen": 230162432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698194583751254, + "loss": 1.6075, + "theoretical_loss": 4.308566253521378, + "tokens_seen": 230227968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004698094282848546, + "loss": 1.7438, + "theoretical_loss": 4.308417138562934, + "tokens_seen": 230293504 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046979939819458377, + "loss": 1.6551, + "theoretical_loss": 4.308268077910806, + "tokens_seen": 230359040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046978936810431295, + "loss": 1.6183, + "theoretical_loss": 4.308119071529774, + "tokens_seen": 230424576 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046977933801404213, + "loss": 1.6842, + "theoretical_loss": 4.3079701193846525, + "tokens_seen": 230490112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697693079237713, + "loss": 1.9469, + "theoretical_loss": 4.307821221440286, + "tokens_seen": 230555648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046975927783350055, + "loss": 1.7384, + "theoretical_loss": 4.3076723776615555, + "tokens_seen": 230621184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697492477432297, + "loss": 1.7963, + "theoretical_loss": 4.307523588013371, + "tokens_seen": 230686720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697392176529589, + "loss": 1.7302, + "theoretical_loss": 4.307374852460679, + "tokens_seen": 230752256 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046972918756268804, + "loss": 1.6983, + "theoretical_loss": 4.307226170968456, + "tokens_seen": 230817792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004697191574724173, + "loss": 1.6656, + "theoretical_loss": 4.307077543501713, + "tokens_seen": 230883328 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046970912738214646, + "loss": 1.8017, + "theoretical_loss": 4.3069289700254885, + "tokens_seen": 230948864 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046969909729187564, + "loss": 1.6961, + "theoretical_loss": 4.306780450504861, + "tokens_seen": 231014400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696890672016048, + "loss": 1.7636, + "theoretical_loss": 4.3066319849049375, + "tokens_seen": 231079936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046967903711133405, + "loss": 1.8597, + "theoretical_loss": 4.306483573190857, + "tokens_seen": 231145472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696690070210632, + "loss": 1.7768, + "theoretical_loss": 4.306335215327791, + "tokens_seen": 231211008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696589769307924, + "loss": 1.7836, + "theoretical_loss": 4.306186911280945, + "tokens_seen": 231276544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046964894684052154, + "loss": 1.6429, + "theoretical_loss": 4.306038661015557, + "tokens_seen": 231342080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696389167502508, + "loss": 1.67, + "theoretical_loss": 4.305890464496895, + "tokens_seen": 231407616 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046962888665997996, + "loss": 1.914, + "theoretical_loss": 4.305742321690261, + "tokens_seen": 231473152 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046961885656970914, + "loss": 1.7408, + "theoretical_loss": 4.305594232560988, + "tokens_seen": 231538688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004696088264794383, + "loss": 1.8727, + "theoretical_loss": 4.305446197074443, + "tokens_seen": 231604224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695987963891675, + "loss": 1.7109, + "theoretical_loss": 4.305298215196023, + "tokens_seen": 231669760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695887662988967, + "loss": 1.779, + "theoretical_loss": 4.305150286891159, + "tokens_seen": 231735296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695787362086259, + "loss": 1.9012, + "theoretical_loss": 4.3050024121253125, + "tokens_seen": 231800832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046956870611835505, + "loss": 1.6472, + "theoretical_loss": 4.304854590863979, + "tokens_seen": 231866368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695586760280843, + "loss": 1.6773, + "theoretical_loss": 4.304706823072683, + "tokens_seen": 231931904 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046954864593781346, + "loss": 1.9157, + "theoretical_loss": 4.304559108716984, + "tokens_seen": 231997440 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046953861584754264, + "loss": 1.8582, + "theoretical_loss": 4.304411447762472, + "tokens_seen": 232062976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695285857572718, + "loss": 1.7488, + "theoretical_loss": 4.304263840174768, + "tokens_seen": 232128512 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469518555667001, + "loss": 1.6057, + "theoretical_loss": 4.304116285919527, + "tokens_seen": 232194048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004695085255767302, + "loss": 1.7725, + "theoretical_loss": 4.303968784962433, + "tokens_seen": 232259584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694984954864594, + "loss": 1.738, + "theoretical_loss": 4.303821337269205, + "tokens_seen": 232325120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046948846539618855, + "loss": 1.7825, + "theoretical_loss": 4.303673942805591, + "tokens_seen": 232390656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694784353059178, + "loss": 1.7641, + "theoretical_loss": 4.303526601537371, + "tokens_seen": 232456192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694684052156469, + "loss": 1.8521, + "theoretical_loss": 4.303379313430359, + "tokens_seen": 232521728 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046945837512537615, + "loss": 1.6732, + "theoretical_loss": 4.303232078450399, + "tokens_seen": 232587264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 91258, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6689767837524414, + "objective/train/theoretical_loss": 4.303084896563364, + "objective/train/tokens_used": 253112800, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694483450351053, + "loss": 1.7642, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694383149448345, + "loss": 1.6466, + "theoretical_loss": 4.302937767735164, + "tokens_seen": 232718336 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004694282848545637, + "loss": 1.7234, + "theoretical_loss": 4.302790691931736, + "tokens_seen": 232783872 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046941825476429287, + "loss": 1.7623, + "theoretical_loss": 4.30264366911905, + "tokens_seen": 232849408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046940822467402205, + "loss": 1.7947, + "theoretical_loss": 4.3024966992631075, + "tokens_seen": 232914944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693981945837513, + "loss": 1.7311, + "theoretical_loss": 4.302349782329942, + "tokens_seen": 232980480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693881644934804, + "loss": 1.7748, + "theoretical_loss": 4.302202918285616, + "tokens_seen": 233046016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046937813440320965, + "loss": 1.8001, + "theoretical_loss": 4.302056107096226, + "tokens_seen": 233111552 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046936810431293883, + "loss": 1.6836, + "theoretical_loss": 4.301909348727899, + "tokens_seen": 233177088 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469358074222668, + "loss": 1.7402, + "theoretical_loss": 4.301762643146792, + "tokens_seen": 233242624 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693480441323972, + "loss": 1.8318, + "theoretical_loss": 4.301615990319094, + "tokens_seen": 233308160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693380140421264, + "loss": 1.7482, + "theoretical_loss": 4.301469390211025, + "tokens_seen": 233373696 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046932798395185555, + "loss": 1.8719, + "theoretical_loss": 4.301322842788839, + "tokens_seen": 233439232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693179538615848, + "loss": 1.6966, + "theoretical_loss": 4.301176348018814, + "tokens_seen": 233504768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004693079237713139, + "loss": 1.7143, + "theoretical_loss": 4.301029905867265, + "tokens_seen": 233570304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046929789368104315, + "loss": 1.6445, + "theoretical_loss": 4.300883516300539, + "tokens_seen": 233635840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692878635907723, + "loss": 1.692, + "theoretical_loss": 4.300737179285008, + "tokens_seen": 233701376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692778335005015, + "loss": 1.7335, + "theoretical_loss": 4.30059089478708, + "tokens_seen": 233766912 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692678034102307, + "loss": 1.7732, + "theoretical_loss": 4.300444662773192, + "tokens_seen": 233832448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692577733199599, + "loss": 1.6952, + "theoretical_loss": 4.30029848320981, + "tokens_seen": 233897984 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046924774322968906, + "loss": 1.8049, + "theoretical_loss": 4.300152356063435, + "tokens_seen": 233963520 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046923771313941824, + "loss": 1.8202, + "theoretical_loss": 4.300006281300598, + "tokens_seen": 234029056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004692276830491474, + "loss": 1.7421, + "theoretical_loss": 4.299860258887856, + "tokens_seen": 234094592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046921765295887666, + "loss": 1.9, + "theoretical_loss": 4.299714288791802, + "tokens_seen": 234160128 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046920762286860584, + "loss": 1.5952, + "theoretical_loss": 4.299568370979056, + "tokens_seen": 234225664 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469197592778335, + "loss": 1.835, + "theoretical_loss": 4.299422505416272, + "tokens_seen": 234291200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046918756268806425, + "loss": 1.8109, + "theoretical_loss": 4.2992766920701335, + "tokens_seen": 234356736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691775325977934, + "loss": 1.7746, + "theoretical_loss": 4.299130930907352, + "tokens_seen": 234422272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691675025075226, + "loss": 1.7552, + "theoretical_loss": 4.2989852218946725, + "tokens_seen": 234487808 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046915747241725174, + "loss": 1.7784, + "theoretical_loss": 4.298839564998869, + "tokens_seen": 234553344 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469147442326981, + "loss": 1.8844, + "theoretical_loss": 4.298693960186747, + "tokens_seen": 234618880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046913741223671016, + "loss": 1.8388, + "theoretical_loss": 4.298548407425141, + "tokens_seen": 234684416 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046912738214643934, + "loss": 1.6763, + "theoretical_loss": 4.298402906680918, + "tokens_seen": 234749952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691173520561685, + "loss": 1.7606, + "theoretical_loss": 4.298257457920973, + "tokens_seen": 234815488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004691073219658977, + "loss": 1.7795, + "theoretical_loss": 4.298112061112232, + "tokens_seen": 234881024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690972918756269, + "loss": 1.8359, + "theoretical_loss": 4.297966716221653, + "tokens_seen": 234946560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690872617853561, + "loss": 1.8068, + "theoretical_loss": 4.297821423216222, + "tokens_seen": 235012096 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046907723169508525, + "loss": 1.7475, + "theoretical_loss": 4.297676182062955, + "tokens_seen": 235077632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690672016048145, + "loss": 1.7601, + "theoretical_loss": 4.297530992728901, + "tokens_seen": 235143168 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046905717151454366, + "loss": 1.755, + "theoretical_loss": 4.297385855181136, + "tokens_seen": 235208704 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046904714142427284, + "loss": 1.7497, + "theoretical_loss": 4.297240769386768, + "tokens_seen": 235274240 + }, + { + "epoch": 0.07, + "learning_rate": 0.000469037111334002, + "loss": 1.8065, + "theoretical_loss": 4.297095735312934, + "tokens_seen": 235339776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690270812437312, + "loss": 1.6931, + "theoretical_loss": 4.296950752926801, + "tokens_seen": 235405312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690170511534604, + "loss": 1.715, + "theoretical_loss": 4.296805822195568, + "tokens_seen": 235470848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004690070210631896, + "loss": 1.7496, + "theoretical_loss": 4.29666094308646, + "tokens_seen": 235536384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046899699097291875, + "loss": 1.7815, + "theoretical_loss": 4.296516115566736, + "tokens_seen": 235601920 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468986960882648, + "loss": 1.8766, + "theoretical_loss": 4.296371339603684, + "tokens_seen": 235667456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689769307923771, + "loss": 1.7151, + "theoretical_loss": 4.296226615164618, + "tokens_seen": 235732992 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046896690070210635, + "loss": 1.7953, + "theoretical_loss": 4.296081942216887, + "tokens_seen": 235798528 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046895687061183553, + "loss": 1.6726, + "theoretical_loss": 4.295937320727866, + "tokens_seen": 235864064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 92471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.865220069885254, + "objective/train/theoretical_loss": 4.295792750664964, + "objective/train/tokens_used": 256389600, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689468405215647, + "loss": 1.7942, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689368104312939, + "loss": 1.7811, + "theoretical_loss": 4.295648231995613, + "tokens_seen": 235995136 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046892678034102307, + "loss": 1.8832, + "theoretical_loss": 4.295503764687281, + "tokens_seen": 236060672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046891675025075225, + "loss": 1.7013, + "theoretical_loss": 4.295359348707462, + "tokens_seen": 236126208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004689067201604815, + "loss": 1.7014, + "theoretical_loss": 4.2952149840236835, + "tokens_seen": 236191744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688966900702106, + "loss": 1.8071, + "theoretical_loss": 4.295070670603496, + "tokens_seen": 236257280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046888665997993985, + "loss": 1.6828, + "theoretical_loss": 4.294926408414486, + "tokens_seen": 236322816 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046887662988966903, + "loss": 1.7212, + "theoretical_loss": 4.294782197424265, + "tokens_seen": 236388352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688665997993982, + "loss": 1.7316, + "theoretical_loss": 4.294638037600478, + "tokens_seen": 236453888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688565697091274, + "loss": 1.8254, + "theoretical_loss": 4.2944939289107955, + "tokens_seen": 236519424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688465396188566, + "loss": 1.8394, + "theoretical_loss": 4.29434987132292, + "tokens_seen": 236584960 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046883650952858575, + "loss": 1.9349, + "theoretical_loss": 4.294205864804581, + "tokens_seen": 236650496 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468826479438315, + "loss": 1.6966, + "theoretical_loss": 4.294061909323539, + "tokens_seen": 236716032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004688164493480441, + "loss": 1.7923, + "theoretical_loss": 4.293918004847585, + "tokens_seen": 236781568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046880641925777335, + "loss": 1.7679, + "theoretical_loss": 4.293774151344537, + "tokens_seen": 236847104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687963891675025, + "loss": 1.8514, + "theoretical_loss": 4.2936303487822425, + "tokens_seen": 236912640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687863590772317, + "loss": 1.7414, + "theoretical_loss": 4.293486597128578, + "tokens_seen": 236978176 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687763289869609, + "loss": 1.6915, + "theoretical_loss": 4.293342896351452, + "tokens_seen": 237043712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687662988966901, + "loss": 1.7218, + "theoretical_loss": 4.293199246418796, + "tokens_seen": 237109248 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046875626880641926, + "loss": 1.7456, + "theoretical_loss": 4.293055647298578, + "tokens_seen": 237174784 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046874623871614844, + "loss": 1.8056, + "theoretical_loss": 4.292912098958791, + "tokens_seen": 237240320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687362086258776, + "loss": 1.7232, + "theoretical_loss": 4.292768601367455, + "tokens_seen": 237305856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046872617853560686, + "loss": 1.6786, + "theoretical_loss": 4.292625154492623, + "tokens_seen": 237371392 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468716148445336, + "loss": 1.6877, + "theoretical_loss": 4.292481758302375, + "tokens_seen": 237436928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004687061183550652, + "loss": 1.6738, + "theoretical_loss": 4.2923384127648205, + "tokens_seen": 237502464 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686960882647944, + "loss": 1.649, + "theoretical_loss": 4.292195117848098, + "tokens_seen": 237568000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686860581745236, + "loss": 1.7228, + "theoretical_loss": 4.292051873520373, + "tokens_seen": 237633536 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046867602808425276, + "loss": 1.7265, + "theoretical_loss": 4.291908679749842, + "tokens_seen": 237699072 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046866599799398194, + "loss": 1.888, + "theoretical_loss": 4.2917655365047285, + "tokens_seen": 237764608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686559679037111, + "loss": 1.6602, + "theoretical_loss": 4.291622443753287, + "tokens_seen": 237830144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046864593781344036, + "loss": 1.8181, + "theoretical_loss": 4.291479401463799, + "tokens_seen": 237895680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686359077231695, + "loss": 1.7468, + "theoretical_loss": 4.291336409604574, + "tokens_seen": 237961216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686258776328987, + "loss": 1.8475, + "theoretical_loss": 4.291193468143952, + "tokens_seen": 238026752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046861584754262785, + "loss": 1.8104, + "theoretical_loss": 4.2910505770503, + "tokens_seen": 238092288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004686058174523571, + "loss": 1.7211, + "theoretical_loss": 4.290907736292016, + "tokens_seen": 238157824 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046859578736208626, + "loss": 1.6934, + "theoretical_loss": 4.290764945837522, + "tokens_seen": 238223360 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046858575727181545, + "loss": 1.7888, + "theoretical_loss": 4.290622205655273, + "tokens_seen": 238288896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685757271815446, + "loss": 1.6595, + "theoretical_loss": 4.290479515713751, + "tokens_seen": 238354432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685656970912738, + "loss": 1.6757, + "theoretical_loss": 4.290336875981465, + "tokens_seen": 238419968 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468555667001003, + "loss": 1.677, + "theoretical_loss": 4.290194286426955, + "tokens_seen": 238485504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685456369107322, + "loss": 1.8435, + "theoretical_loss": 4.290051747018785, + "tokens_seen": 238551040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046853560682046135, + "loss": 1.8452, + "theoretical_loss": 4.289909257725553, + "tokens_seen": 238616576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004685255767301906, + "loss": 1.7591, + "theoretical_loss": 4.289766818515882, + "tokens_seen": 238682112 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046851554663991977, + "loss": 1.8861, + "theoretical_loss": 4.2896244293584225, + "tokens_seen": 238747648 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046850551654964895, + "loss": 1.7665, + "theoretical_loss": 4.2894820902218544, + "tokens_seen": 238813184 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046849548645937813, + "loss": 1.7108, + "theoretical_loss": 4.289339801074888, + "tokens_seen": 238878720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684854563691073, + "loss": 1.7128, + "theoretical_loss": 4.289197561886258, + "tokens_seen": 238944256 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684754262788365, + "loss": 1.8543, + "theoretical_loss": 4.289055372624729, + "tokens_seen": 239009792 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046846539618856573, + "loss": 1.7916, + "theoretical_loss": 4.288913233259092, + "tokens_seen": 239075328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684553660982949, + "loss": 1.7389, + "theoretical_loss": 4.28877114375817, + "tokens_seen": 239140864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 93570, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8372766971588135, + "objective/train/theoretical_loss": 4.288629104090811, + "objective/train/tokens_used": 259666400, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684453360080241, + "loss": 1.8221, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046843530591775327, + "loss": 1.7441, + "theoretical_loss": 4.28848711422589, + "tokens_seen": 239271936 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046842527582748245, + "loss": 1.735, + "theoretical_loss": 4.288345174132313, + "tokens_seen": 239337472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684152457372117, + "loss": 1.8853, + "theoretical_loss": 4.288203283779012, + "tokens_seen": 239403008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004684052156469408, + "loss": 1.8019, + "theoretical_loss": 4.288061443134947, + "tokens_seen": 239468544 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046839518555667005, + "loss": 1.8089, + "theoretical_loss": 4.287919652169107, + "tokens_seen": 239534080 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046838515546639923, + "loss": 1.6598, + "theoretical_loss": 4.287777910850507, + "tokens_seen": 239599616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683751253761284, + "loss": 1.7456, + "theoretical_loss": 4.287636219148191, + "tokens_seen": 239665152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683650952858576, + "loss": 1.8097, + "theoretical_loss": 4.287494577031231, + "tokens_seen": 239730688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683550651955868, + "loss": 1.7664, + "theoretical_loss": 4.287352984468727, + "tokens_seen": 239796224 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046834503510531595, + "loss": 1.8247, + "theoretical_loss": 4.287211441429805, + "tokens_seen": 239861760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683350050150452, + "loss": 1.8351, + "theoretical_loss": 4.2870699478836185, + "tokens_seen": 239927296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683249749247743, + "loss": 1.7484, + "theoretical_loss": 4.286928503799352, + "tokens_seen": 239992832 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046831494483450355, + "loss": 1.6657, + "theoretical_loss": 4.286787109146213, + "tokens_seen": 240058368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004683049147442327, + "loss": 1.7811, + "theoretical_loss": 4.286645763893442, + "tokens_seen": 240123904 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682948846539619, + "loss": 1.8011, + "theoretical_loss": 4.2865044680103015, + "tokens_seen": 240189440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682848545636911, + "loss": 1.849, + "theoretical_loss": 4.286363221466084, + "tokens_seen": 240254976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682748244734203, + "loss": 1.7761, + "theoretical_loss": 4.286222024230111, + "tokens_seen": 240320512 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046826479438314946, + "loss": 1.6687, + "theoretical_loss": 4.286080876271729, + "tokens_seen": 240386048 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046825476429287864, + "loss": 1.7366, + "theoretical_loss": 4.285939777560312, + "tokens_seen": 240451584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682447342026078, + "loss": 1.8503, + "theoretical_loss": 4.285798728065264, + "tokens_seen": 240517120 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046823470411233706, + "loss": 1.8351, + "theoretical_loss": 4.285657727756012, + "tokens_seen": 240582656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682246740220662, + "loss": 1.8559, + "theoretical_loss": 4.285516776602016, + "tokens_seen": 240648192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682146439317954, + "loss": 1.7073, + "theoretical_loss": 4.2853758745727575, + "tokens_seen": 240713728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004682046138415246, + "loss": 1.6188, + "theoretical_loss": 4.285235021637748, + "tokens_seen": 240779264 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681945837512538, + "loss": 1.6775, + "theoretical_loss": 4.285094217766527, + "tokens_seen": 240844800 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046818455366098296, + "loss": 1.6866, + "theoretical_loss": 4.2849534629286605, + "tokens_seen": 240910336 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046817452357071214, + "loss": 1.7361, + "theoretical_loss": 4.284812757093741, + "tokens_seen": 240975872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681644934804413, + "loss": 1.6433, + "theoretical_loss": 4.284672100231388, + "tokens_seen": 241041408 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046815446339017056, + "loss": 1.7946, + "theoretical_loss": 4.284531492311248, + "tokens_seen": 241106944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681444332998997, + "loss": 1.6748, + "theoretical_loss": 4.284390933302997, + "tokens_seen": 241172480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681344032096289, + "loss": 1.7032, + "theoretical_loss": 4.284250423176335, + "tokens_seen": 241238016 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046812437311935805, + "loss": 1.7192, + "theoretical_loss": 4.28410996190099, + "tokens_seen": 241303552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004681143430290873, + "loss": 1.7434, + "theoretical_loss": 4.283969549446717, + "tokens_seen": 241369088 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046810431293881646, + "loss": 1.8071, + "theoretical_loss": 4.2838291857833, + "tokens_seen": 241434624 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046809428284854565, + "loss": 1.7661, + "theoretical_loss": 4.283688870880545, + "tokens_seen": 241500160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680842527582748, + "loss": 1.6525, + "theoretical_loss": 4.28354860470829, + "tokens_seen": 241565696 + }, + { + "epoch": 0.07, + "learning_rate": 0.000468074222668004, + "loss": 1.7428, + "theoretical_loss": 4.283408387236397, + "tokens_seen": 241631232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680641925777332, + "loss": 1.7331, + "theoretical_loss": 4.283268218434755, + "tokens_seen": 241696768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680541624874624, + "loss": 1.7032, + "theoretical_loss": 4.283128098273281, + "tokens_seen": 241762304 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046804413239719155, + "loss": 1.7226, + "theoretical_loss": 4.282988026721917, + "tokens_seen": 241827840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004680341023069208, + "loss": 1.607, + "theoretical_loss": 4.282848003750634, + "tokens_seen": 241893376 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046802407221664997, + "loss": 1.7131, + "theoretical_loss": 4.282708029329427, + "tokens_seen": 241958912 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046801404212637915, + "loss": 1.811, + "theoretical_loss": 4.2825681034283205, + "tokens_seen": 242024448 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046800401203610833, + "loss": 1.8229, + "theoretical_loss": 4.2824282260173625, + "tokens_seen": 242089984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679939819458375, + "loss": 1.8511, + "theoretical_loss": 4.282288397066631, + "tokens_seen": 242155520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679839518555667, + "loss": 1.7403, + "theoretical_loss": 4.282148616546227, + "tokens_seen": 242221056 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046797392176529593, + "loss": 1.7163, + "theoretical_loss": 4.282008884426281, + "tokens_seen": 242286592 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046796389167502505, + "loss": 1.5831, + "theoretical_loss": 4.281869200676948, + "tokens_seen": 242352128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679538615847543, + "loss": 1.8088, + "theoretical_loss": 4.281729565268412, + "tokens_seen": 242417664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 94878, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7779464721679688, + "objective/train/theoretical_loss": 4.28158997817088, + "objective/train/tokens_used": 262943200, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679438314944834, + "loss": 1.7374, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046793380140421265, + "loss": 1.77, + "theoretical_loss": 4.281450439354587, + "tokens_seen": 242548736 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046792377131394183, + "loss": 1.7439, + "theoretical_loss": 4.281310948789796, + "tokens_seen": 242614272 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467913741223671, + "loss": 1.6673, + "theoretical_loss": 4.2811715064467935, + "tokens_seen": 242679808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004679037111334002, + "loss": 1.7833, + "theoretical_loss": 4.281032112295894, + "tokens_seen": 242745344 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046789368104312943, + "loss": 1.7438, + "theoretical_loss": 4.2808927663074385, + "tokens_seen": 242810880 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046788365095285856, + "loss": 1.7403, + "theoretical_loss": 4.280753468451793, + "tokens_seen": 242876416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678736208625878, + "loss": 1.6936, + "theoretical_loss": 4.28061421869935, + "tokens_seen": 242941952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678635907723169, + "loss": 1.7599, + "theoretical_loss": 4.28047501702053, + "tokens_seen": 243007488 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046785356068204616, + "loss": 1.7525, + "theoretical_loss": 4.280335863385778, + "tokens_seen": 243073024 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046784353059177534, + "loss": 1.671, + "theoretical_loss": 4.280196757765564, + "tokens_seen": 243138560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678335005015045, + "loss": 1.6893, + "theoretical_loss": 4.280057700130388, + "tokens_seen": 243204096 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678234704112337, + "loss": 1.7426, + "theoretical_loss": 4.279918690450771, + "tokens_seen": 243269632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004678134403209629, + "loss": 1.7584, + "theoretical_loss": 4.279779728697263, + "tokens_seen": 243335168 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046780341023069206, + "loss": 1.7437, + "theoretical_loss": 4.279640814840441, + "tokens_seen": 243400704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677933801404213, + "loss": 1.7288, + "theoretical_loss": 4.279501948850905, + "tokens_seen": 243466240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677833500501504, + "loss": 1.6896, + "theoretical_loss": 4.279363130699284, + "tokens_seen": 243531776 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046777331995987966, + "loss": 1.8287, + "theoretical_loss": 4.279224360356231, + "tokens_seen": 243597312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677632898696088, + "loss": 1.8024, + "theoretical_loss": 4.279085637792425, + "tokens_seen": 243662848 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467753259779338, + "loss": 1.8412, + "theoretical_loss": 4.278946962978573, + "tokens_seen": 243728384 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677432296890672, + "loss": 1.7531, + "theoretical_loss": 4.278808335885403, + "tokens_seen": 243793920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677331995987964, + "loss": 1.703, + "theoretical_loss": 4.278669756483674, + "tokens_seen": 243859456 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046772316950852556, + "loss": 1.7639, + "theoretical_loss": 4.278531224744167, + "tokens_seen": 243924992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004677131394182548, + "loss": 1.7102, + "theoretical_loss": 4.278392740637694, + "tokens_seen": 243990528 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467703109327984, + "loss": 1.7373, + "theoretical_loss": 4.278254304135085, + "tokens_seen": 244056064 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046769307923771316, + "loss": 1.8409, + "theoretical_loss": 4.2781159152072, + "tokens_seen": 244121600 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046768304914744234, + "loss": 1.7234, + "theoretical_loss": 4.277977573824928, + "tokens_seen": 244187136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676730190571715, + "loss": 1.7673, + "theoretical_loss": 4.277839279959177, + "tokens_seen": 244252672 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046766298896690076, + "loss": 1.6539, + "theoretical_loss": 4.277701033580884, + "tokens_seen": 244318208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676529588766299, + "loss": 1.5551, + "theoretical_loss": 4.277562834661013, + "tokens_seen": 244383744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676429287863591, + "loss": 1.7054, + "theoretical_loss": 4.277424683170549, + "tokens_seen": 244449280 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046763289869608825, + "loss": 1.6691, + "theoretical_loss": 4.277286579080508, + "tokens_seen": 244514816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004676228686058175, + "loss": 1.7285, + "theoretical_loss": 4.277148522361927, + "tokens_seen": 244580352 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046761283851554666, + "loss": 1.7075, + "theoretical_loss": 4.277010512985871, + "tokens_seen": 244645888 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046760280842527585, + "loss": 1.8156, + "theoretical_loss": 4.2768725509234296, + "tokens_seen": 244711424 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467592778335005, + "loss": 1.6701, + "theoretical_loss": 4.276734636145717, + "tokens_seen": 244776960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675827482447342, + "loss": 1.8362, + "theoretical_loss": 4.276596768623875, + "tokens_seen": 244842496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675727181544634, + "loss": 1.7328, + "theoretical_loss": 4.276458948329068, + "tokens_seen": 244908032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675626880641926, + "loss": 1.7706, + "theoretical_loss": 4.276321175232487, + "tokens_seen": 244973568 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046755265797392175, + "loss": 1.6684, + "theoretical_loss": 4.276183449305348, + "tokens_seen": 245039104 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467542627883651, + "loss": 1.8219, + "theoretical_loss": 4.276045770518895, + "tokens_seen": 245104640 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046753259779338017, + "loss": 1.9357, + "theoretical_loss": 4.275908138844392, + "tokens_seen": 245170176 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046752256770310935, + "loss": 1.7096, + "theoretical_loss": 4.275770554253132, + "tokens_seen": 245235712 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046751253761283853, + "loss": 1.6805, + "theoretical_loss": 4.275633016716432, + "tokens_seen": 245301248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004675025075225677, + "loss": 1.7709, + "theoretical_loss": 4.275495526205635, + "tokens_seen": 245366784 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674924774322969, + "loss": 1.7258, + "theoretical_loss": 4.275358082692106, + "tokens_seen": 245432320 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046748244734202613, + "loss": 1.7807, + "theoretical_loss": 4.275220686147241, + "tokens_seen": 245497856 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046747241725175525, + "loss": 1.6143, + "theoretical_loss": 4.275083336542455, + "tokens_seen": 245563392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674623871614845, + "loss": 1.7359, + "theoretical_loss": 4.274946033849189, + "tokens_seen": 245628928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674523570712136, + "loss": 1.7242, + "theoretical_loss": 4.2748087780389135, + "tokens_seen": 245694464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 96014, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8354618549346924, + "objective/train/theoretical_loss": 4.27467156908312, + "objective/train/tokens_used": 266220000, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046744232698094285, + "loss": 1.7114, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046743229689067203, + "loss": 1.7216, + "theoretical_loss": 4.274534406953325, + "tokens_seen": 245825536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674222668004012, + "loss": 1.717, + "theoretical_loss": 4.274397291621073, + "tokens_seen": 245891072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004674122367101304, + "loss": 1.6857, + "theoretical_loss": 4.274260223057928, + "tokens_seen": 245956608 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046740220661985963, + "loss": 1.7827, + "theoretical_loss": 4.274123201235483, + "tokens_seen": 246022144 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046739217652958876, + "loss": 1.7393, + "theoretical_loss": 4.273986226125356, + "tokens_seen": 246087680 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467382146439318, + "loss": 1.5358, + "theoretical_loss": 4.273849297699186, + "tokens_seen": 246153216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673721163490471, + "loss": 1.795, + "theoretical_loss": 4.273712415928643, + "tokens_seen": 246218752 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046736208625877636, + "loss": 1.7781, + "theoretical_loss": 4.273575580785415, + "tokens_seen": 246284288 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046735205616850554, + "loss": 1.7755, + "theoretical_loss": 4.27343879224122, + "tokens_seen": 246349824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673420260782347, + "loss": 1.7681, + "theoretical_loss": 4.273302050267796, + "tokens_seen": 246415360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673319959879639, + "loss": 1.806, + "theoretical_loss": 4.27316535483691, + "tokens_seen": 246480896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673219658976931, + "loss": 1.7504, + "theoretical_loss": 4.273028705920351, + "tokens_seen": 246546432 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046731193580742226, + "loss": 1.6896, + "theoretical_loss": 4.272892103489934, + "tokens_seen": 246611968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004673019057171515, + "loss": 1.7387, + "theoretical_loss": 4.272755547517495, + "tokens_seen": 246677504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672918756268806, + "loss": 1.6291, + "theoretical_loss": 4.272619037974902, + "tokens_seen": 246743040 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046728184553660986, + "loss": 1.6365, + "theoretical_loss": 4.272482574834039, + "tokens_seen": 246808576 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467271815446339, + "loss": 1.6585, + "theoretical_loss": 4.272346158066821, + "tokens_seen": 246874112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672617853560682, + "loss": 1.8164, + "theoretical_loss": 4.272209787645183, + "tokens_seen": 246939648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672517552657974, + "loss": 1.743, + "theoretical_loss": 4.272073463541087, + "tokens_seen": 247005184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672417251755266, + "loss": 1.6828, + "theoretical_loss": 4.271937185726519, + "tokens_seen": 247070720 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046723169508525576, + "loss": 1.7311, + "theoretical_loss": 4.2718009541734885, + "tokens_seen": 247136256 + }, + { + "epoch": 0.07, + "learning_rate": 0.000467221664994985, + "loss": 1.6964, + "theoretical_loss": 4.27166476885403, + "tokens_seen": 247201792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004672116349047141, + "loss": 1.841, + "theoretical_loss": 4.271528629740201, + "tokens_seen": 247267328 + }, + { + "epoch": 0.07, + "learning_rate": 0.00046720160481444336, + "loss": 1.7587, + "theoretical_loss": 4.271392536804088, + "tokens_seen": 247332864 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004671915747241725, + "loss": 1.7501, + "theoretical_loss": 4.2712564900177945, + "tokens_seen": 247398400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0004671815446339017, + "loss": 1.7155, + "theoretical_loss": 4.271120489353454, + "tokens_seen": 247463936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671715145436309, + "loss": 1.6635, + "theoretical_loss": 4.270984534783221, + "tokens_seen": 247529472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671614844533601, + "loss": 1.7568, + "theoretical_loss": 4.270848626279275, + "tokens_seen": 247595008 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046715145436308927, + "loss": 1.757, + "theoretical_loss": 4.270712763813822, + "tokens_seen": 247660544 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046714142427281845, + "loss": 1.7619, + "theoretical_loss": 4.270576947359088, + "tokens_seen": 247726080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046713139418254763, + "loss": 1.7809, + "theoretical_loss": 4.270441176887327, + "tokens_seen": 247791616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046712136409227686, + "loss": 1.7185, + "theoretical_loss": 4.270305452370814, + "tokens_seen": 247857152 + }, + { + "epoch": 0.08, + "learning_rate": 0.000467111334002006, + "loss": 1.664, + "theoretical_loss": 4.27016977378185, + "tokens_seen": 247922688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004671013039117352, + "loss": 1.8179, + "theoretical_loss": 4.270034141092759, + "tokens_seen": 247988224 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046709127382146435, + "loss": 1.7353, + "theoretical_loss": 4.269898554275889, + "tokens_seen": 248053760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670812437311936, + "loss": 1.7081, + "theoretical_loss": 4.269763013303613, + "tokens_seen": 248119296 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046707121364092277, + "loss": 1.6895, + "theoretical_loss": 4.269627518148326, + "tokens_seen": 248184832 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046706118355065195, + "loss": 1.7299, + "theoretical_loss": 4.26949206878245, + "tokens_seen": 248250368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046705115346038113, + "loss": 1.8317, + "theoretical_loss": 4.269356665178426, + "tokens_seen": 248315904 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046704112337011037, + "loss": 1.9058, + "theoretical_loss": 4.269221307308725, + "tokens_seen": 248381440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670310932798395, + "loss": 1.7984, + "theoretical_loss": 4.269085995145837, + "tokens_seen": 248446976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046702106318956873, + "loss": 1.7507, + "theoretical_loss": 4.268950728662278, + "tokens_seen": 248512512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046701103309929786, + "loss": 2.0917, + "theoretical_loss": 4.268815507830587, + "tokens_seen": 248578048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004670010030090271, + "loss": 1.6631, + "theoretical_loss": 4.268680332623328, + "tokens_seen": 248643584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669909729187563, + "loss": 1.675, + "theoretical_loss": 4.268545203013085, + "tokens_seen": 248709120 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046698094282848545, + "loss": 1.7378, + "theoretical_loss": 4.26841011897247, + "tokens_seen": 248774656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046697091273821464, + "loss": 1.8384, + "theoretical_loss": 4.268275080474118, + "tokens_seen": 248840192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669608826479438, + "loss": 1.7106, + "theoretical_loss": 4.268140087490686, + "tokens_seen": 248905728 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046695085255767305, + "loss": 1.8061, + "theoretical_loss": 4.2680051399948535, + "tokens_seen": 248971264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 97104, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.738494873046875, + "objective/train/theoretical_loss": 4.267870237959327, + "objective/train/tokens_used": 269496800, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046694082246740223, + "loss": 1.7155, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669307923771314, + "loss": 1.7241, + "theoretical_loss": 4.267735381356835, + "tokens_seen": 249102336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004669207622868606, + "loss": 1.7735, + "theoretical_loss": 4.26760057016013, + "tokens_seen": 249167872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046691073219658983, + "loss": 1.8859, + "theoretical_loss": 4.2674658043419855, + "tokens_seen": 249233408 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046690070210631896, + "loss": 1.6898, + "theoretical_loss": 4.267331083875202, + "tokens_seen": 249298944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668906720160482, + "loss": 1.4933, + "theoretical_loss": 4.267196408732601, + "tokens_seen": 249364480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668806419257773, + "loss": 1.6614, + "theoretical_loss": 4.267061778887029, + "tokens_seen": 249430016 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046687061183550656, + "loss": 1.7496, + "theoretical_loss": 4.2669271943113545, + "tokens_seen": 249495552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046686058174523574, + "loss": 1.6707, + "theoretical_loss": 4.26679265497847, + "tokens_seen": 249561088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668505516549649, + "loss": 1.7323, + "theoretical_loss": 4.266658160861293, + "tokens_seen": 249626624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668405215646941, + "loss": 1.86, + "theoretical_loss": 4.266523711932761, + "tokens_seen": 249692160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668304914744233, + "loss": 1.8458, + "theoretical_loss": 4.266389308165837, + "tokens_seen": 249757696 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046682046138415246, + "loss": 1.7389, + "theoretical_loss": 4.266254949533507, + "tokens_seen": 249823232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668104312938817, + "loss": 1.7727, + "theoretical_loss": 4.26612063600878, + "tokens_seen": 249888768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004668004012036108, + "loss": 1.892, + "theoretical_loss": 4.265986367564688, + "tokens_seen": 249954304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046679037111334006, + "loss": 1.787, + "theoretical_loss": 4.265852144174286, + "tokens_seen": 250019840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667803410230692, + "loss": 1.6235, + "theoretical_loss": 4.265717965810653, + "tokens_seen": 250085376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667703109327984, + "loss": 1.6716, + "theoretical_loss": 4.265583832446892, + "tokens_seen": 250150912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667602808425276, + "loss": 1.7713, + "theoretical_loss": 4.265449744056126, + "tokens_seen": 250216448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667502507522568, + "loss": 1.7758, + "theoretical_loss": 4.265315700611504, + "tokens_seen": 250281984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046674022066198596, + "loss": 1.869, + "theoretical_loss": 4.2651817020861955, + "tokens_seen": 250347520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667301905717152, + "loss": 1.785, + "theoretical_loss": 4.265047748453396, + "tokens_seen": 250413056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667201604814443, + "loss": 1.8014, + "theoretical_loss": 4.264913839686323, + "tokens_seen": 250478592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046671013039117356, + "loss": 1.6997, + "theoretical_loss": 4.264779975758215, + "tokens_seen": 250544128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004667001003009027, + "loss": 1.6393, + "theoretical_loss": 4.2646461566423355, + "tokens_seen": 250609664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666900702106319, + "loss": 1.7098, + "theoretical_loss": 4.26451238231197, + "tokens_seen": 250675200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666800401203611, + "loss": 1.8663, + "theoretical_loss": 4.2643786527404295, + "tokens_seen": 250740736 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666700100300903, + "loss": 1.7673, + "theoretical_loss": 4.264244967901043, + "tokens_seen": 250806272 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046665997993981947, + "loss": 1.6894, + "theoretical_loss": 4.264111327767164, + "tokens_seen": 250871808 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046664994984954865, + "loss": 1.9099, + "theoretical_loss": 4.263977732312174, + "tokens_seen": 250937344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046663991975927783, + "loss": 1.6887, + "theoretical_loss": 4.26384418150947, + "tokens_seen": 251002880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046662988966900706, + "loss": 1.8296, + "theoretical_loss": 4.263710675332476, + "tokens_seen": 251068416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004666198595787362, + "loss": 1.8505, + "theoretical_loss": 4.263577213754638, + "tokens_seen": 251133952 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046660982948846543, + "loss": 1.8563, + "theoretical_loss": 4.2634437967494225, + "tokens_seen": 251199488 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046659979939819455, + "loss": 1.7764, + "theoretical_loss": 4.2633104242903235, + "tokens_seen": 251265024 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665897693079238, + "loss": 1.8095, + "theoretical_loss": 4.263177096350852, + "tokens_seen": 251330560 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046657973921765297, + "loss": 1.7206, + "theoretical_loss": 4.263043812904547, + "tokens_seen": 251396096 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046656970912738215, + "loss": 1.9784, + "theoretical_loss": 4.262910573924965, + "tokens_seen": 251461632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046655967903711133, + "loss": 1.622, + "theoretical_loss": 4.26277737938569, + "tokens_seen": 251527168 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046654964894684057, + "loss": 1.802, + "theoretical_loss": 4.262644229260324, + "tokens_seen": 251592704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665396188565697, + "loss": 1.8709, + "theoretical_loss": 4.262511123522495, + "tokens_seen": 251658240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046652958876629893, + "loss": 1.6729, + "theoretical_loss": 4.262378062145853, + "tokens_seen": 251723776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046651955867602806, + "loss": 1.7891, + "theoretical_loss": 4.262245045104068, + "tokens_seen": 251789312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004665095285857573, + "loss": 1.6938, + "theoretical_loss": 4.262112072370837, + "tokens_seen": 251854848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664994984954865, + "loss": 1.8459, + "theoretical_loss": 4.261979143919873, + "tokens_seen": 251920384 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046648946840521565, + "loss": 1.7857, + "theoretical_loss": 4.261846259724917, + "tokens_seen": 251985920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046647943831494484, + "loss": 1.7943, + "theoretical_loss": 4.261713419759731, + "tokens_seen": 252051456 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466469408224674, + "loss": 1.7697, + "theoretical_loss": 4.2615806239981, + "tokens_seen": 252116992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664593781344032, + "loss": 1.8036, + "theoretical_loss": 4.261447872413827, + "tokens_seen": 252182528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046644934804413243, + "loss": 1.7101, + "theoretical_loss": 4.261315164980743, + "tokens_seen": 252248064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 98257, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7829054594039917, + "objective/train/theoretical_loss": 4.261182501672698, + "objective/train/tokens_used": 272773600, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046643931795386156, + "loss": 1.7102, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664292878635908, + "loss": 1.8172, + "theoretical_loss": 4.261049882463565, + "tokens_seen": 252379136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004664192577733199, + "loss": 1.8366, + "theoretical_loss": 4.26091730732724, + "tokens_seen": 252444672 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046640922768304916, + "loss": 1.6736, + "theoretical_loss": 4.2607847762376405, + "tokens_seen": 252510208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046639919759277834, + "loss": 1.7671, + "theoretical_loss": 4.260652289168706, + "tokens_seen": 252575744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663891675025075, + "loss": 1.8, + "theoretical_loss": 4.260519846094399, + "tokens_seen": 252641280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663791374122367, + "loss": 1.8162, + "theoretical_loss": 4.2603874469887035, + "tokens_seen": 252706816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046636910732196594, + "loss": 1.8163, + "theoretical_loss": 4.260255091825625, + "tokens_seen": 252772352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046635907723169506, + "loss": 1.8245, + "theoretical_loss": 4.260122780579193, + "tokens_seen": 252837888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663490471414243, + "loss": 1.7111, + "theoretical_loss": 4.259990513223458, + "tokens_seen": 252903424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004663390170511534, + "loss": 1.8529, + "theoretical_loss": 4.259858289732491, + "tokens_seen": 252968960 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046632898696088266, + "loss": 1.7128, + "theoretical_loss": 4.259726110080388, + "tokens_seen": 253034496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046631895687061184, + "loss": 1.6867, + "theoretical_loss": 4.259593974241264, + "tokens_seen": 253100032 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466308926780341, + "loss": 1.7382, + "theoretical_loss": 4.25946188218926, + "tokens_seen": 253165568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662988966900702, + "loss": 1.6858, + "theoretical_loss": 4.259329833898534, + "tokens_seen": 253231104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662888665997994, + "loss": 1.686, + "theoretical_loss": 4.25919782934327, + "tokens_seen": 253296640 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046627883650952857, + "loss": 1.8233, + "theoretical_loss": 4.259065868497672, + "tokens_seen": 253362176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662688064192578, + "loss": 1.7494, + "theoretical_loss": 4.258933951335965, + "tokens_seen": 253427712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046625877632898693, + "loss": 1.8873, + "theoretical_loss": 4.258802077832399, + "tokens_seen": 253493248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046624874623871616, + "loss": 1.743, + "theoretical_loss": 4.258670247961243, + "tokens_seen": 253558784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046623871614844535, + "loss": 1.7487, + "theoretical_loss": 4.258538461696787, + "tokens_seen": 253624320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662286860581745, + "loss": 1.7036, + "theoretical_loss": 4.258406719013347, + "tokens_seen": 253689856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662186559679037, + "loss": 1.8295, + "theoretical_loss": 4.258275019885257, + "tokens_seen": 253755392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004662086258776329, + "loss": 1.8013, + "theoretical_loss": 4.258143364286875, + "tokens_seen": 253820928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661985957873621, + "loss": 1.666, + "theoretical_loss": 4.258011752192579, + "tokens_seen": 253886464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661885656970913, + "loss": 1.7318, + "theoretical_loss": 4.257880183576768, + "tokens_seen": 253952000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661785356068205, + "loss": 1.7341, + "theoretical_loss": 4.257748658413867, + "tokens_seen": 254017536 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046616850551654967, + "loss": 1.7085, + "theoretical_loss": 4.257617176678317, + "tokens_seen": 254083072 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046615847542627885, + "loss": 1.7104, + "theoretical_loss": 4.2574857383445845, + "tokens_seen": 254148608 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046614844533600803, + "loss": 1.7958, + "theoretical_loss": 4.257354343387156, + "tokens_seen": 254214144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046613841524573727, + "loss": 1.7613, + "theoretical_loss": 4.257222991780541, + "tokens_seen": 254279680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004661283851554664, + "loss": 1.8681, + "theoretical_loss": 4.257091683499268, + "tokens_seen": 254345216 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046611835506519563, + "loss": 1.9302, + "theoretical_loss": 4.25696041851789, + "tokens_seen": 254410752 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046610832497492475, + "loss": 1.812, + "theoretical_loss": 4.256829196810978, + "tokens_seen": 254476288 + }, + { + "epoch": 0.08, + "learning_rate": 0.000466098294884654, + "loss": 1.6694, + "theoretical_loss": 4.256698018353129, + "tokens_seen": 254541824 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046608826479438317, + "loss": 1.8038, + "theoretical_loss": 4.256566883118958, + "tokens_seen": 254607360 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046607823470411235, + "loss": 1.742, + "theoretical_loss": 4.256435791083101, + "tokens_seen": 254672896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046606820461384153, + "loss": 1.8249, + "theoretical_loss": 4.256304742220219, + "tokens_seen": 254738432 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046605817452357077, + "loss": 1.8049, + "theoretical_loss": 4.25617373650499, + "tokens_seen": 254803968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660481444332999, + "loss": 1.6153, + "theoretical_loss": 4.256042773912117, + "tokens_seen": 254869504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046603811434302913, + "loss": 1.754, + "theoretical_loss": 4.255911854416323, + "tokens_seen": 254935040 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046602808425275826, + "loss": 1.8075, + "theoretical_loss": 4.255780977992352, + "tokens_seen": 255000576 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660180541624875, + "loss": 1.6279, + "theoretical_loss": 4.255650144614968, + "tokens_seen": 255066112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004660080240722167, + "loss": 1.7851, + "theoretical_loss": 4.255519354258959, + "tokens_seen": 255131648 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046599799398194586, + "loss": 1.7626, + "theoretical_loss": 4.255388606899133, + "tokens_seen": 255197184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046598796389167504, + "loss": 1.725, + "theoretical_loss": 4.255257902510319, + "tokens_seen": 255262720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659779338014042, + "loss": 1.7794, + "theoretical_loss": 4.255127241067368, + "tokens_seen": 255328256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659679037111334, + "loss": 1.8078, + "theoretical_loss": 4.25499662254515, + "tokens_seen": 255393792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046595787362086263, + "loss": 1.7124, + "theoretical_loss": 4.254866046918559, + "tokens_seen": 255459328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046594784353059176, + "loss": 1.7934, + "theoretical_loss": 4.254735514162508, + "tokens_seen": 255524864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 99403, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7988343238830566, + "objective/train/theoretical_loss": 4.254605024251933, + "objective/train/tokens_used": 276050400, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465937813440321, + "loss": 1.7328, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004659277833500501, + "loss": 1.7352, + "theoretical_loss": 4.254474577161789, + "tokens_seen": 255655936 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046591775325977936, + "loss": 1.5886, + "theoretical_loss": 4.254344172867053, + "tokens_seen": 255721472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046590772316950854, + "loss": 1.76, + "theoretical_loss": 4.254213811342723, + "tokens_seen": 255787008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658976930792377, + "loss": 1.7598, + "theoretical_loss": 4.254083492563819, + "tokens_seen": 255852544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658876629889669, + "loss": 1.6742, + "theoretical_loss": 4.253953216505381, + "tokens_seen": 255918080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046587763289869614, + "loss": 1.8131, + "theoretical_loss": 4.253822983142468, + "tokens_seen": 255983616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046586760280842526, + "loss": 1.6697, + "theoretical_loss": 4.253692792450165, + "tokens_seen": 256049152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658575727181545, + "loss": 1.755, + "theoretical_loss": 4.253562644403573, + "tokens_seen": 256114688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658475426278836, + "loss": 1.746, + "theoretical_loss": 4.253432538977816, + "tokens_seen": 256180224 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046583751253761286, + "loss": 1.6707, + "theoretical_loss": 4.253302476148039, + "tokens_seen": 256245760 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046582748244734204, + "loss": 1.7828, + "theoretical_loss": 4.253172455889407, + "tokens_seen": 256311296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658174523570712, + "loss": 1.7254, + "theoretical_loss": 4.2530424781771075, + "tokens_seen": 256376832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004658074222668004, + "loss": 1.8882, + "theoretical_loss": 4.252912542986345, + "tokens_seen": 256442368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657973921765296, + "loss": 1.6075, + "theoretical_loss": 4.2527826502923505, + "tokens_seen": 256507904 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046578736208625877, + "loss": 1.7679, + "theoretical_loss": 4.25265280007037, + "tokens_seen": 256573440 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465777331995988, + "loss": 1.6765, + "theoretical_loss": 4.252522992295676, + "tokens_seen": 256638976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046576730190571713, + "loss": 1.6702, + "theoretical_loss": 4.252393226943555, + "tokens_seen": 256704512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046575727181544636, + "loss": 1.7171, + "theoretical_loss": 4.252263503989321, + "tokens_seen": 256770048 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046574724172517555, + "loss": 1.7216, + "theoretical_loss": 4.252133823408303, + "tokens_seen": 256835584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657372116349047, + "loss": 1.7939, + "theoretical_loss": 4.2520041851758545, + "tokens_seen": 256901120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657271815446339, + "loss": 1.5922, + "theoretical_loss": 4.251874589267347, + "tokens_seen": 256966656 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004657171514543631, + "loss": 1.7302, + "theoretical_loss": 4.251745035658176, + "tokens_seen": 257032192 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046570712136409227, + "loss": 1.8358, + "theoretical_loss": 4.251615524323752, + "tokens_seen": 257097728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656970912738215, + "loss": 1.8373, + "theoretical_loss": 4.251486055239512, + "tokens_seen": 257163264 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046568706118355063, + "loss": 1.8293, + "theoretical_loss": 4.251356628380911, + "tokens_seen": 257228800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046567703109327987, + "loss": 1.8597, + "theoretical_loss": 4.251227243723424, + "tokens_seen": 257294336 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465667001003009, + "loss": 1.7287, + "theoretical_loss": 4.251097901242546, + "tokens_seen": 257359872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046565697091273823, + "loss": 1.878, + "theoretical_loss": 4.250968600913795, + "tokens_seen": 257425408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656469408224674, + "loss": 1.7376, + "theoretical_loss": 4.250839342712707, + "tokens_seen": 257490944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656369107321966, + "loss": 1.6428, + "theoretical_loss": 4.250710126614838, + "tokens_seen": 257556480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004656268806419258, + "loss": 1.8846, + "theoretical_loss": 4.250580952595769, + "tokens_seen": 257622016 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046561685055165495, + "loss": 1.7793, + "theoretical_loss": 4.250451820631095, + "tokens_seen": 257687552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046560682046138414, + "loss": 1.6975, + "theoretical_loss": 4.2503227306964355, + "tokens_seen": 257753088 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046559679037111337, + "loss": 1.7541, + "theoretical_loss": 4.25019368276743, + "tokens_seen": 257818624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655867602808425, + "loss": 1.7202, + "theoretical_loss": 4.250064676819735, + "tokens_seen": 257884160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046557673019057173, + "loss": 1.6112, + "theoretical_loss": 4.249935712829033, + "tokens_seen": 257949696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655667001003009, + "loss": 1.9671, + "theoretical_loss": 4.249806790771022, + "tokens_seen": 258015232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655566700100301, + "loss": 1.7527, + "theoretical_loss": 4.249677910621422, + "tokens_seen": 258080768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655466399197593, + "loss": 1.8401, + "theoretical_loss": 4.249549072355974, + "tokens_seen": 258146304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046553660982948846, + "loss": 1.7622, + "theoretical_loss": 4.2494202759504365, + "tokens_seen": 258211840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046552657973921764, + "loss": 1.7514, + "theoretical_loss": 4.249291521380591, + "tokens_seen": 258277376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004655165496489469, + "loss": 1.7739, + "theoretical_loss": 4.249162808622237, + "tokens_seen": 258342912 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465506519558676, + "loss": 1.8573, + "theoretical_loss": 4.249034137651197, + "tokens_seen": 258408448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046549648946840524, + "loss": 1.698, + "theoretical_loss": 4.248905508443311, + "tokens_seen": 258473984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046548645937813436, + "loss": 1.731, + "theoretical_loss": 4.248776920974439, + "tokens_seen": 258539520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654764292878636, + "loss": 1.6788, + "theoretical_loss": 4.248648375220464, + "tokens_seen": 258605056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654663991975928, + "loss": 1.6928, + "theoretical_loss": 4.248519871157285, + "tokens_seen": 258670592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046545636910732196, + "loss": 1.8293, + "theoretical_loss": 4.248391408760824, + "tokens_seen": 258736128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654463390170512, + "loss": 1.6628, + "theoretical_loss": 4.248262988007022, + "tokens_seen": 258801664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 100739, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6191655397415161, + "objective/train/theoretical_loss": 4.248134608871841, + "objective/train/tokens_used": 279327200, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654363089267803, + "loss": 1.7002, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046542627883650956, + "loss": 1.8079, + "theoretical_loss": 4.24800627133126, + "tokens_seen": 258932736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046541624874623874, + "loss": 1.8607, + "theoretical_loss": 4.247877975361281, + "tokens_seen": 258998272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004654062186559679, + "loss": 1.6984, + "theoretical_loss": 4.247749720937925, + "tokens_seen": 259063808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653961885656971, + "loss": 1.7662, + "theoretical_loss": 4.247621508037232, + "tokens_seen": 259129344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046538615847542634, + "loss": 1.7372, + "theoretical_loss": 4.247493336635264, + "tokens_seen": 259194880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046537612838515546, + "loss": 1.7048, + "theoretical_loss": 4.2473652067081, + "tokens_seen": 259260416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653660982948847, + "loss": 1.6958, + "theoretical_loss": 4.247237118231841, + "tokens_seen": 259325952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653560682046138, + "loss": 1.7623, + "theoretical_loss": 4.247109071182607, + "tokens_seen": 259391488 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046534603811434306, + "loss": 1.7838, + "theoretical_loss": 4.246981065536538, + "tokens_seen": 259457024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046533600802407224, + "loss": 1.7954, + "theoretical_loss": 4.246853101269793, + "tokens_seen": 259522560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653259779338014, + "loss": 1.9012, + "theoretical_loss": 4.2467251783585525, + "tokens_seen": 259588096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653159478435306, + "loss": 1.757, + "theoretical_loss": 4.246597296779014, + "tokens_seen": 259653632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004653059177532598, + "loss": 1.7933, + "theoretical_loss": 4.246469456507398, + "tokens_seen": 259719168 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046529588766298897, + "loss": 1.7609, + "theoretical_loss": 4.246341657519943, + "tokens_seen": 259784704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652858575727182, + "loss": 1.802, + "theoretical_loss": 4.246213899792905, + "tokens_seen": 259850240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046527582748244733, + "loss": 1.7663, + "theoretical_loss": 4.246086183302564, + "tokens_seen": 259915776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046526579739217656, + "loss": 1.5906, + "theoretical_loss": 4.2459585080252165, + "tokens_seen": 259981312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046525576730190575, + "loss": 1.7124, + "theoretical_loss": 4.245830873937179, + "tokens_seen": 260046848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652457372116349, + "loss": 1.8328, + "theoretical_loss": 4.24570328101479, + "tokens_seen": 260112384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652357071213641, + "loss": 1.7474, + "theoretical_loss": 4.245575729234403, + "tokens_seen": 260177920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652256770310933, + "loss": 1.8164, + "theoretical_loss": 4.245448218572396, + "tokens_seen": 260243456 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046521564694082247, + "loss": 1.8183, + "theoretical_loss": 4.245320749005163, + "tokens_seen": 260308992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004652056168505517, + "loss": 1.6819, + "theoretical_loss": 4.245193320509118, + "tokens_seen": 260374528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046519558676028083, + "loss": 1.7871, + "theoretical_loss": 4.245065933060697, + "tokens_seen": 260440064 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046518555667001007, + "loss": 1.7288, + "theoretical_loss": 4.244938586636351, + "tokens_seen": 260505600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651755265797392, + "loss": 1.7974, + "theoretical_loss": 4.244811281212556, + "tokens_seen": 260571136 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046516549648946843, + "loss": 1.6762, + "theoretical_loss": 4.244684016765802, + "tokens_seen": 260636672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651554663991976, + "loss": 1.8278, + "theoretical_loss": 4.2445567932726025, + "tokens_seen": 260702208 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004651454363089268, + "loss": 1.851, + "theoretical_loss": 4.244429610709488, + "tokens_seen": 260767744 + }, + { + "epoch": 0.08, + "learning_rate": 0.000465135406218656, + "loss": 1.7451, + "theoretical_loss": 4.244302469053007, + "tokens_seen": 260833280 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046512537612838515, + "loss": 1.7118, + "theoretical_loss": 4.244175368279733, + "tokens_seen": 260898816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046511534603811434, + "loss": 1.6197, + "theoretical_loss": 4.244048308366253, + "tokens_seen": 260964352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046510531594784357, + "loss": 1.6275, + "theoretical_loss": 4.243921289289175, + "tokens_seen": 261029888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650952858575727, + "loss": 1.694, + "theoretical_loss": 4.243794311025128, + "tokens_seen": 261095424 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046508525576730193, + "loss": 1.7494, + "theoretical_loss": 4.243667373550757, + "tokens_seen": 261160960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650752256770311, + "loss": 1.7788, + "theoretical_loss": 4.24354047684273, + "tokens_seen": 261226496 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650651955867603, + "loss": 1.8104, + "theoretical_loss": 4.243413620877732, + "tokens_seen": 261292032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650551654964895, + "loss": 1.6446, + "theoretical_loss": 4.243286805632467, + "tokens_seen": 261357568 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046504513540621866, + "loss": 1.7558, + "theoretical_loss": 4.243160031083658, + "tokens_seen": 261423104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046503510531594784, + "loss": 1.8443, + "theoretical_loss": 4.24303329720805, + "tokens_seen": 261488640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650250752256771, + "loss": 1.8766, + "theoretical_loss": 4.242906603982402, + "tokens_seen": 261554176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004650150451354062, + "loss": 1.8172, + "theoretical_loss": 4.242779951383497, + "tokens_seen": 261619712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046500501504513544, + "loss": 1.7704, + "theoretical_loss": 4.242653339388135, + "tokens_seen": 261685248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046499498495486456, + "loss": 1.7434, + "theoretical_loss": 4.242526767973134, + "tokens_seen": 261750784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649849548645938, + "loss": 1.8131, + "theoretical_loss": 4.242400237115333, + "tokens_seen": 261816320 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464974924774323, + "loss": 1.734, + "theoretical_loss": 4.242273746791589, + "tokens_seen": 261881856 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046496489468405216, + "loss": 1.6897, + "theoretical_loss": 4.242147296978778, + "tokens_seen": 261947392 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046495486459378134, + "loss": 1.6299, + "theoretical_loss": 4.242020887653795, + "tokens_seen": 262012928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649448345035105, + "loss": 1.6238, + "theoretical_loss": 4.241894518793555, + "tokens_seen": 262078464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 101934, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8843626976013184, + "objective/train/theoretical_loss": 4.24176819037499, + "objective/train/tokens_used": 282604000, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649348044132397, + "loss": 1.8011, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046492477432296894, + "loss": 1.7469, + "theoretical_loss": 4.2416419023750525, + "tokens_seen": 262209536 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046491474423269807, + "loss": 1.6595, + "theoretical_loss": 4.241515654770713, + "tokens_seen": 262275072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004649047141424273, + "loss": 1.73, + "theoretical_loss": 4.241389447538961, + "tokens_seen": 262340608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648946840521565, + "loss": 1.7598, + "theoretical_loss": 4.241263280656806, + "tokens_seen": 262406144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046488465396188566, + "loss": 1.7611, + "theoretical_loss": 4.241137154101275, + "tokens_seen": 262471680 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046487462387161485, + "loss": 1.845, + "theoretical_loss": 4.2410110678494135, + "tokens_seen": 262537216 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464864593781344, + "loss": 1.6581, + "theoretical_loss": 4.240885021878286, + "tokens_seen": 262602752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648545636910732, + "loss": 1.8448, + "theoretical_loss": 4.240759016164978, + "tokens_seen": 262668288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046484453360080244, + "loss": 1.75, + "theoretical_loss": 4.240633050686592, + "tokens_seen": 262733824 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046483450351053157, + "loss": 1.7078, + "theoretical_loss": 4.2405071254202475, + "tokens_seen": 262799360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004648244734202608, + "loss": 1.7263, + "theoretical_loss": 4.240381240343085, + "tokens_seen": 262864896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046481444332998993, + "loss": 1.6317, + "theoretical_loss": 4.240255395432264, + "tokens_seen": 262930432 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046480441323971917, + "loss": 1.6311, + "theoretical_loss": 4.240129590664961, + "tokens_seen": 262995968 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046479438314944835, + "loss": 1.7277, + "theoretical_loss": 4.240003826018373, + "tokens_seen": 263061504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046478435305917753, + "loss": 1.687, + "theoretical_loss": 4.239878101469713, + "tokens_seen": 263127040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647743229689067, + "loss": 1.7349, + "theoretical_loss": 4.239752416996215, + "tokens_seen": 263192576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046476429287863595, + "loss": 1.6493, + "theoretical_loss": 4.239626772575131, + "tokens_seen": 263258112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647542627883651, + "loss": 1.8507, + "theoretical_loss": 4.239501168183732, + "tokens_seen": 263323648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647442326980943, + "loss": 1.6862, + "theoretical_loss": 4.239375603799305, + "tokens_seen": 263389184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046473420260782344, + "loss": 1.7616, + "theoretical_loss": 4.2392500793991585, + "tokens_seen": 263454720 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046472417251755267, + "loss": 1.8334, + "theoretical_loss": 4.239124594960619, + "tokens_seen": 263520256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004647141424272819, + "loss": 1.7984, + "theoretical_loss": 4.23899915046103, + "tokens_seen": 263585792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046470411233701103, + "loss": 1.7591, + "theoretical_loss": 4.238873745877756, + "tokens_seen": 263651328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046469408224674027, + "loss": 1.749, + "theoretical_loss": 4.238748381188176, + "tokens_seen": 263716864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646840521564694, + "loss": 1.7648, + "theoretical_loss": 4.238623056369692, + "tokens_seen": 263782400 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046467402206619863, + "loss": 1.7872, + "theoretical_loss": 4.238497771399721, + "tokens_seen": 263847936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646639919759278, + "loss": 1.7237, + "theoretical_loss": 4.2383725262556995, + "tokens_seen": 263913472 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464653961885657, + "loss": 1.7632, + "theoretical_loss": 4.238247320915083, + "tokens_seen": 263979008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646439317953862, + "loss": 1.7173, + "theoretical_loss": 4.238122155355345, + "tokens_seen": 264044544 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046463390170511535, + "loss": 1.829, + "theoretical_loss": 4.237997029553977, + "tokens_seen": 264110080 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046462387161484454, + "loss": 1.8432, + "theoretical_loss": 4.237871943488488, + "tokens_seen": 264175616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046461384152457377, + "loss": 1.7322, + "theoretical_loss": 4.237746897136408, + "tokens_seen": 264241152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004646038114343029, + "loss": 1.7707, + "theoretical_loss": 4.237621890475284, + "tokens_seen": 264306688 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046459378134403213, + "loss": 1.7163, + "theoretical_loss": 4.237496923482678, + "tokens_seen": 264372224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645837512537613, + "loss": 1.678, + "theoretical_loss": 4.237371996136175, + "tokens_seen": 264437760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645737211634905, + "loss": 1.7331, + "theoretical_loss": 4.237247108413376, + "tokens_seen": 264503296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645636910732197, + "loss": 1.5832, + "theoretical_loss": 4.237122260291901, + "tokens_seen": 264568832 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046455366098294886, + "loss": 1.742, + "theoretical_loss": 4.236997451749387, + "tokens_seen": 264634368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046454363089267804, + "loss": 1.6319, + "theoretical_loss": 4.236872682763489, + "tokens_seen": 264699904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645336008024073, + "loss": 1.7331, + "theoretical_loss": 4.236747953311884, + "tokens_seen": 264765440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004645235707121364, + "loss": 1.7917, + "theoretical_loss": 4.236623263372261, + "tokens_seen": 264830976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046451354062186564, + "loss": 1.8082, + "theoretical_loss": 4.236498612922331, + "tokens_seen": 264896512 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046450351053159476, + "loss": 1.7643, + "theoretical_loss": 4.236374001939822, + "tokens_seen": 264962048 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464493480441324, + "loss": 1.8368, + "theoretical_loss": 4.2362494304024825, + "tokens_seen": 265027584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644834503510532, + "loss": 1.8147, + "theoretical_loss": 4.236124898288074, + "tokens_seen": 265093120 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046447342026078236, + "loss": 1.726, + "theoretical_loss": 4.236000405574382, + "tokens_seen": 265158656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046446339017051154, + "loss": 1.7549, + "theoretical_loss": 4.235875952239204, + "tokens_seen": 265224192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644533600802407, + "loss": 1.6786, + "theoretical_loss": 4.235751538260359, + "tokens_seen": 265289728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644433299899699, + "loss": 1.808, + "theoretical_loss": 4.2356271636156855, + "tokens_seen": 265355264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 103026, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5964581966400146, + "objective/train/theoretical_loss": 4.235502828283035, + "objective/train/tokens_used": 285880800, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046443329989969914, + "loss": 1.7543, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046442326980942827, + "loss": 1.8498, + "theoretical_loss": 4.235378532240283, + "tokens_seen": 265486336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644132397191575, + "loss": 1.8498, + "theoretical_loss": 4.235254275465317, + "tokens_seen": 265551872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004644032096288867, + "loss": 1.641, + "theoretical_loss": 4.235130057936045, + "tokens_seen": 265617408 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046439317953861586, + "loss": 1.6584, + "theoretical_loss": 4.235005879630395, + "tokens_seen": 265682944 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046438314944834505, + "loss": 1.7548, + "theoretical_loss": 4.23488174052631, + "tokens_seen": 265748480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004643731193580742, + "loss": 1.7025, + "theoretical_loss": 4.234757640601751, + "tokens_seen": 265814016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004643630892678034, + "loss": 1.731, + "theoretical_loss": 4.234633579834698, + "tokens_seen": 265879552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046435305917753264, + "loss": 1.6832, + "theoretical_loss": 4.234509558203149, + "tokens_seen": 265945088 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046434302908726177, + "loss": 1.7919, + "theoretical_loss": 4.234385575685116, + "tokens_seen": 266010624 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464332998996991, + "loss": 1.7264, + "theoretical_loss": 4.234261632258636, + "tokens_seen": 266076160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046432296890672013, + "loss": 1.7395, + "theoretical_loss": 4.234137727901757, + "tokens_seen": 266141696 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046431293881644937, + "loss": 1.8492, + "theoretical_loss": 4.234013862592548, + "tokens_seen": 266207232 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046430290872617855, + "loss": 1.7352, + "theoretical_loss": 4.233890036309095, + "tokens_seen": 266272768 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046429287863590773, + "loss": 1.6185, + "theoretical_loss": 4.233766249029501, + "tokens_seen": 266338304 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642828485456369, + "loss": 1.7621, + "theoretical_loss": 4.233642500731889, + "tokens_seen": 266403840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046427281845536615, + "loss": 1.8522, + "theoretical_loss": 4.233518791394396, + "tokens_seen": 266469376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642627883650953, + "loss": 1.8293, + "theoretical_loss": 4.233395120995181, + "tokens_seen": 266534912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642527582748245, + "loss": 1.773, + "theoretical_loss": 4.233271489512416, + "tokens_seen": 266600448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046424272818455364, + "loss": 1.7598, + "theoretical_loss": 4.233147896924294, + "tokens_seen": 266665984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046423269809428287, + "loss": 1.828, + "theoretical_loss": 4.233024343209023, + "tokens_seen": 266731520 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046422266800401205, + "loss": 1.756, + "theoretical_loss": 4.232900828344833, + "tokens_seen": 266797056 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046421263791374123, + "loss": 1.6804, + "theoretical_loss": 4.232777352309966, + "tokens_seen": 266862592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004642026078234704, + "loss": 1.7952, + "theoretical_loss": 4.232653915082684, + "tokens_seen": 266928128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641925777331996, + "loss": 1.741, + "theoretical_loss": 4.232530516641267, + "tokens_seen": 266993664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641825476429288, + "loss": 1.7267, + "theoretical_loss": 4.232407156964014, + "tokens_seen": 267059200 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464172517552658, + "loss": 1.709, + "theoretical_loss": 4.232283836029237, + "tokens_seen": 267124736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046416248746238714, + "loss": 1.6933, + "theoretical_loss": 4.232160553815269, + "tokens_seen": 267190272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641524573721164, + "loss": 1.7704, + "theoretical_loss": 4.232037310300458, + "tokens_seen": 267255808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641424272818455, + "loss": 1.7058, + "theoretical_loss": 4.231914105463173, + "tokens_seen": 267321344 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046413239719157474, + "loss": 1.8327, + "theoretical_loss": 4.231790939281797, + "tokens_seen": 267386880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641223671013039, + "loss": 1.7832, + "theoretical_loss": 4.231667811734732, + "tokens_seen": 267452416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641123370110331, + "loss": 1.8292, + "theoretical_loss": 4.231544722800395, + "tokens_seen": 267517952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004641023069207623, + "loss": 1.6926, + "theoretical_loss": 4.231421672457225, + "tokens_seen": 267583488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640922768304915, + "loss": 1.8496, + "theoretical_loss": 4.231298660683675, + "tokens_seen": 267649024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046408224674022064, + "loss": 1.6857, + "theoretical_loss": 4.231175687458214, + "tokens_seen": 267714560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640722166499499, + "loss": 1.7576, + "theoretical_loss": 4.231052752759332, + "tokens_seen": 267780096 + }, + { + "epoch": 0.08, + "learning_rate": 0.000464062186559679, + "loss": 1.6505, + "theoretical_loss": 4.230929856565535, + "tokens_seen": 267845632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046405215646940824, + "loss": 1.7048, + "theoretical_loss": 4.230806998855343, + "tokens_seen": 267911168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640421263791374, + "loss": 1.7751, + "theoretical_loss": 4.230684179607297, + "tokens_seen": 267976704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640320962888666, + "loss": 1.7347, + "theoretical_loss": 4.2305613987999555, + "tokens_seen": 268042240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004640220661985958, + "loss": 1.7086, + "theoretical_loss": 4.230438656411891, + "tokens_seen": 268107776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046401203610832496, + "loss": 1.7179, + "theoretical_loss": 4.230315952421697, + "tokens_seen": 268173312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046400200601805414, + "loss": 1.7292, + "theoretical_loss": 4.230193286807981, + "tokens_seen": 268238848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639919759277834, + "loss": 1.7664, + "theoretical_loss": 4.230070659549368, + "tokens_seen": 268304384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639819458375125, + "loss": 1.752, + "theoretical_loss": 4.229948070624502, + "tokens_seen": 268369920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046397191574724174, + "loss": 1.7228, + "theoretical_loss": 4.229825520012042, + "tokens_seen": 268435456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639618856569709, + "loss": 1.8388, + "theoretical_loss": 4.2297030076906665, + "tokens_seen": 268500992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639518555667001, + "loss": 1.7911, + "theoretical_loss": 4.229580533639068, + "tokens_seen": 268566528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046394182547642934, + "loss": 1.7111, + "theoretical_loss": 4.229458097835959, + "tokens_seen": 268632064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 104173, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6287827491760254, + "objective/train/theoretical_loss": 4.2293357002600676, + "objective/train/tokens_used": 289157600, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046393179538615847, + "loss": 1.6299, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639217652958877, + "loss": 1.7138, + "theoretical_loss": 4.229213340890139, + "tokens_seen": 268763136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004639117352056169, + "loss": 1.7581, + "theoretical_loss": 4.229091019704935, + "tokens_seen": 268828672 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046390170511534606, + "loss": 1.8302, + "theoretical_loss": 4.2289687366832345, + "tokens_seen": 268894208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046389167502507525, + "loss": 1.7544, + "theoretical_loss": 4.228846491803835, + "tokens_seen": 268959744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638816449348044, + "loss": 1.7829, + "theoretical_loss": 4.228724285045549, + "tokens_seen": 269025280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638716148445336, + "loss": 1.6978, + "theoretical_loss": 4.228602116387206, + "tokens_seen": 269090816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046386158475426284, + "loss": 1.7405, + "theoretical_loss": 4.228479985807653, + "tokens_seen": 269156352 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046385155466399197, + "loss": 1.5477, + "theoretical_loss": 4.228357893285755, + "tokens_seen": 269221888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004638415245737212, + "loss": 1.7279, + "theoretical_loss": 4.2282358388003916, + "tokens_seen": 269287424 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046383149448345033, + "loss": 1.6576, + "theoretical_loss": 4.228113822330461, + "tokens_seen": 269352960 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046382146439317957, + "loss": 1.6152, + "theoretical_loss": 4.227991843854878, + "tokens_seen": 269418496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046381143430290875, + "loss": 1.6935, + "theoretical_loss": 4.227869903352572, + "tokens_seen": 269484032 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046380140421263793, + "loss": 1.6617, + "theoretical_loss": 4.227748000802493, + "tokens_seen": 269549568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637913741223671, + "loss": 1.7716, + "theoretical_loss": 4.227626136183606, + "tokens_seen": 269615104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046378134403209635, + "loss": 1.7391, + "theoretical_loss": 4.227504309474892, + "tokens_seen": 269680640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637713139418255, + "loss": 1.738, + "theoretical_loss": 4.227382520655348, + "tokens_seen": 269746176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637612838515547, + "loss": 1.7938, + "theoretical_loss": 4.227260769703991, + "tokens_seen": 269811712 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046375125376128384, + "loss": 1.6186, + "theoretical_loss": 4.227139056599851, + "tokens_seen": 269877248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046374122367101307, + "loss": 1.6778, + "theoretical_loss": 4.2270173813219785, + "tokens_seen": 269942784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046373119358074225, + "loss": 1.6778, + "theoretical_loss": 4.226895743849437, + "tokens_seen": 270008320 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046372116349047143, + "loss": 1.7754, + "theoretical_loss": 4.226774144161309, + "tokens_seen": 270073856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637111334002006, + "loss": 1.7534, + "theoretical_loss": 4.226652582236694, + "tokens_seen": 270139392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004637011033099298, + "loss": 1.7687, + "theoretical_loss": 4.226531058054706, + "tokens_seen": 270204928 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463691073219659, + "loss": 1.777, + "theoretical_loss": 4.226409571594477, + "tokens_seen": 270270464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636810431293882, + "loss": 1.7608, + "theoretical_loss": 4.226288122835156, + "tokens_seen": 270336000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046367101303911734, + "loss": 1.6661, + "theoretical_loss": 4.226166711755906, + "tokens_seen": 270401536 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636609829488466, + "loss": 1.6667, + "theoretical_loss": 4.2260453383359105, + "tokens_seen": 270467072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636509528585757, + "loss": 1.8417, + "theoretical_loss": 4.2259240025543665, + "tokens_seen": 270532608 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046364092276830494, + "loss": 1.8021, + "theoretical_loss": 4.225802704390488, + "tokens_seen": 270598144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636308926780341, + "loss": 1.6292, + "theoretical_loss": 4.2256814438235075, + "tokens_seen": 270663680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636208625877633, + "loss": 1.6469, + "theoretical_loss": 4.225560220832672, + "tokens_seen": 270729216 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636108324974925, + "loss": 1.6053, + "theoretical_loss": 4.225439035397246, + "tokens_seen": 270794752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004636008024072217, + "loss": 1.7165, + "theoretical_loss": 4.225317887496508, + "tokens_seen": 270860288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046359077231695084, + "loss": 1.6475, + "theoretical_loss": 4.2251967771097565, + "tokens_seen": 270925824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635807422266801, + "loss": 1.7979, + "theoretical_loss": 4.225075704216305, + "tokens_seen": 270991360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635707121364092, + "loss": 1.6364, + "theoretical_loss": 4.224954668795483, + "tokens_seen": 271056896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046356068204613844, + "loss": 1.7271, + "theoretical_loss": 4.224833670826636, + "tokens_seen": 271122432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635506519558676, + "loss": 1.7709, + "theoretical_loss": 4.224712710289127, + "tokens_seen": 271187968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635406218655968, + "loss": 1.809, + "theoretical_loss": 4.2245917871623355, + "tokens_seen": 271253504 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463530591775326, + "loss": 1.6643, + "theoretical_loss": 4.224470901425655, + "tokens_seen": 271319040 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046352056168505516, + "loss": 1.8533, + "theoretical_loss": 4.224350053058499, + "tokens_seen": 271384576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046351053159478434, + "loss": 1.7803, + "theoretical_loss": 4.224229242040295, + "tokens_seen": 271450112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004635005015045136, + "loss": 1.7077, + "theoretical_loss": 4.224108468350485, + "tokens_seen": 271515648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634904714142427, + "loss": 1.8428, + "theoretical_loss": 4.223987731968532, + "tokens_seen": 271581184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046348044132397194, + "loss": 1.8534, + "theoretical_loss": 4.223867032873912, + "tokens_seen": 271646720 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046347041123370107, + "loss": 1.7623, + "theoretical_loss": 4.223746371046116, + "tokens_seen": 271712256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634603811434303, + "loss": 1.6828, + "theoretical_loss": 4.223625746464656, + "tokens_seen": 271777792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634503510531595, + "loss": 1.8432, + "theoretical_loss": 4.223505159109054, + "tokens_seen": 271843328 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046344032096288867, + "loss": 1.6782, + "theoretical_loss": 4.223384608958855, + "tokens_seen": 271908864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 105443, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6679344177246094, + "objective/train/theoretical_loss": 4.223264095993615, + "objective/train/tokens_used": 292434400, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046343029087261785, + "loss": 1.7294, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634202607823471, + "loss": 1.611, + "theoretical_loss": 4.223143620192907, + "tokens_seen": 272039936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004634102306920762, + "loss": 1.7004, + "theoretical_loss": 4.223023181536323, + "tokens_seen": 272105472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046340020060180545, + "loss": 1.686, + "theoretical_loss": 4.2229027800034675, + "tokens_seen": 272171008 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046339017051153457, + "loss": 1.6688, + "theoretical_loss": 4.222782415573962, + "tokens_seen": 272236544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633801404212638, + "loss": 1.69, + "theoretical_loss": 4.222662088227447, + "tokens_seen": 272302080 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463370110330993, + "loss": 1.8354, + "theoretical_loss": 4.222541797943576, + "tokens_seen": 272367616 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046336008024072217, + "loss": 1.7367, + "theoretical_loss": 4.222421544702018, + "tokens_seen": 272433152 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046335005015045135, + "loss": 1.7158, + "theoretical_loss": 4.2223013284824615, + "tokens_seen": 272498688 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046334002006018053, + "loss": 1.6651, + "theoretical_loss": 4.222181149264608, + "tokens_seen": 272564224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633299899699097, + "loss": 1.7032, + "theoretical_loss": 4.222061007028175, + "tokens_seen": 272629760 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046331995987963895, + "loss": 1.7256, + "theoretical_loss": 4.221940901752899, + "tokens_seen": 272695296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004633099297893681, + "loss": 1.7898, + "theoretical_loss": 4.221820833418529, + "tokens_seen": 272760832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632998996990973, + "loss": 1.7349, + "theoretical_loss": 4.221700802004832, + "tokens_seen": 272826368 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046328986960882644, + "loss": 1.8399, + "theoretical_loss": 4.2215808074915895, + "tokens_seen": 272891904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632798395185557, + "loss": 1.8253, + "theoretical_loss": 4.221460849858601, + "tokens_seen": 272957440 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046326980942828485, + "loss": 1.7463, + "theoretical_loss": 4.221340929085679, + "tokens_seen": 273022976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046325977933801404, + "loss": 1.6424, + "theoretical_loss": 4.221221045152655, + "tokens_seen": 273088512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632497492477432, + "loss": 1.8493, + "theoretical_loss": 4.221101198039375, + "tokens_seen": 273154048 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046323971915747245, + "loss": 1.8312, + "theoretical_loss": 4.2209813877257, + "tokens_seen": 273219584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632296890672016, + "loss": 1.671, + "theoretical_loss": 4.220861614191508, + "tokens_seen": 273285120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004632196589769308, + "loss": 1.6828, + "theoretical_loss": 4.220741877416691, + "tokens_seen": 273350656 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046320962888666, + "loss": 1.6565, + "theoretical_loss": 4.22062217738116, + "tokens_seen": 273416192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631995987963892, + "loss": 1.7211, + "theoretical_loss": 4.22050251406484, + "tokens_seen": 273481728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631895687061184, + "loss": 1.8285, + "theoretical_loss": 4.220382887447671, + "tokens_seen": 273547264 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046317953861584754, + "loss": 1.8129, + "theoretical_loss": 4.22026329750961, + "tokens_seen": 273612800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631695085255768, + "loss": 1.6919, + "theoretical_loss": 4.220143744230629, + "tokens_seen": 273678336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631594784353059, + "loss": 1.7385, + "theoretical_loss": 4.220024227590717, + "tokens_seen": 273743872 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046314944834503514, + "loss": 1.7049, + "theoretical_loss": 4.219904747569877, + "tokens_seen": 273809408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631394182547643, + "loss": 1.7988, + "theoretical_loss": 4.219785304148129, + "tokens_seen": 273874944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631293881644935, + "loss": 1.8251, + "theoretical_loss": 4.219665897305506, + "tokens_seen": 273940480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631193580742227, + "loss": 1.7383, + "theoretical_loss": 4.219546527022062, + "tokens_seen": 274006016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004631093279839519, + "loss": 1.7249, + "theoretical_loss": 4.219427193277861, + "tokens_seen": 274071552 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046309929789368104, + "loss": 1.716, + "theoretical_loss": 4.219307896052987, + "tokens_seen": 274137088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630892678034103, + "loss": 1.8143, + "theoretical_loss": 4.219188635327535, + "tokens_seen": 274202624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630792377131394, + "loss": 1.733, + "theoretical_loss": 4.219069411081622, + "tokens_seen": 274268160 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046306920762286864, + "loss": 1.6222, + "theoretical_loss": 4.218950223295374, + "tokens_seen": 274333696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630591775325978, + "loss": 1.6462, + "theoretical_loss": 4.218831071948936, + "tokens_seen": 274399232 + }, + { + "epoch": 0.08, + "learning_rate": 0.000463049147442327, + "loss": 1.7766, + "theoretical_loss": 4.218711957022469, + "tokens_seen": 274464768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630391173520562, + "loss": 1.6205, + "theoretical_loss": 4.218592878496148, + "tokens_seen": 274530304 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046302908726178536, + "loss": 1.7276, + "theoretical_loss": 4.218473836350164, + "tokens_seen": 274595840 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046301905717151455, + "loss": 1.6554, + "theoretical_loss": 4.2183548305647225, + "tokens_seen": 274661376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004630090270812438, + "loss": 1.7496, + "theoretical_loss": 4.218235861120047, + "tokens_seen": 274726912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629989969909729, + "loss": 1.8331, + "theoretical_loss": 4.218116927996375, + "tokens_seen": 274792448 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046298896690070214, + "loss": 1.6639, + "theoretical_loss": 4.217998031173959, + "tokens_seen": 274857984 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046297893681043127, + "loss": 1.722, + "theoretical_loss": 4.217879170633067, + "tokens_seen": 274923520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629689067201605, + "loss": 1.763, + "theoretical_loss": 4.217760346353984, + "tokens_seen": 274989056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629588766298897, + "loss": 1.7952, + "theoretical_loss": 4.217641558317008, + "tokens_seen": 275054592 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046294884653961887, + "loss": 1.616, + "theoretical_loss": 4.217522806502455, + "tokens_seen": 275120128 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046293881644934805, + "loss": 1.6642, + "theoretical_loss": 4.217404090890653, + "tokens_seen": 275185664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 106703, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6659315824508667, + "objective/train/theoretical_loss": 4.217285411461949, + "objective/train/tokens_used": 295711200, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629287863590773, + "loss": 1.6553, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004629187562688064, + "loss": 1.6874, + "theoretical_loss": 4.217166768196704, + "tokens_seen": 275316736 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046290872617853565, + "loss": 1.7951, + "theoretical_loss": 4.217048161075294, + "tokens_seen": 275382272 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046289869608826477, + "loss": 1.6611, + "theoretical_loss": 4.2169295900781085, + "tokens_seen": 275447808 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462888665997994, + "loss": 1.8, + "theoretical_loss": 4.216811055185556, + "tokens_seen": 275513344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628786359077232, + "loss": 1.7543, + "theoretical_loss": 4.216692556378058, + "tokens_seen": 275578880 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046286860581745237, + "loss": 1.709, + "theoretical_loss": 4.216574093636053, + "tokens_seen": 275644416 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046285857572718155, + "loss": 1.7493, + "theoretical_loss": 4.216455666939991, + "tokens_seen": 275709952 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046284854563691073, + "loss": 1.7447, + "theoretical_loss": 4.216337276270341, + "tokens_seen": 275775488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628385155466399, + "loss": 1.619, + "theoretical_loss": 4.216218921607586, + "tokens_seen": 275841024 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046282848545636915, + "loss": 1.813, + "theoretical_loss": 4.216100602932224, + "tokens_seen": 275906560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628184553660983, + "loss": 1.7676, + "theoretical_loss": 4.21598232022477, + "tokens_seen": 275972096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004628084252758275, + "loss": 1.841, + "theoretical_loss": 4.215864073465749, + "tokens_seen": 276037632 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046279839518555664, + "loss": 1.7917, + "theoretical_loss": 4.215745862635708, + "tokens_seen": 276103168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627883650952859, + "loss": 1.7191, + "theoretical_loss": 4.2156276877152035, + "tokens_seen": 276168704 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046277833500501505, + "loss": 1.892, + "theoretical_loss": 4.21550954868481, + "tokens_seen": 276234240 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046276830491474424, + "loss": 1.7294, + "theoretical_loss": 4.2153914455251185, + "tokens_seen": 276299776 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627582748244734, + "loss": 1.8211, + "theoretical_loss": 4.215273378216731, + "tokens_seen": 276365312 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046274824473420265, + "loss": 1.7903, + "theoretical_loss": 4.215155346740268, + "tokens_seen": 276430848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627382146439318, + "loss": 1.8372, + "theoretical_loss": 4.215037351076363, + "tokens_seen": 276496384 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462728184553661, + "loss": 1.5834, + "theoretical_loss": 4.214919391205665, + "tokens_seen": 276561920 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046271815446339014, + "loss": 1.7651, + "theoretical_loss": 4.21480146710884, + "tokens_seen": 276627456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004627081243731194, + "loss": 1.8217, + "theoretical_loss": 4.214683578766566, + "tokens_seen": 276692992 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046269809428284856, + "loss": 1.7232, + "theoretical_loss": 4.214565726159539, + "tokens_seen": 276758528 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046268806419257774, + "loss": 1.7275, + "theoretical_loss": 4.214447909268467, + "tokens_seen": 276824064 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626780341023069, + "loss": 1.7933, + "theoretical_loss": 4.214330128074074, + "tokens_seen": 276889600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626680040120361, + "loss": 1.6391, + "theoretical_loss": 4.214212382557101, + "tokens_seen": 276955136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626579739217653, + "loss": 1.6952, + "theoretical_loss": 4.214094672698302, + "tokens_seen": 277020672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626479438314945, + "loss": 1.656, + "theoretical_loss": 4.213976998478445, + "tokens_seen": 277086208 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046263791374122364, + "loss": 1.7792, + "theoretical_loss": 4.213859359878316, + "tokens_seen": 277151744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004626278836509529, + "loss": 1.784, + "theoretical_loss": 4.213741756878713, + "tokens_seen": 277217280 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462617853560682, + "loss": 1.685, + "theoretical_loss": 4.213624189460449, + "tokens_seen": 277282816 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046260782347041124, + "loss": 1.7008, + "theoretical_loss": 4.213506657604356, + "tokens_seen": 277348352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625977933801404, + "loss": 1.704, + "theoretical_loss": 4.213389161291273, + "tokens_seen": 277413888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625877632898696, + "loss": 1.6782, + "theoretical_loss": 4.213271700502062, + "tokens_seen": 277479424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625777331995988, + "loss": 1.816, + "theoretical_loss": 4.213154275217597, + "tokens_seen": 277544960 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462567703109328, + "loss": 1.7763, + "theoretical_loss": 4.213036885418763, + "tokens_seen": 277610496 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046255767301905715, + "loss": 1.7162, + "theoretical_loss": 4.212919531086466, + "tokens_seen": 277676032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625476429287864, + "loss": 1.7335, + "theoretical_loss": 4.2128022122016215, + "tokens_seen": 277741568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625376128385155, + "loss": 1.7434, + "theoretical_loss": 4.212684928745163, + "tokens_seen": 277807104 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046252758274824475, + "loss": 1.7753, + "theoretical_loss": 4.212567680698038, + "tokens_seen": 277872640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625175526579739, + "loss": 1.7365, + "theoretical_loss": 4.212450468041208, + "tokens_seen": 277938176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004625075225677031, + "loss": 1.8201, + "theoretical_loss": 4.21233329075565, + "tokens_seen": 278003712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624974924774323, + "loss": 1.6838, + "theoretical_loss": 4.212216148822355, + "tokens_seen": 278069248 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046248746238716147, + "loss": 1.6808, + "theoretical_loss": 4.212099042222331, + "tokens_seen": 278134784 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046247743229689065, + "loss": 1.8519, + "theoretical_loss": 4.211981970936595, + "tokens_seen": 278200320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624674022066199, + "loss": 1.6956, + "theoretical_loss": 4.2118649349461865, + "tokens_seen": 278265856 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046245737211634907, + "loss": 1.7904, + "theoretical_loss": 4.211747934232153, + "tokens_seen": 278331392 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046244734202607825, + "loss": 1.779, + "theoretical_loss": 4.211630968775561, + "tokens_seen": 278396928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624373119358075, + "loss": 1.7246, + "theoretical_loss": 4.211514038557488, + "tokens_seen": 278462464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 107929, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7468466758728027, + "objective/train/theoretical_loss": 4.211397143559029, + "objective/train/tokens_used": 298988000, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004624272818455366, + "loss": 1.753, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046241725175526585, + "loss": 1.7569, + "theoretical_loss": 4.211280283761294, + "tokens_seen": 278593536 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462407221664995, + "loss": 1.7741, + "theoretical_loss": 4.211163459145403, + "tokens_seen": 278659072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623971915747242, + "loss": 1.7807, + "theoretical_loss": 4.211046669692496, + "tokens_seen": 278724608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623871614844534, + "loss": 1.8074, + "theoretical_loss": 4.2109299153837245, + "tokens_seen": 278790144 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046237713139418257, + "loss": 1.8297, + "theoretical_loss": 4.210813196200255, + "tokens_seen": 278855680 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046236710130391175, + "loss": 1.6808, + "theoretical_loss": 4.21069651212327, + "tokens_seen": 278921216 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046235707121364093, + "loss": 1.8495, + "theoretical_loss": 4.2105798631339635, + "tokens_seen": 278986752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623470411233701, + "loss": 1.7331, + "theoretical_loss": 4.2104632492135465, + "tokens_seen": 279052288 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046233701103309935, + "loss": 1.6243, + "theoretical_loss": 4.210346670343244, + "tokens_seen": 279117824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623269809428285, + "loss": 1.727, + "theoretical_loss": 4.210230126504296, + "tokens_seen": 279183360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004623169508525577, + "loss": 1.7862, + "theoretical_loss": 4.210113617677955, + "tokens_seen": 279248896 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046230692076228684, + "loss": 1.8817, + "theoretical_loss": 4.20999714384549, + "tokens_seen": 279314432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622968906720161, + "loss": 1.9009, + "theoretical_loss": 4.209880704988181, + "tokens_seen": 279379968 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046228686058174525, + "loss": 1.7983, + "theoretical_loss": 4.209764301087327, + "tokens_seen": 279445504 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046227683049147444, + "loss": 1.7582, + "theoretical_loss": 4.20964793212424, + "tokens_seen": 279511040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622668004012036, + "loss": 1.7498, + "theoretical_loss": 4.209531598080245, + "tokens_seen": 279576576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046225677031093285, + "loss": 1.8344, + "theoretical_loss": 4.20941529893668, + "tokens_seen": 279642112 + }, + { + "epoch": 0.08, + "learning_rate": 0.000462246740220662, + "loss": 1.653, + "theoretical_loss": 4.209299034674902, + "tokens_seen": 279707648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622367101303912, + "loss": 1.76, + "theoretical_loss": 4.209182805276278, + "tokens_seen": 279773184 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046222668004012034, + "loss": 1.7163, + "theoretical_loss": 4.209066610722192, + "tokens_seen": 279838720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004622166499498496, + "loss": 1.8927, + "theoretical_loss": 4.208950450994041, + "tokens_seen": 279904256 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046220661985957876, + "loss": 1.8409, + "theoretical_loss": 4.2088343260732355, + "tokens_seen": 279969792 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046219658976930794, + "loss": 1.6564, + "theoretical_loss": 4.208718235941204, + "tokens_seen": 280035328 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621865596790371, + "loss": 1.7814, + "theoretical_loss": 4.2086021805793825, + "tokens_seen": 280100864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621765295887663, + "loss": 1.7455, + "theoretical_loss": 4.208486159969229, + "tokens_seen": 280166400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621664994984955, + "loss": 1.6727, + "theoretical_loss": 4.20837017409221, + "tokens_seen": 280231936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621564694082247, + "loss": 1.7125, + "theoretical_loss": 4.20825422292981, + "tokens_seen": 280297472 + }, + { + "epoch": 0.08, + "learning_rate": 0.00046214643931795384, + "loss": 1.7954, + "theoretical_loss": 4.208138306463524, + "tokens_seen": 280363008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621364092276831, + "loss": 1.7675, + "theoretical_loss": 4.208022424674864, + "tokens_seen": 280428544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0004621263791374122, + "loss": 1.8743, + "theoretical_loss": 4.207906577545356, + "tokens_seen": 280494080 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046211634904714144, + "loss": 1.6758, + "theoretical_loss": 4.207790765056538, + "tokens_seen": 280559616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004621063189568706, + "loss": 1.7673, + "theoretical_loss": 4.207674987189965, + "tokens_seen": 280625152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620962888665998, + "loss": 1.7147, + "theoretical_loss": 4.2075592439272045, + "tokens_seen": 280690688 + }, + { + "epoch": 0.09, + "learning_rate": 0.000462086258776329, + "loss": 1.7895, + "theoretical_loss": 4.207443535249838, + "tokens_seen": 280756224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620762286860582, + "loss": 1.6485, + "theoretical_loss": 4.207327861139461, + "tokens_seen": 280821760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046206619859578735, + "loss": 1.6835, + "theoretical_loss": 4.207212221577684, + "tokens_seen": 280887296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620561685055166, + "loss": 1.8451, + "theoretical_loss": 4.207096616546132, + "tokens_seen": 280952832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620461384152457, + "loss": 1.714, + "theoretical_loss": 4.206981046026442, + "tokens_seen": 281018368 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046203610832497495, + "loss": 1.7309, + "theoretical_loss": 4.2068655100002665, + "tokens_seen": 281083904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620260782347041, + "loss": 1.7048, + "theoretical_loss": 4.206750008449272, + "tokens_seen": 281149440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620160481444333, + "loss": 1.7187, + "theoretical_loss": 4.206634541355138, + "tokens_seen": 281214976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004620060180541625, + "loss": 1.6603, + "theoretical_loss": 4.20651910869956, + "tokens_seen": 281280512 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046199598796389167, + "loss": 1.8562, + "theoretical_loss": 4.2064037104642455, + "tokens_seen": 281346048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046198595787362085, + "loss": 1.7743, + "theoretical_loss": 4.206288346630917, + "tokens_seen": 281411584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619759277833501, + "loss": 1.6713, + "theoretical_loss": 4.206173017181309, + "tokens_seen": 281477120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619658976930792, + "loss": 1.6564, + "theoretical_loss": 4.206057722097174, + "tokens_seen": 281542656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046195586760280845, + "loss": 1.6428, + "theoretical_loss": 4.205942461360276, + "tokens_seen": 281608192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046194583751253763, + "loss": 1.7067, + "theoretical_loss": 4.205827234952391, + "tokens_seen": 281673728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619358074222668, + "loss": 1.7988, + "theoretical_loss": 4.205712042855312, + "tokens_seen": 281739264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 109222, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6372196674346924, + "objective/train/theoretical_loss": 4.205596885050846, + "objective/train/tokens_used": 302264800, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461925777331996, + "loss": 1.7245, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004619157472417252, + "loss": 1.7359, + "theoretical_loss": 4.205481761520811, + "tokens_seen": 281870336 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046190571715145435, + "loss": 1.6675, + "theoretical_loss": 4.20536667224704, + "tokens_seen": 281935872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618956870611836, + "loss": 1.7055, + "theoretical_loss": 4.205251617211382, + "tokens_seen": 282001408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618856569709127, + "loss": 1.65, + "theoretical_loss": 4.205136596395698, + "tokens_seen": 282066944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046187562688064195, + "loss": 1.7063, + "theoretical_loss": 4.205021609781861, + "tokens_seen": 282132480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618655967903711, + "loss": 1.7029, + "theoretical_loss": 4.204906657351763, + "tokens_seen": 282198016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618555667001003, + "loss": 1.7243, + "theoretical_loss": 4.2047917390873035, + "tokens_seen": 282263552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618455366098295, + "loss": 1.792, + "theoretical_loss": 4.204676854970401, + "tokens_seen": 282329088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618355065195587, + "loss": 1.7241, + "theoretical_loss": 4.204562004982986, + "tokens_seen": 282394624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046182547642928786, + "loss": 1.6774, + "theoretical_loss": 4.204447189107, + "tokens_seen": 282460160 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046181544633901704, + "loss": 1.7294, + "theoretical_loss": 4.204332407324403, + "tokens_seen": 282525696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004618054162487462, + "loss": 1.7359, + "theoretical_loss": 4.204217659617164, + "tokens_seen": 282591232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046179538615847545, + "loss": 1.7927, + "theoretical_loss": 4.2041029459672705, + "tokens_seen": 282656768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617853560682046, + "loss": 1.6621, + "theoretical_loss": 4.2039882663567205, + "tokens_seen": 282722304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617753259779338, + "loss": 1.7615, + "theoretical_loss": 4.203873620767526, + "tokens_seen": 282787840 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461765295887663, + "loss": 1.7172, + "theoretical_loss": 4.203759009181714, + "tokens_seen": 282853376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617552657973922, + "loss": 1.6926, + "theoretical_loss": 4.203644431581323, + "tokens_seen": 282918912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046174523570712136, + "loss": 1.7332, + "theoretical_loss": 4.203529887948408, + "tokens_seen": 282984448 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046173520561685054, + "loss": 1.735, + "theoretical_loss": 4.203415378265034, + "tokens_seen": 283049984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004617251755265797, + "loss": 1.7046, + "theoretical_loss": 4.203300902513284, + "tokens_seen": 283115520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046171514543630896, + "loss": 1.8186, + "theoretical_loss": 4.203186460675251, + "tokens_seen": 283181056 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046170511534603814, + "loss": 1.6966, + "theoretical_loss": 4.203072052733043, + "tokens_seen": 283246592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616950852557673, + "loss": 1.7019, + "theoretical_loss": 4.202957678668782, + "tokens_seen": 283312128 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616850551654965, + "loss": 1.6874, + "theoretical_loss": 4.202843338464602, + "tokens_seen": 283377664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616750250752257, + "loss": 1.8784, + "theoretical_loss": 4.202729032102653, + "tokens_seen": 283443200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616649949849549, + "loss": 1.7366, + "theoretical_loss": 4.202614759565096, + "tokens_seen": 283508736 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046165496489468404, + "loss": 1.7899, + "theoretical_loss": 4.202500520834106, + "tokens_seen": 283574272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616449348044133, + "loss": 1.7269, + "theoretical_loss": 4.202386315891874, + "tokens_seen": 283639808 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616349047141424, + "loss": 1.7169, + "theoretical_loss": 4.202272144720602, + "tokens_seen": 283705344 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046162487462387164, + "loss": 1.734, + "theoretical_loss": 4.202158007302506, + "tokens_seen": 283770880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004616148445336008, + "loss": 1.7491, + "theoretical_loss": 4.202043903619814, + "tokens_seen": 283836416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046160481444333, + "loss": 1.6737, + "theoretical_loss": 4.201929833654772, + "tokens_seen": 283901952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615947843530592, + "loss": 1.6712, + "theoretical_loss": 4.201815797389634, + "tokens_seen": 283967488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615847542627884, + "loss": 1.6987, + "theoretical_loss": 4.201701794806672, + "tokens_seen": 284033024 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046157472417251755, + "loss": 1.8621, + "theoretical_loss": 4.201587825888167, + "tokens_seen": 284098560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615646940822468, + "loss": 1.6125, + "theoretical_loss": 4.201473890616418, + "tokens_seen": 284164096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615546639919759, + "loss": 1.8721, + "theoretical_loss": 4.201359988973733, + "tokens_seen": 284229632 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046154463390170515, + "loss": 1.7649, + "theoretical_loss": 4.201246120942438, + "tokens_seen": 284295168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615346038114343, + "loss": 1.8232, + "theoretical_loss": 4.201132286504867, + "tokens_seen": 284360704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615245737211635, + "loss": 1.8138, + "theoretical_loss": 4.201018485643373, + "tokens_seen": 284426240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004615145436308927, + "loss": 1.7212, + "theoretical_loss": 4.200904718340317, + "tokens_seen": 284491776 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046150451354062187, + "loss": 1.7844, + "theoretical_loss": 4.200790984578078, + "tokens_seen": 284557312 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046149448345035105, + "loss": 1.7123, + "theoretical_loss": 4.200677284339045, + "tokens_seen": 284622848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614844533600803, + "loss": 1.8333, + "theoretical_loss": 4.2005636176056225, + "tokens_seen": 284688384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614744232698094, + "loss": 1.7715, + "theoretical_loss": 4.200449984360226, + "tokens_seen": 284753920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046146439317953865, + "loss": 1.7826, + "theoretical_loss": 4.200336384585286, + "tokens_seen": 284819456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046145436308926783, + "loss": 1.7551, + "theoretical_loss": 4.200222818263246, + "tokens_seen": 284884992 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461444332998997, + "loss": 1.7554, + "theoretical_loss": 4.200109285376563, + "tokens_seen": 284950528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614343029087262, + "loss": 1.6863, + "theoretical_loss": 4.1999957859077055, + "tokens_seen": 285016064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 110428, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6771039962768555, + "objective/train/theoretical_loss": 4.1998823198391575, + "objective/train/tokens_used": 305541600, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614242728184554, + "loss": 1.73, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046141424272818455, + "loss": 1.7079, + "theoretical_loss": 4.199768887153415, + "tokens_seen": 285147136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004614042126379138, + "loss": 1.8076, + "theoretical_loss": 4.199655487832987, + "tokens_seen": 285212672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613941825476429, + "loss": 1.7907, + "theoretical_loss": 4.199542121860397, + "tokens_seen": 285278208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046138415245737215, + "loss": 1.7309, + "theoretical_loss": 4.199428789218181, + "tokens_seen": 285343744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613741223671013, + "loss": 1.7189, + "theoretical_loss": 4.199315489888885, + "tokens_seen": 285409280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613640922768305, + "loss": 1.7262, + "theoretical_loss": 4.199202223855075, + "tokens_seen": 285474816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613540621865597, + "loss": 1.7436, + "theoretical_loss": 4.199088991099323, + "tokens_seen": 285540352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613440320962889, + "loss": 1.6428, + "theoretical_loss": 4.19897579160422, + "tokens_seen": 285605888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046133400200601806, + "loss": 1.751, + "theoretical_loss": 4.1988626253523655, + "tokens_seen": 285671424 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046132397191574724, + "loss": 1.7595, + "theoretical_loss": 4.198749492326374, + "tokens_seen": 285736960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004613139418254764, + "loss": 1.7753, + "theoretical_loss": 4.198636392508874, + "tokens_seen": 285802496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046130391173520566, + "loss": 1.7111, + "theoretical_loss": 4.198523325882505, + "tokens_seen": 285868032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612938816449348, + "loss": 1.681, + "theoretical_loss": 4.198410292429922, + "tokens_seen": 285933568 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461283851554664, + "loss": 1.6532, + "theoretical_loss": 4.19829729213379, + "tokens_seen": 285999104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612738214643932, + "loss": 1.7138, + "theoretical_loss": 4.198184324976791, + "tokens_seen": 286064640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612637913741224, + "loss": 1.7137, + "theoretical_loss": 4.198071390941616, + "tokens_seen": 286130176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046125376128385156, + "loss": 1.6704, + "theoretical_loss": 4.197958490010971, + "tokens_seen": 286195712 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046124373119358074, + "loss": 1.7474, + "theoretical_loss": 4.197845622167575, + "tokens_seen": 286261248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612337011033099, + "loss": 1.7649, + "theoretical_loss": 4.197732787394161, + "tokens_seen": 286326784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046122367101303916, + "loss": 1.7412, + "theoretical_loss": 4.197619985673471, + "tokens_seen": 286392320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612136409227683, + "loss": 1.6827, + "theoretical_loss": 4.197507216988265, + "tokens_seen": 286457856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004612036108324975, + "loss": 1.6642, + "theoretical_loss": 4.197394481321312, + "tokens_seen": 286523392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046119358074222665, + "loss": 1.8059, + "theoretical_loss": 4.1972817786553955, + "tokens_seen": 286588928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611835506519559, + "loss": 1.7107, + "theoretical_loss": 4.197169108973313, + "tokens_seen": 286654464 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046117352056168506, + "loss": 1.7486, + "theoretical_loss": 4.197056472257874, + "tokens_seen": 286720000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046116349047141425, + "loss": 1.8019, + "theoretical_loss": 4.196943868491899, + "tokens_seen": 286785536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611534603811434, + "loss": 1.7269, + "theoretical_loss": 4.196831297658225, + "tokens_seen": 286851072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611434302908726, + "loss": 1.7408, + "theoretical_loss": 4.1967187597396975, + "tokens_seen": 286916608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611334002006018, + "loss": 1.6744, + "theoretical_loss": 4.196606254719179, + "tokens_seen": 286982144 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461123370110331, + "loss": 1.7665, + "theoretical_loss": 4.196493782579544, + "tokens_seen": 287047680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046111334002006015, + "loss": 1.8181, + "theoretical_loss": 4.196381343303676, + "tokens_seen": 287113216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004611033099297894, + "loss": 1.7154, + "theoretical_loss": 4.1962689368744766, + "tokens_seen": 287178752 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046109327983951857, + "loss": 1.6479, + "theoretical_loss": 4.196156563274858, + "tokens_seen": 287244288 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046108324974924775, + "loss": 1.6657, + "theoretical_loss": 4.1960442224877434, + "tokens_seen": 287309824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046107321965897693, + "loss": 1.6802, + "theoretical_loss": 4.195931914496072, + "tokens_seen": 287375360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610631895687061, + "loss": 1.6672, + "theoretical_loss": 4.195819639282792, + "tokens_seen": 287440896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610531594784353, + "loss": 1.8124, + "theoretical_loss": 4.195707396830869, + "tokens_seen": 287506432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610431293881645, + "loss": 1.6786, + "theoretical_loss": 4.195595187123277, + "tokens_seen": 287571968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046103309929789365, + "loss": 1.9345, + "theoretical_loss": 4.195483010143006, + "tokens_seen": 287637504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004610230692076229, + "loss": 1.8441, + "theoretical_loss": 4.1953708658730555, + "tokens_seen": 287703040 + }, + { + "epoch": 0.09, + "learning_rate": 0.000461013039117352, + "loss": 1.7531, + "theoretical_loss": 4.195258754296441, + "tokens_seen": 287768576 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046100300902708125, + "loss": 1.7673, + "theoretical_loss": 4.195146675396189, + "tokens_seen": 287834112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046099297893681043, + "loss": 1.7202, + "theoretical_loss": 4.195034629155338, + "tokens_seen": 287899648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609829488465396, + "loss": 1.6534, + "theoretical_loss": 4.194922615556941, + "tokens_seen": 287965184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609729187562688, + "loss": 1.732, + "theoretical_loss": 4.194810634584061, + "tokens_seen": 288030720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046096288866599803, + "loss": 1.733, + "theoretical_loss": 4.194698686219778, + "tokens_seen": 288096256 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609528585757272, + "loss": 1.7037, + "theoretical_loss": 4.194586770447179, + "tokens_seen": 288161792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609428284854564, + "loss": 1.6987, + "theoretical_loss": 4.194474887249369, + "tokens_seen": 288227328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609327983951856, + "loss": 1.6782, + "theoretical_loss": 4.19436303660946, + "tokens_seen": 288292864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 111591, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6751337051391602, + "objective/train/theoretical_loss": 4.194251218510583, + "objective/train/tokens_used": 308818400, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046092276830491475, + "loss": 1.6473, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460912738214644, + "loss": 1.7404, + "theoretical_loss": 4.194139432935877, + "tokens_seen": 288423936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004609027081243731, + "loss": 1.7521, + "theoretical_loss": 4.194027679868495, + "tokens_seen": 288489472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046089267803410235, + "loss": 1.6987, + "theoretical_loss": 4.193915959291601, + "tokens_seen": 288555008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608826479438315, + "loss": 1.794, + "theoretical_loss": 4.193804271188376, + "tokens_seen": 288620544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608726178535607, + "loss": 1.8073, + "theoretical_loss": 4.193692615542007, + "tokens_seen": 288686080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608625877632899, + "loss": 1.8507, + "theoretical_loss": 4.193580992335699, + "tokens_seen": 288751616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608525576730191, + "loss": 1.8172, + "theoretical_loss": 4.193469401552667, + "tokens_seen": 288817152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046084252758274826, + "loss": 1.7347, + "theoretical_loss": 4.193357843176139, + "tokens_seen": 288882688 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046083249749247744, + "loss": 1.6888, + "theoretical_loss": 4.193246317189356, + "tokens_seen": 288948224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004608224674022066, + "loss": 1.7896, + "theoretical_loss": 4.19313482357557, + "tokens_seen": 289013760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046081243731193586, + "loss": 1.9245, + "theoretical_loss": 4.193023362318047, + "tokens_seen": 289079296 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460802407221665, + "loss": 1.7733, + "theoretical_loss": 4.1929119334000635, + "tokens_seen": 289144832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607923771313942, + "loss": 1.7208, + "theoretical_loss": 4.192800536804912, + "tokens_seen": 289210368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607823470411234, + "loss": 1.7358, + "theoretical_loss": 4.192689172515893, + "tokens_seen": 289275904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607723169508526, + "loss": 1.7328, + "theoretical_loss": 4.192577840516323, + "tokens_seen": 289341440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046076228686058176, + "loss": 1.7377, + "theoretical_loss": 4.192466540789528, + "tokens_seen": 289406976 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046075225677031094, + "loss": 1.7404, + "theoretical_loss": 4.19235527331885, + "tokens_seen": 289472512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607422266800401, + "loss": 1.702, + "theoretical_loss": 4.19224403808764, + "tokens_seen": 289538048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046073219658976936, + "loss": 1.8364, + "theoretical_loss": 4.192132835079262, + "tokens_seen": 289603584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607221664994985, + "loss": 1.7757, + "theoretical_loss": 4.192021664277094, + "tokens_seen": 289669120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004607121364092277, + "loss": 1.9331, + "theoretical_loss": 4.191910525664524, + "tokens_seen": 289734656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046070210631895685, + "loss": 1.7059, + "theoretical_loss": 4.191799419224955, + "tokens_seen": 289800192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606920762286861, + "loss": 1.9316, + "theoretical_loss": 4.1916883449418, + "tokens_seen": 289865728 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046068204613841526, + "loss": 1.9051, + "theoretical_loss": 4.1915773027984855, + "tokens_seen": 289931264 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046067201604814445, + "loss": 1.8029, + "theoretical_loss": 4.19146629277845, + "tokens_seen": 289996800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606619859578736, + "loss": 1.838, + "theoretical_loss": 4.1913553148651435, + "tokens_seen": 290062336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606519558676028, + "loss": 1.7711, + "theoretical_loss": 4.191244369042029, + "tokens_seen": 290127872 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460641925777332, + "loss": 1.8115, + "theoretical_loss": 4.191133455292583, + "tokens_seen": 290193408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606318956870612, + "loss": 1.8239, + "theoretical_loss": 4.191022573600293, + "tokens_seen": 290258944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046062186559679035, + "loss": 1.8091, + "theoretical_loss": 4.190911723948657, + "tokens_seen": 290324480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004606118355065196, + "loss": 1.7197, + "theoretical_loss": 4.190800906321188, + "tokens_seen": 290390016 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046060180541624877, + "loss": 1.8309, + "theoretical_loss": 4.190690120701412, + "tokens_seen": 290455552 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046059177532597795, + "loss": 1.6697, + "theoretical_loss": 4.190579367072861, + "tokens_seen": 290521088 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046058174523570713, + "loss": 1.7263, + "theoretical_loss": 4.190468645419086, + "tokens_seen": 290586624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605717151454363, + "loss": 1.8109, + "theoretical_loss": 4.190357955723649, + "tokens_seen": 290652160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605616850551655, + "loss": 1.6511, + "theoretical_loss": 4.190247297970121, + "tokens_seen": 290717696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605516549648947, + "loss": 1.7568, + "theoretical_loss": 4.190136672142088, + "tokens_seen": 290783232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046054162487462385, + "loss": 1.8173, + "theoretical_loss": 4.190026078223145, + "tokens_seen": 290848768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605315947843531, + "loss": 1.7668, + "theoretical_loss": 4.189915516196904, + "tokens_seen": 290914304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004605215646940822, + "loss": 1.7497, + "theoretical_loss": 4.189804986046984, + "tokens_seen": 290979840 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046051153460381145, + "loss": 1.8419, + "theoretical_loss": 4.18969448775702, + "tokens_seen": 291045376 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046050150451354063, + "loss": 1.8942, + "theoretical_loss": 4.189584021310657, + "tokens_seen": 291110912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604914744232698, + "loss": 1.7146, + "theoretical_loss": 4.189473586691554, + "tokens_seen": 291176448 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460481444332999, + "loss": 1.7057, + "theoretical_loss": 4.189363183883377, + "tokens_seen": 291241984 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046047141424272823, + "loss": 1.7022, + "theoretical_loss": 4.189252812869811, + "tokens_seen": 291307520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046046138415245736, + "loss": 1.7989, + "theoretical_loss": 4.18914247363455, + "tokens_seen": 291373056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604513540621866, + "loss": 1.7319, + "theoretical_loss": 4.189032166161297, + "tokens_seen": 291438592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604413239719157, + "loss": 1.7014, + "theoretical_loss": 4.188921890433772, + "tokens_seen": 291504128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046043129388164495, + "loss": 1.7416, + "theoretical_loss": 4.1888116464357035, + "tokens_seen": 291569664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 112844, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7077699899673462, + "objective/train/theoretical_loss": 4.188701434150834, + "objective/train/tokens_used": 312095200, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046042126379137414, + "loss": 1.8038, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604112337011033, + "loss": 1.7536, + "theoretical_loss": 4.188591253562918, + "tokens_seen": 291700736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004604012036108325, + "loss": 1.6568, + "theoretical_loss": 4.18848110465572, + "tokens_seen": 291766272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603911735205617, + "loss": 1.6176, + "theoretical_loss": 4.188370987413018, + "tokens_seen": 291831808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046038114343029086, + "loss": 1.8506, + "theoretical_loss": 4.188260901818602, + "tokens_seen": 291897344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603711133400201, + "loss": 1.7808, + "theoretical_loss": 4.1881508478562735, + "tokens_seen": 291962880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603610832497492, + "loss": 1.7446, + "theoretical_loss": 4.188040825509845, + "tokens_seen": 292028416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046035105315947846, + "loss": 1.7531, + "theoretical_loss": 4.1879308347631445, + "tokens_seen": 292093952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603410230692076, + "loss": 1.7696, + "theoretical_loss": 4.187820875600006, + "tokens_seen": 292159488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603309929789368, + "loss": 1.7291, + "theoretical_loss": 4.187710948004281, + "tokens_seen": 292225024 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460320962888666, + "loss": 1.7625, + "theoretical_loss": 4.187601051959831, + "tokens_seen": 292290560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004603109327983952, + "loss": 1.7819, + "theoretical_loss": 4.187491187450527, + "tokens_seen": 292356096 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046030090270812436, + "loss": 1.8002, + "theoretical_loss": 4.187381354460255, + "tokens_seen": 292421632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602908726178536, + "loss": 1.7842, + "theoretical_loss": 4.187271552972911, + "tokens_seen": 292487168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602808425275827, + "loss": 1.7397, + "theoretical_loss": 4.187161782972405, + "tokens_seen": 292552704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046027081243731196, + "loss": 1.748, + "theoretical_loss": 4.187052044442655, + "tokens_seen": 292618240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602607823470411, + "loss": 1.8988, + "theoretical_loss": 4.1869423373675945, + "tokens_seen": 292683776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602507522567703, + "loss": 1.8293, + "theoretical_loss": 4.1868326617311675, + "tokens_seen": 292749312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602407221664995, + "loss": 1.7187, + "theoretical_loss": 4.18672301751733, + "tokens_seen": 292814848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602306920762287, + "loss": 1.7112, + "theoretical_loss": 4.186613404710048, + "tokens_seen": 292880384 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046022066198595787, + "loss": 1.8505, + "theoretical_loss": 4.186503823293303, + "tokens_seen": 292945920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046021063189568705, + "loss": 1.6734, + "theoretical_loss": 4.186394273251084, + "tokens_seen": 293011456 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004602006018054163, + "loss": 1.8131, + "theoretical_loss": 4.186284754567395, + "tokens_seen": 293076992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046019057171514546, + "loss": 1.8493, + "theoretical_loss": 4.1861752672262496, + "tokens_seen": 293142528 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046018054162487465, + "loss": 1.8679, + "theoretical_loss": 4.186065811211675, + "tokens_seen": 293208064 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601705115346038, + "loss": 1.7376, + "theoretical_loss": 4.185956386507708, + "tokens_seen": 293273600 + }, + { + "epoch": 0.09, + "learning_rate": 0.000460160481444333, + "loss": 1.7945, + "theoretical_loss": 4.185846993098398, + "tokens_seen": 293339136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601504513540622, + "loss": 1.7043, + "theoretical_loss": 4.185737630967808, + "tokens_seen": 293404672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601404212637914, + "loss": 1.8728, + "theoretical_loss": 4.18562830010001, + "tokens_seen": 293470208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046013039117352055, + "loss": 1.7321, + "theoretical_loss": 4.185519000479089, + "tokens_seen": 293535744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004601203610832498, + "loss": 1.7587, + "theoretical_loss": 4.18540973208914, + "tokens_seen": 293601280 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046011033099297897, + "loss": 1.7612, + "theoretical_loss": 4.185300494914273, + "tokens_seen": 293666816 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046010030090270815, + "loss": 1.8121, + "theoretical_loss": 4.185191288938606, + "tokens_seen": 293732352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046009027081243733, + "loss": 1.6018, + "theoretical_loss": 4.185082114146271, + "tokens_seen": 293797888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600802407221665, + "loss": 1.7593, + "theoretical_loss": 4.184972970521409, + "tokens_seen": 293863424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600702106318957, + "loss": 1.7636, + "theoretical_loss": 4.184863858048177, + "tokens_seen": 293928960 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046006018054162493, + "loss": 1.7821, + "theoretical_loss": 4.18475477671074, + "tokens_seen": 293994496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046005015045135405, + "loss": 1.6943, + "theoretical_loss": 4.184645726493275, + "tokens_seen": 294060032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600401203610833, + "loss": 1.8314, + "theoretical_loss": 4.184536707379972, + "tokens_seen": 294125568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004600300902708124, + "loss": 1.7154, + "theoretical_loss": 4.184427719355031, + "tokens_seen": 294191104 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046002006018054165, + "loss": 1.8133, + "theoretical_loss": 4.184318762402664, + "tokens_seen": 294256640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046001003009027083, + "loss": 1.7441, + "theoretical_loss": 4.184209836507096, + "tokens_seen": 294322176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00046, + "loss": 1.7069, + "theoretical_loss": 4.18410094165256, + "tokens_seen": 294387712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599899699097292, + "loss": 1.8391, + "theoretical_loss": 4.183992077823305, + "tokens_seen": 294453248 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045997993981945843, + "loss": 1.7286, + "theoretical_loss": 4.183883245003588, + "tokens_seen": 294518784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045996990972918756, + "loss": 1.7911, + "theoretical_loss": 4.183774443177679, + "tokens_seen": 294584320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599598796389168, + "loss": 1.8657, + "theoretical_loss": 4.183665672329861, + "tokens_seen": 294649856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599498495486459, + "loss": 1.7307, + "theoretical_loss": 4.183556932444424, + "tokens_seen": 294715392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045993981945837515, + "loss": 1.7398, + "theoretical_loss": 4.183448223505673, + "tokens_seen": 294780928 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045992978936810434, + "loss": 1.7348, + "theoretical_loss": 4.183339545497924, + "tokens_seen": 294846464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 114046, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7685636281967163, + "objective/train/theoretical_loss": 4.1832308984055055, + "objective/train/tokens_used": 315372000, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599197592778335, + "loss": 1.7597, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004599097291875627, + "loss": 1.6648, + "theoretical_loss": 4.183122282212754, + "tokens_seen": 294977536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598996990972919, + "loss": 1.9014, + "theoretical_loss": 4.183013696904018, + "tokens_seen": 295043072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045988966900702106, + "loss": 1.6793, + "theoretical_loss": 4.182905142463662, + "tokens_seen": 295108608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598796389167503, + "loss": 1.7323, + "theoretical_loss": 4.182796618876057, + "tokens_seen": 295174144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598696088264794, + "loss": 1.8093, + "theoretical_loss": 4.182688126125587, + "tokens_seen": 295239680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045985957873620866, + "loss": 1.7917, + "theoretical_loss": 4.182579664196648, + "tokens_seen": 295305216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598495486459378, + "loss": 1.7184, + "theoretical_loss": 4.182471233073647, + "tokens_seen": 295370752 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459839518555667, + "loss": 1.7672, + "theoretical_loss": 4.182362832741001, + "tokens_seen": 295436288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598294884653962, + "loss": 1.9142, + "theoretical_loss": 4.182254463183139, + "tokens_seen": 295501824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004598194583751254, + "loss": 1.6455, + "theoretical_loss": 4.182146124384504, + "tokens_seen": 295567360 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045980942828485456, + "loss": 1.8299, + "theoretical_loss": 4.182037816329547, + "tokens_seen": 295632896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597993981945838, + "loss": 1.7518, + "theoretical_loss": 4.181929539002731, + "tokens_seen": 295698432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597893681043129, + "loss": 1.7786, + "theoretical_loss": 4.1818212923885305, + "tokens_seen": 295763968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045977933801404216, + "loss": 1.6891, + "theoretical_loss": 4.181713076471432, + "tokens_seen": 295829504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597693079237713, + "loss": 1.7203, + "theoretical_loss": 4.181604891235933, + "tokens_seen": 295895040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597592778335005, + "loss": 1.7009, + "theoretical_loss": 4.181496736666542, + "tokens_seen": 295960576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597492477432297, + "loss": 1.6751, + "theoretical_loss": 4.181388612747778, + "tokens_seen": 296026112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004597392176529589, + "loss": 1.8065, + "theoretical_loss": 4.181280519464172, + "tokens_seen": 296091648 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045972918756268807, + "loss": 1.7155, + "theoretical_loss": 4.181172456800267, + "tokens_seen": 296157184 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045971915747241725, + "loss": 1.7499, + "theoretical_loss": 4.181064424740617, + "tokens_seen": 296222720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045970912738214643, + "loss": 1.7376, + "theoretical_loss": 4.180956423269784, + "tokens_seen": 296288256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045969909729187566, + "loss": 1.6907, + "theoretical_loss": 4.180848452372347, + "tokens_seen": 296353792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596890672016048, + "loss": 1.7661, + "theoretical_loss": 4.180740512032891, + "tokens_seen": 296419328 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459679037111334, + "loss": 1.7294, + "theoretical_loss": 4.180632602236014, + "tokens_seen": 296484864 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045966900702106315, + "loss": 1.7684, + "theoretical_loss": 4.180524722966327, + "tokens_seen": 296550400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596589769307924, + "loss": 1.6849, + "theoretical_loss": 4.180416874208449, + "tokens_seen": 296615936 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045964894684052157, + "loss": 1.8653, + "theoretical_loss": 4.180309055947012, + "tokens_seen": 296681472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045963891675025075, + "loss": 1.7673, + "theoretical_loss": 4.18020126816666, + "tokens_seen": 296747008 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045962888665997993, + "loss": 1.5964, + "theoretical_loss": 4.180093510852045, + "tokens_seen": 296812544 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045961885656970917, + "loss": 1.7756, + "theoretical_loss": 4.179985783987834, + "tokens_seen": 296878080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004596088264794383, + "loss": 1.7359, + "theoretical_loss": 4.179878087558702, + "tokens_seen": 296943616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045959879638916753, + "loss": 1.8246, + "theoretical_loss": 4.179770421549336, + "tokens_seen": 297009152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045958876629889666, + "loss": 1.7098, + "theoretical_loss": 4.179662785944435, + "tokens_seen": 297074688 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595787362086259, + "loss": 1.7797, + "theoretical_loss": 4.179555180728708, + "tokens_seen": 297140224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595687061183551, + "loss": 1.652, + "theoretical_loss": 4.179447605886876, + "tokens_seen": 297205760 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045955867602808425, + "loss": 1.542, + "theoretical_loss": 4.1793400614036695, + "tokens_seen": 297271296 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045954864593781344, + "loss": 1.6272, + "theoretical_loss": 4.179232547263831, + "tokens_seen": 297336832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595386158475426, + "loss": 1.7163, + "theoretical_loss": 4.1791250634521155, + "tokens_seen": 297402368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004595285857572718, + "loss": 1.7457, + "theoretical_loss": 4.179017609953287, + "tokens_seen": 297467904 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045951855566700103, + "loss": 1.7553, + "theoretical_loss": 4.17891018675212, + "tokens_seen": 297533440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045950852557673016, + "loss": 1.718, + "theoretical_loss": 4.178802793833402, + "tokens_seen": 297598976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594984954864594, + "loss": 1.7492, + "theoretical_loss": 4.17869543118193, + "tokens_seen": 297664512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594884653961885, + "loss": 1.6774, + "theoretical_loss": 4.178588098782513, + "tokens_seen": 297730048 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045947843530591776, + "loss": 1.8119, + "theoretical_loss": 4.178480796619971, + "tokens_seen": 297795584 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045946840521564694, + "loss": 1.7151, + "theoretical_loss": 4.178373524679134, + "tokens_seen": 297861120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594583751253761, + "loss": 1.7089, + "theoretical_loss": 4.1782662829448425, + "tokens_seen": 297926656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045944834503510536, + "loss": 1.7951, + "theoretical_loss": 4.17815907140195, + "tokens_seen": 297992192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045943831494483454, + "loss": 1.7379, + "theoretical_loss": 4.178051890035319, + "tokens_seen": 298057728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594282848545637, + "loss": 1.6651, + "theoretical_loss": 4.177944738829825, + "tokens_seen": 298123264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 115247, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6735737323760986, + "objective/train/theoretical_loss": 4.177837617770352, + "objective/train/tokens_used": 318648800, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594182547642929, + "loss": 1.7488, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004594082246740221, + "loss": 1.6502, + "theoretical_loss": 4.1777305268417955, + "tokens_seen": 298254336 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045939819458375126, + "loss": 1.7854, + "theoretical_loss": 4.177623466029064, + "tokens_seen": 298319872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593881644934805, + "loss": 1.8249, + "theoretical_loss": 4.177516435317074, + "tokens_seen": 298385408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593781344032096, + "loss": 1.7849, + "theoretical_loss": 4.177409434690754, + "tokens_seen": 298450944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045936810431293886, + "loss": 1.6794, + "theoretical_loss": 4.177302464135044, + "tokens_seen": 298516480 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459358074222668, + "loss": 1.7807, + "theoretical_loss": 4.177195523634895, + "tokens_seen": 298582016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593480441323972, + "loss": 1.7188, + "theoretical_loss": 4.1770886131752665, + "tokens_seen": 298647552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593380140421264, + "loss": 1.8098, + "theoretical_loss": 4.176981732741131, + "tokens_seen": 298713088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004593279839518556, + "loss": 1.7966, + "theoretical_loss": 4.176874882317472, + "tokens_seen": 298778624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045931795386158476, + "loss": 1.9028, + "theoretical_loss": 4.176768061889283, + "tokens_seen": 298844160 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459307923771314, + "loss": 1.9133, + "theoretical_loss": 4.176661271441568, + "tokens_seen": 298909696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592978936810431, + "loss": 1.7485, + "theoretical_loss": 4.176554510959342, + "tokens_seen": 298975232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045928786359077236, + "loss": 1.7351, + "theoretical_loss": 4.1764477804276305, + "tokens_seen": 299040768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592778335005015, + "loss": 1.6874, + "theoretical_loss": 4.176341079831471, + "tokens_seen": 299106304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592678034102307, + "loss": 1.809, + "theoretical_loss": 4.176234409155912, + "tokens_seen": 299171840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592577733199599, + "loss": 1.7136, + "theoretical_loss": 4.176127768386008, + "tokens_seen": 299237376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004592477432296891, + "loss": 1.7844, + "theoretical_loss": 4.176021157506832, + "tokens_seen": 299302912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045923771313941827, + "loss": 1.7083, + "theoretical_loss": 4.1759145765034615, + "tokens_seen": 299368448 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045922768304914745, + "loss": 1.639, + "theoretical_loss": 4.175808025360988, + "tokens_seen": 299433984 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045921765295887663, + "loss": 1.7633, + "theoretical_loss": 4.175701504064511, + "tokens_seen": 299499520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045920762286860586, + "loss": 1.6579, + "theoretical_loss": 4.175595012599143, + "tokens_seen": 299565056 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459197592778335, + "loss": 1.7101, + "theoretical_loss": 4.175488550950007, + "tokens_seen": 299630592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591875626880642, + "loss": 1.7907, + "theoretical_loss": 4.175382119102235, + "tokens_seen": 299696128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045917753259779335, + "loss": 1.6852, + "theoretical_loss": 4.175275717040972, + "tokens_seen": 299761664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591675025075226, + "loss": 1.7104, + "theoretical_loss": 4.175169344751371, + "tokens_seen": 299827200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045915747241725177, + "loss": 1.8359, + "theoretical_loss": 4.175063002218598, + "tokens_seen": 299892736 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045914744232698095, + "loss": 1.8138, + "theoretical_loss": 4.174956689427829, + "tokens_seen": 299958272 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045913741223671013, + "loss": 1.8247, + "theoretical_loss": 4.17485040636425, + "tokens_seen": 300023808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045912738214643937, + "loss": 1.7234, + "theoretical_loss": 4.174744153013057, + "tokens_seen": 300089344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004591173520561685, + "loss": 1.7831, + "theoretical_loss": 4.174637929359458, + "tokens_seen": 300154880 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045910732196589773, + "loss": 1.683, + "theoretical_loss": 4.174531735388673, + "tokens_seen": 300220416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045909729187562686, + "loss": 1.7035, + "theoretical_loss": 4.174425571085927, + "tokens_seen": 300285952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590872617853561, + "loss": 1.7458, + "theoretical_loss": 4.174319436436462, + "tokens_seen": 300351488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590772316950853, + "loss": 1.6189, + "theoretical_loss": 4.174213331425527, + "tokens_seen": 300417024 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045906720160481445, + "loss": 1.7098, + "theoretical_loss": 4.174107256038383, + "tokens_seen": 300482560 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045905717151454364, + "loss": 1.7541, + "theoretical_loss": 4.1740012102603, + "tokens_seen": 300548096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590471414242728, + "loss": 1.8076, + "theoretical_loss": 4.17389519407656, + "tokens_seen": 300613632 + }, + { + "epoch": 0.09, + "learning_rate": 0.000459037111334002, + "loss": 1.6024, + "theoretical_loss": 4.173789207472455, + "tokens_seen": 300679168 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045902708124373123, + "loss": 1.8573, + "theoretical_loss": 4.173683250433287, + "tokens_seen": 300744704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045901705115346036, + "loss": 1.7435, + "theoretical_loss": 4.17357732294437, + "tokens_seen": 300810240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004590070210631896, + "loss": 1.735, + "theoretical_loss": 4.1734714249910265, + "tokens_seen": 300875776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589969909729187, + "loss": 1.9031, + "theoretical_loss": 4.173365556558591, + "tokens_seen": 300941312 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045898696088264796, + "loss": 1.7382, + "theoretical_loss": 4.1732597176324076, + "tokens_seen": 301006848 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045897693079237714, + "loss": 1.6303, + "theoretical_loss": 4.173153908197831, + "tokens_seen": 301072384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589669007021063, + "loss": 1.8774, + "theoretical_loss": 4.173048128240227, + "tokens_seen": 301137920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589568706118355, + "loss": 1.6654, + "theoretical_loss": 4.172942377744972, + "tokens_seen": 301203456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045894684052156474, + "loss": 1.6541, + "theoretical_loss": 4.17283665669745, + "tokens_seen": 301268992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045893681043129386, + "loss": 1.639, + "theoretical_loss": 4.172730965083061, + "tokens_seen": 301334528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589267803410231, + "loss": 1.7215, + "theoretical_loss": 4.1726253028872105, + "tokens_seen": 301400064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 116418, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6323739290237427, + "objective/train/theoretical_loss": 4.172519670095315, + "objective/train/tokens_used": 321925600, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004589167502507522, + "loss": 1.8442, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045890672016048146, + "loss": 1.6468, + "theoretical_loss": 4.172414066692804, + "tokens_seen": 301531136 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045889669007021064, + "loss": 1.7098, + "theoretical_loss": 4.172308492665114, + "tokens_seen": 301596672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588866599799398, + "loss": 1.7696, + "theoretical_loss": 4.172202947997695, + "tokens_seen": 301662208 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458876629889669, + "loss": 1.6835, + "theoretical_loss": 4.172097432676006, + "tokens_seen": 301727744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588665997993982, + "loss": 1.8659, + "theoretical_loss": 4.171991946685516, + "tokens_seen": 301793280 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045885656970912737, + "loss": 1.6916, + "theoretical_loss": 4.171886490011705, + "tokens_seen": 301858816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588465396188566, + "loss": 1.7195, + "theoretical_loss": 4.171781062640063, + "tokens_seen": 301924352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045883650952858573, + "loss": 1.7178, + "theoretical_loss": 4.17167566455609, + "tokens_seen": 301989888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045882647943831496, + "loss": 1.7391, + "theoretical_loss": 4.171570295745297, + "tokens_seen": 302055424 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045881644934804415, + "loss": 1.9095, + "theoretical_loss": 4.171464956193205, + "tokens_seen": 302120960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004588064192577733, + "loss": 1.7495, + "theoretical_loss": 4.171359645885346, + "tokens_seen": 302186496 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587963891675025, + "loss": 1.7355, + "theoretical_loss": 4.17125436480726, + "tokens_seen": 302252032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587863590772317, + "loss": 1.8128, + "theoretical_loss": 4.171149112944501, + "tokens_seen": 302317568 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045877632898696087, + "loss": 1.705, + "theoretical_loss": 4.171043890282629, + "tokens_seen": 302383104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587662988966901, + "loss": 1.7621, + "theoretical_loss": 4.170938696807218, + "tokens_seen": 302448640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045875626880641923, + "loss": 1.6744, + "theoretical_loss": 4.17083353250385, + "tokens_seen": 302514176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045874623871614847, + "loss": 1.7491, + "theoretical_loss": 4.170728397358118, + "tokens_seen": 302579712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587362086258776, + "loss": 1.7505, + "theoretical_loss": 4.170623291355626, + "tokens_seen": 302645248 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045872617853560683, + "loss": 1.7535, + "theoretical_loss": 4.1705182144819855, + "tokens_seen": 302710784 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458716148445336, + "loss": 1.7462, + "theoretical_loss": 4.170413166722822, + "tokens_seen": 302776320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004587061183550652, + "loss": 1.6871, + "theoretical_loss": 4.170308148063769, + "tokens_seen": 302841856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586960882647944, + "loss": 1.8163, + "theoretical_loss": 4.1702031584904695, + "tokens_seen": 302907392 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045868605817452355, + "loss": 1.6774, + "theoretical_loss": 4.170098197988578, + "tokens_seen": 302972928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586760280842528, + "loss": 1.8311, + "theoretical_loss": 4.169993266543759, + "tokens_seen": 303038464 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045866599799398197, + "loss": 1.733, + "theoretical_loss": 4.169888364141689, + "tokens_seen": 303104000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045865596790371115, + "loss": 1.7289, + "theoretical_loss": 4.169783490768049, + "tokens_seen": 303169536 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045864593781344033, + "loss": 1.7523, + "theoretical_loss": 4.169678646408536, + "tokens_seen": 303235072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045863590772316957, + "loss": 1.6523, + "theoretical_loss": 4.169573831048854, + "tokens_seen": 303300608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004586258776328987, + "loss": 1.8231, + "theoretical_loss": 4.16946904467472, + "tokens_seen": 303366144 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045861584754262793, + "loss": 1.7213, + "theoretical_loss": 4.169364287271858, + "tokens_seen": 303431680 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045860581745235706, + "loss": 1.7167, + "theoretical_loss": 4.169259558826003, + "tokens_seen": 303497216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585957873620863, + "loss": 1.8948, + "theoretical_loss": 4.169154859322901, + "tokens_seen": 303562752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585857572718155, + "loss": 1.8385, + "theoretical_loss": 4.169050188748307, + "tokens_seen": 303628288 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045857572718154465, + "loss": 1.7809, + "theoretical_loss": 4.168945547087987, + "tokens_seen": 303693824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045856569709127384, + "loss": 1.7514, + "theoretical_loss": 4.168840934327717, + "tokens_seen": 303759360 + }, + { + "epoch": 0.09, + "learning_rate": 0.000458555667001003, + "loss": 1.6723, + "theoretical_loss": 4.1687363504532815, + "tokens_seen": 303824896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585456369107322, + "loss": 1.623, + "theoretical_loss": 4.168631795450478, + "tokens_seen": 303890432 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045853560682046143, + "loss": 1.7077, + "theoretical_loss": 4.168527269305112, + "tokens_seen": 303955968 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045852557673019056, + "loss": 1.7915, + "theoretical_loss": 4.168422772002998, + "tokens_seen": 304021504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585155466399198, + "loss": 1.737, + "theoretical_loss": 4.168318303529963, + "tokens_seen": 304087040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004585055165496489, + "loss": 1.7462, + "theoretical_loss": 4.168213863871843, + "tokens_seen": 304152576 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045849548645937816, + "loss": 1.7435, + "theoretical_loss": 4.168109453014484, + "tokens_seen": 304218112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045848545636910734, + "loss": 1.6851, + "theoretical_loss": 4.1680050709437415, + "tokens_seen": 304283648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584754262788365, + "loss": 1.6634, + "theoretical_loss": 4.167900717645482, + "tokens_seen": 304349184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584653961885657, + "loss": 1.679, + "theoretical_loss": 4.16779639310558, + "tokens_seen": 304414720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045845536609829494, + "loss": 1.7298, + "theoretical_loss": 4.167692097309923, + "tokens_seen": 304480256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045844533600802406, + "loss": 1.561, + "theoretical_loss": 4.167587830244408, + "tokens_seen": 304545792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584353059177533, + "loss": 1.6179, + "theoretical_loss": 4.167483591894937, + "tokens_seen": 304611328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004584252758274824, + "loss": 1.8292, + "theoretical_loss": 4.167379382247429, + "tokens_seen": 304676864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 117554, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6285362243652344, + "objective/train/theoretical_loss": 4.167275201287808, + "objective/train/tokens_used": 325202400, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045841524573721166, + "loss": 1.6327, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045840521564694084, + "loss": 1.8165, + "theoretical_loss": 4.16717104900201, + "tokens_seen": 304807936 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045839518555667, + "loss": 1.9179, + "theoretical_loss": 4.167066925375981, + "tokens_seen": 304873472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583851554663992, + "loss": 1.7041, + "theoretical_loss": 4.1669628303956765, + "tokens_seen": 304939008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583751253761284, + "loss": 1.7289, + "theoretical_loss": 4.166858764047062, + "tokens_seen": 305004544 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045836509528585757, + "loss": 1.7396, + "theoretical_loss": 4.166754726316111, + "tokens_seen": 305070080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583550651955868, + "loss": 1.6827, + "theoretical_loss": 4.16665071718881, + "tokens_seen": 305135616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045834503510531593, + "loss": 1.824, + "theoretical_loss": 4.1665467366511555, + "tokens_seen": 305201152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045833500501504516, + "loss": 1.7547, + "theoretical_loss": 4.166442784689149, + "tokens_seen": 305266688 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045832497492477435, + "loss": 1.7571, + "theoretical_loss": 4.166338861288808, + "tokens_seen": 305332224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583149448345035, + "loss": 1.7867, + "theoretical_loss": 4.166234966436155, + "tokens_seen": 305397760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004583049147442327, + "loss": 1.7247, + "theoretical_loss": 4.166131100117226, + "tokens_seen": 305463296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582948846539619, + "loss": 1.7373, + "theoretical_loss": 4.166027262318064, + "tokens_seen": 305528832 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045828485456369107, + "loss": 1.7914, + "theoretical_loss": 4.165923453024724, + "tokens_seen": 305594368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582748244734203, + "loss": 1.7758, + "theoretical_loss": 4.165819672223269, + "tokens_seen": 305659904 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045826479438314943, + "loss": 1.8473, + "theoretical_loss": 4.165715919899773, + "tokens_seen": 305725440 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045825476429287867, + "loss": 1.6507, + "theoretical_loss": 4.165612196040319, + "tokens_seen": 305790976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582447342026078, + "loss": 1.6894, + "theoretical_loss": 4.1655085006310015, + "tokens_seen": 305856512 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045823470411233703, + "loss": 1.8435, + "theoretical_loss": 4.165404833657923, + "tokens_seen": 305922048 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582246740220662, + "loss": 1.743, + "theoretical_loss": 4.165301195107195, + "tokens_seen": 305987584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004582146439317954, + "loss": 1.7833, + "theoretical_loss": 4.1651975849649405, + "tokens_seen": 306053120 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045820461384152457, + "loss": 1.684, + "theoretical_loss": 4.165094003217293, + "tokens_seen": 306118656 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045819458375125375, + "loss": 1.6945, + "theoretical_loss": 4.164990449850393, + "tokens_seen": 306184192 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045818455366098294, + "loss": 1.8713, + "theoretical_loss": 4.164886924850393, + "tokens_seen": 306249728 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045817452357071217, + "loss": 1.7592, + "theoretical_loss": 4.164783428203453, + "tokens_seen": 306315264 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581644934804413, + "loss": 1.6914, + "theoretical_loss": 4.1646799598957465, + "tokens_seen": 306380800 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045815446339017053, + "loss": 1.5983, + "theoretical_loss": 4.164576519913453, + "tokens_seen": 306446336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581444332998997, + "loss": 1.688, + "theoretical_loss": 4.164473108242763, + "tokens_seen": 306511872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581344032096289, + "loss": 1.9784, + "theoretical_loss": 4.164369724869878, + "tokens_seen": 306577408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004581243731193581, + "loss": 1.5192, + "theoretical_loss": 4.164266369781005, + "tokens_seen": 306642944 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045811434302908726, + "loss": 1.7567, + "theoretical_loss": 4.164163042962366, + "tokens_seen": 306708480 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045810431293881644, + "loss": 1.6597, + "theoretical_loss": 4.1640597444001894, + "tokens_seen": 306774016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580942828485457, + "loss": 1.7406, + "theoretical_loss": 4.163956474080715, + "tokens_seen": 306839552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580842527582748, + "loss": 1.7818, + "theoretical_loss": 4.16385323199019, + "tokens_seen": 306905088 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045807422266800404, + "loss": 1.7134, + "theoretical_loss": 4.1637500181148726, + "tokens_seen": 306970624 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045806419257773316, + "loss": 1.8131, + "theoretical_loss": 4.163646832441031, + "tokens_seen": 307036160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580541624874624, + "loss": 1.7649, + "theoretical_loss": 4.163543674954942, + "tokens_seen": 307101696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580441323971916, + "loss": 1.8489, + "theoretical_loss": 4.163440545642894, + "tokens_seen": 307167232 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045803410230692076, + "loss": 1.7858, + "theoretical_loss": 4.163337444491183, + "tokens_seen": 307232768 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045802407221664994, + "loss": 1.7925, + "theoretical_loss": 4.163234371486114, + "tokens_seen": 307298304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580140421263791, + "loss": 1.7065, + "theoretical_loss": 4.163131326614003, + "tokens_seen": 307363840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004580040120361083, + "loss": 1.7153, + "theoretical_loss": 4.163028309861176, + "tokens_seen": 307429376 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045799398194583754, + "loss": 1.8127, + "theoretical_loss": 4.162925321213968, + "tokens_seen": 307494912 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045798395185556667, + "loss": 1.6983, + "theoretical_loss": 4.162822360658723, + "tokens_seen": 307560448 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579739217652959, + "loss": 1.8042, + "theoretical_loss": 4.162719428181794, + "tokens_seen": 307625984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579638916750251, + "loss": 1.6032, + "theoretical_loss": 4.162616523769547, + "tokens_seen": 307691520 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045795386158475426, + "loss": 1.7349, + "theoretical_loss": 4.1625136474083515, + "tokens_seen": 307757056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579438314944835, + "loss": 1.69, + "theoretical_loss": 4.1624107990845935, + "tokens_seen": 307822592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579338014042126, + "loss": 1.6385, + "theoretical_loss": 4.162307978784662, + "tokens_seen": 307888128 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045792377131394186, + "loss": 1.7932, + "theoretical_loss": 4.162205186494962, + "tokens_seen": 307953664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 118729, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7779209613800049, + "objective/train/theoretical_loss": 4.162102422201901, + "objective/train/tokens_used": 328479200, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045791374122367104, + "loss": 1.7149, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004579037111334002, + "loss": 1.8031, + "theoretical_loss": 4.161999685891901, + "tokens_seen": 308084736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578936810431294, + "loss": 1.7687, + "theoretical_loss": 4.161896977551393, + "tokens_seen": 308150272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578836509528586, + "loss": 1.6635, + "theoretical_loss": 4.161794297166815, + "tokens_seen": 308215808 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045787362086258777, + "loss": 1.6686, + "theoretical_loss": 4.161691644724616, + "tokens_seen": 308281344 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457863590772317, + "loss": 1.8219, + "theoretical_loss": 4.161589020211256, + "tokens_seen": 308346880 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045785356068204613, + "loss": 1.6937, + "theoretical_loss": 4.1614864236132, + "tokens_seen": 308412416 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045784353059177536, + "loss": 1.8087, + "theoretical_loss": 4.161383854916926, + "tokens_seen": 308477952 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045783350050150455, + "loss": 1.7514, + "theoretical_loss": 4.161281314108923, + "tokens_seen": 308543488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578234704112337, + "loss": 1.7446, + "theoretical_loss": 4.1611788011756845, + "tokens_seen": 308609024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578134403209629, + "loss": 1.793, + "theoretical_loss": 4.161076316103717, + "tokens_seen": 308674560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004578034102306921, + "loss": 1.7762, + "theoretical_loss": 4.160973858879535, + "tokens_seen": 308740096 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045779338014042127, + "loss": 1.6674, + "theoretical_loss": 4.160871429489663, + "tokens_seen": 308805632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577833500501505, + "loss": 1.7007, + "theoretical_loss": 4.160769027920635, + "tokens_seen": 308871168 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045777331995987963, + "loss": 1.7318, + "theoretical_loss": 4.160666654158993, + "tokens_seen": 308936704 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045776328986960887, + "loss": 1.7488, + "theoretical_loss": 4.160564308191289, + "tokens_seen": 309002240 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457753259779338, + "loss": 1.843, + "theoretical_loss": 4.1604619900040865, + "tokens_seen": 309067776 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045774322968906723, + "loss": 1.7843, + "theoretical_loss": 4.160359699583955, + "tokens_seen": 309133312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577331995987964, + "loss": 1.7252, + "theoretical_loss": 4.160257436917476, + "tokens_seen": 309198848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577231695085256, + "loss": 1.7734, + "theoretical_loss": 4.160155201991238, + "tokens_seen": 309264384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004577131394182548, + "loss": 1.7997, + "theoretical_loss": 4.160052994791841, + "tokens_seen": 309329920 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045770310932798395, + "loss": 1.6944, + "theoretical_loss": 4.159950815305893, + "tokens_seen": 309395456 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045769307923771314, + "loss": 1.6799, + "theoretical_loss": 4.1598486635200125, + "tokens_seen": 309460992 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045768304914744237, + "loss": 1.6406, + "theoretical_loss": 4.159746539420825, + "tokens_seen": 309526528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576730190571715, + "loss": 1.8032, + "theoretical_loss": 4.159644442994967, + "tokens_seen": 309592064 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045766298896690073, + "loss": 1.7368, + "theoretical_loss": 4.159542374229084, + "tokens_seen": 309657600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576529588766299, + "loss": 1.6711, + "theoretical_loss": 4.159440333109832, + "tokens_seen": 309723136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576429287863591, + "loss": 1.6963, + "theoretical_loss": 4.159338319623874, + "tokens_seen": 309788672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576328986960883, + "loss": 1.6373, + "theoretical_loss": 4.1592363337578835, + "tokens_seen": 309854208 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045762286860581746, + "loss": 1.6186, + "theoretical_loss": 4.159134375498542, + "tokens_seen": 309919744 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045761283851554664, + "loss": 1.6934, + "theoretical_loss": 4.159032444832543, + "tokens_seen": 309985280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004576028084252759, + "loss": 1.705, + "theoretical_loss": 4.158930541746587, + "tokens_seen": 310050816 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457592778335005, + "loss": 1.701, + "theoretical_loss": 4.158828666227383, + "tokens_seen": 310116352 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045758274824473424, + "loss": 1.7023, + "theoretical_loss": 4.158726818261652, + "tokens_seen": 310181888 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045757271815446336, + "loss": 1.6697, + "theoretical_loss": 4.158624997836123, + "tokens_seen": 310247424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575626880641926, + "loss": 1.6581, + "theoretical_loss": 4.158523204937531, + "tokens_seen": 310312960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575526579739218, + "loss": 1.7557, + "theoretical_loss": 4.158421439552626, + "tokens_seen": 310378496 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045754262788365096, + "loss": 1.6799, + "theoretical_loss": 4.158319701668162, + "tokens_seen": 310444032 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045753259779338014, + "loss": 1.5919, + "theoretical_loss": 4.158217991270906, + "tokens_seen": 310509568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575225677031093, + "loss": 1.8028, + "theoretical_loss": 4.158116308347632, + "tokens_seen": 310575104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004575125376128385, + "loss": 1.8906, + "theoretical_loss": 4.1580146528851225, + "tokens_seen": 310640640 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045750250752256774, + "loss": 1.6854, + "theoretical_loss": 4.157913024870172, + "tokens_seen": 310706176 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045749247743229687, + "loss": 1.7105, + "theoretical_loss": 4.157811424289582, + "tokens_seen": 310771712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574824473420261, + "loss": 1.6618, + "theoretical_loss": 4.1577098511301624, + "tokens_seen": 310837248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574724172517553, + "loss": 1.8392, + "theoretical_loss": 4.157608305378735, + "tokens_seen": 310902784 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045746238716148446, + "loss": 1.7032, + "theoretical_loss": 4.157506787022128, + "tokens_seen": 310968320 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045745235707121364, + "loss": 1.8757, + "theoretical_loss": 4.15740529604718, + "tokens_seen": 311033856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574423269809428, + "loss": 1.7918, + "theoretical_loss": 4.15730383244074, + "tokens_seen": 311099392 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457432296890672, + "loss": 1.8129, + "theoretical_loss": 4.157202396189662, + "tokens_seen": 311164928 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045742226680040124, + "loss": 1.4955, + "theoretical_loss": 4.157100987280814, + "tokens_seen": 311230464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 120076, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8498172760009766, + "objective/train/theoretical_loss": 4.15699960570107, + "objective/train/tokens_used": 331756000, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045741223671013037, + "loss": 1.7183, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004574022066198596, + "loss": 1.7392, + "theoretical_loss": 4.156898251437312, + "tokens_seen": 311361536 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045739217652958873, + "loss": 1.6874, + "theoretical_loss": 4.156796924476435, + "tokens_seen": 311427072 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045738214643931797, + "loss": 1.7452, + "theoretical_loss": 4.156695624805341, + "tokens_seen": 311492608 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045737211634904715, + "loss": 1.7951, + "theoretical_loss": 4.15659435241094, + "tokens_seen": 311558144 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045736208625877633, + "loss": 1.7371, + "theoretical_loss": 4.15649310728015, + "tokens_seen": 311623680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004573520561685055, + "loss": 1.6923, + "theoretical_loss": 4.156391889399904, + "tokens_seen": 311689216 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045734202607823475, + "loss": 1.5725, + "theoretical_loss": 4.156290698757139, + "tokens_seen": 311754752 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045733199598796387, + "loss": 1.7602, + "theoretical_loss": 4.156189535338799, + "tokens_seen": 311820288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004573219658976931, + "loss": 1.7182, + "theoretical_loss": 4.156088399131844, + "tokens_seen": 311885824 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045731193580742223, + "loss": 1.7896, + "theoretical_loss": 4.155987290123236, + "tokens_seen": 311951360 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045730190571715147, + "loss": 1.7264, + "theoretical_loss": 4.15588620829995, + "tokens_seen": 312016896 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045729187562688065, + "loss": 1.7679, + "theoretical_loss": 4.15578515364897, + "tokens_seen": 312082432 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045728184553660983, + "loss": 1.8018, + "theoretical_loss": 4.155684126157286, + "tokens_seen": 312147968 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457271815446339, + "loss": 1.8268, + "theoretical_loss": 4.1555831258119005, + "tokens_seen": 312213504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572617853560682, + "loss": 1.8954, + "theoretical_loss": 4.155482152599823, + "tokens_seen": 312279040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572517552657974, + "loss": 1.5859, + "theoretical_loss": 4.155381206508071, + "tokens_seen": 312344576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572417251755266, + "loss": 1.7467, + "theoretical_loss": 4.155280287523674, + "tokens_seen": 312410112 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045723169508525574, + "loss": 1.6365, + "theoretical_loss": 4.155179395633667, + "tokens_seen": 312475648 + }, + { + "epoch": 0.09, + "learning_rate": 0.000457221664994985, + "loss": 1.7485, + "theoretical_loss": 4.1550785308250955, + "tokens_seen": 312541184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004572116349047141, + "loss": 1.556, + "theoretical_loss": 4.154977693085016, + "tokens_seen": 312606720 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045720160481444334, + "loss": 1.6838, + "theoretical_loss": 4.154876882400489, + "tokens_seen": 312672256 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045719157472417257, + "loss": 1.7156, + "theoretical_loss": 4.154776098758589, + "tokens_seen": 312737792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571815446339017, + "loss": 1.6348, + "theoretical_loss": 4.154675342146397, + "tokens_seen": 312803328 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045717151454363093, + "loss": 1.7981, + "theoretical_loss": 4.154574612551001, + "tokens_seen": 312868864 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571614844533601, + "loss": 1.7167, + "theoretical_loss": 4.154473909959503, + "tokens_seen": 312934400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571514543630893, + "loss": 1.8256, + "theoretical_loss": 4.154373234359007, + "tokens_seen": 312999936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571414242728185, + "loss": 1.8542, + "theoretical_loss": 4.154272585736632, + "tokens_seen": 313065472 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045713139418254766, + "loss": 1.7683, + "theoretical_loss": 4.154171964079504, + "tokens_seen": 313131008 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045712136409227684, + "loss": 1.5632, + "theoretical_loss": 4.154071369374755, + "tokens_seen": 313196544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571113340020061, + "loss": 1.6558, + "theoretical_loss": 4.15397080160953, + "tokens_seen": 313262080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0004571013039117352, + "loss": 1.8449, + "theoretical_loss": 4.153870260770979, + "tokens_seen": 313327616 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045709127382146444, + "loss": 1.7979, + "theoretical_loss": 4.153769746846265, + "tokens_seen": 313393152 + }, + { + "epoch": 0.09, + "learning_rate": 0.00045708124373119356, + "loss": 1.7176, + "theoretical_loss": 4.153669259822556, + "tokens_seen": 313458688 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570712136409228, + "loss": 1.7315, + "theoretical_loss": 4.153568799687031, + "tokens_seen": 313524224 + }, + { + "epoch": 0.1, + "learning_rate": 0.000457061183550652, + "loss": 1.7531, + "theoretical_loss": 4.153468366426877, + "tokens_seen": 313589760 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045705115346038116, + "loss": 1.7006, + "theoretical_loss": 4.153367960029288, + "tokens_seen": 313655296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045704112337011034, + "loss": 1.7166, + "theoretical_loss": 4.153267580481472, + "tokens_seen": 313720832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570310932798395, + "loss": 1.9519, + "theoretical_loss": 4.15316722777064, + "tokens_seen": 313786368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004570210631895687, + "loss": 1.7564, + "theoretical_loss": 4.153066901884016, + "tokens_seen": 313851904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045701103309929794, + "loss": 1.8335, + "theoretical_loss": 4.15296660280883, + "tokens_seen": 313917440 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045700100300902707, + "loss": 1.7622, + "theoretical_loss": 4.152866330532321, + "tokens_seen": 313982976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569909729187563, + "loss": 1.7645, + "theoretical_loss": 4.152766085041737, + "tokens_seen": 314048512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569809428284855, + "loss": 1.6912, + "theoretical_loss": 4.152665866324337, + "tokens_seen": 314114048 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045697091273821466, + "loss": 1.7683, + "theoretical_loss": 4.152565674367387, + "tokens_seen": 314179584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045696088264794384, + "loss": 1.7509, + "theoretical_loss": 4.15246550915816, + "tokens_seen": 314245120 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456950852557673, + "loss": 1.7163, + "theoretical_loss": 4.15236537068394, + "tokens_seen": 314310656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569408224674022, + "loss": 1.7936, + "theoretical_loss": 4.152265258932019, + "tokens_seen": 314376192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045693079237713144, + "loss": 1.7169, + "theoretical_loss": 4.152165173889697, + "tokens_seen": 314441728 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045692076228686057, + "loss": 1.632, + "theoretical_loss": 4.152065115544284, + "tokens_seen": 314507264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 121169, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7491862773895264, + "objective/train/theoretical_loss": 4.151965083883098, + "objective/train/tokens_used": 335032800, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004569107321965898, + "loss": 1.6729, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045690070210631893, + "loss": 1.712, + "theoretical_loss": 4.151865078893466, + "tokens_seen": 314638336 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045689067201604817, + "loss": 1.6764, + "theoretical_loss": 4.151765100562723, + "tokens_seen": 314703872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045688064192577735, + "loss": 1.6999, + "theoretical_loss": 4.151665148878212, + "tokens_seen": 314769408 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045687061183550653, + "loss": 1.8232, + "theoretical_loss": 4.151565223827288, + "tokens_seen": 314834944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004568605817452357, + "loss": 1.6798, + "theoretical_loss": 4.15146532539731, + "tokens_seen": 314900480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045685055165496495, + "loss": 1.7496, + "theoretical_loss": 4.151365453575648, + "tokens_seen": 314966016 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045684052156469407, + "loss": 1.7245, + "theoretical_loss": 4.151265608349682, + "tokens_seen": 315031552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004568304914744233, + "loss": 1.7816, + "theoretical_loss": 4.151165789706798, + "tokens_seen": 315097088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045682046138415243, + "loss": 1.7426, + "theoretical_loss": 4.151065997634393, + "tokens_seen": 315162624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045681043129388167, + "loss": 1.6897, + "theoretical_loss": 4.1509662321198695, + "tokens_seen": 315228160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045680040120361085, + "loss": 1.8119, + "theoretical_loss": 4.150866493150642, + "tokens_seen": 315293696 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045679037111334003, + "loss": 1.7503, + "theoretical_loss": 4.150766780714131, + "tokens_seen": 315359232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567803410230692, + "loss": 1.716, + "theoretical_loss": 4.150667094797766, + "tokens_seen": 315424768 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567703109327984, + "loss": 1.7032, + "theoretical_loss": 4.150567435388988, + "tokens_seen": 315490304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567602808425276, + "loss": 1.707, + "theoretical_loss": 4.150467802475241, + "tokens_seen": 315555840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567502507522568, + "loss": 1.7129, + "theoretical_loss": 4.150368196043983, + "tokens_seen": 315621376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045674022066198594, + "loss": 1.7331, + "theoretical_loss": 4.150268616082678, + "tokens_seen": 315686912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567301905717152, + "loss": 1.6171, + "theoretical_loss": 4.150169062578798, + "tokens_seen": 315752448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567201604814443, + "loss": 1.7001, + "theoretical_loss": 4.150069535519825, + "tokens_seen": 315817984 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045671013039117354, + "loss": 1.6913, + "theoretical_loss": 4.149970034893249, + "tokens_seen": 315883520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004567001003009027, + "loss": 1.7334, + "theoretical_loss": 4.149870560686567, + "tokens_seen": 315949056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566900702106319, + "loss": 2.052, + "theoretical_loss": 4.149771112887288, + "tokens_seen": 316014592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566800401203611, + "loss": 1.7827, + "theoretical_loss": 4.149671691482926, + "tokens_seen": 316080128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566700100300903, + "loss": 1.6888, + "theoretical_loss": 4.149572296461005, + "tokens_seen": 316145664 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045665997993981944, + "loss": 1.6652, + "theoretical_loss": 4.149472927809057, + "tokens_seen": 316211200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566499498495487, + "loss": 1.7076, + "theoretical_loss": 4.149373585514625, + "tokens_seen": 316276736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566399197592778, + "loss": 1.5902, + "theoretical_loss": 4.149274269565256, + "tokens_seen": 316342272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045662988966900704, + "loss": 1.6942, + "theoretical_loss": 4.149174979948507, + "tokens_seen": 316407808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566198595787362, + "loss": 1.5834, + "theoretical_loss": 4.1490757166519465, + "tokens_seen": 316473344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004566098294884654, + "loss": 1.8103, + "theoretical_loss": 4.148976479663148, + "tokens_seen": 316538880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565997993981946, + "loss": 1.7062, + "theoretical_loss": 4.148877268969694, + "tokens_seen": 316604416 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045658976930792376, + "loss": 1.7728, + "theoretical_loss": 4.1487780845591775, + "tokens_seen": 316669952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045657973921765294, + "loss": 1.6899, + "theoretical_loss": 4.148678926419197, + "tokens_seen": 316735488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565697091273822, + "loss": 1.7059, + "theoretical_loss": 4.148579794537362, + "tokens_seen": 316801024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565596790371113, + "loss": 1.7936, + "theoretical_loss": 4.148480688901288, + "tokens_seen": 316866560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045654964894684054, + "loss": 1.6922, + "theoretical_loss": 4.148381609498602, + "tokens_seen": 316932096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045653961885656967, + "loss": 1.7304, + "theoretical_loss": 4.148282556316936, + "tokens_seen": 316997632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565295887662989, + "loss": 1.7689, + "theoretical_loss": 4.148183529343932, + "tokens_seen": 317063168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004565195586760281, + "loss": 1.6766, + "theoretical_loss": 4.148084528567241, + "tokens_seen": 317128704 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045650952858575727, + "loss": 1.7009, + "theoretical_loss": 4.147985553974522, + "tokens_seen": 317194240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045649949849548645, + "loss": 1.6719, + "theoretical_loss": 4.14788660555344, + "tokens_seen": 317259776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564894684052157, + "loss": 1.6964, + "theoretical_loss": 4.147787683291672, + "tokens_seen": 317325312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564794383149448, + "loss": 1.6682, + "theoretical_loss": 4.147688787176902, + "tokens_seen": 317390848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045646940822467405, + "loss": 1.8279, + "theoretical_loss": 4.147589917196821, + "tokens_seen": 317456384 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045645937813440317, + "loss": 1.7187, + "theoretical_loss": 4.14749107333913, + "tokens_seen": 317521920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004564493480441324, + "loss": 1.6058, + "theoretical_loss": 4.147392255591538, + "tokens_seen": 317587456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045643931795386164, + "loss": 1.6786, + "theoretical_loss": 4.147293463941762, + "tokens_seen": 317652992 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045642928786359077, + "loss": 1.6614, + "theoretical_loss": 4.147194698377526, + "tokens_seen": 317718528 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045641925777332, + "loss": 1.8719, + "theoretical_loss": 4.147095958886566, + "tokens_seen": 317784064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 122440, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5624806880950928, + "objective/train/theoretical_loss": 4.146997245456623, + "objective/train/tokens_used": 338309600, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045640922768304913, + "loss": 1.7126, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045639919759277837, + "loss": 1.7863, + "theoretical_loss": 4.146898558075446, + "tokens_seen": 317915136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045638916750250755, + "loss": 1.6925, + "theoretical_loss": 4.146799896730794, + "tokens_seen": 317980672 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045637913741223673, + "loss": 1.6439, + "theoretical_loss": 4.146701261410435, + "tokens_seen": 318046208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004563691073219659, + "loss": 1.5878, + "theoretical_loss": 4.1466026521021435, + "tokens_seen": 318111744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045635907723169515, + "loss": 1.7604, + "theoretical_loss": 4.146504068793703, + "tokens_seen": 318177280 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045634904714142427, + "loss": 1.5894, + "theoretical_loss": 4.146405511472905, + "tokens_seen": 318242816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004563390170511535, + "loss": 1.683, + "theoretical_loss": 4.146306980127548, + "tokens_seen": 318308352 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045632898696088263, + "loss": 1.69, + "theoretical_loss": 4.146208474745442, + "tokens_seen": 318373888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045631895687061187, + "loss": 1.6703, + "theoretical_loss": 4.1461099953144025, + "tokens_seen": 318439424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045630892678034105, + "loss": 1.7591, + "theoretical_loss": 4.146011541822253, + "tokens_seen": 318504960 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045629889669007023, + "loss": 1.8026, + "theoretical_loss": 4.145913114256829, + "tokens_seen": 318570496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562888665997994, + "loss": 1.7543, + "theoretical_loss": 4.145814712605969, + "tokens_seen": 318636032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562788365095286, + "loss": 1.7388, + "theoretical_loss": 4.1457163368575225, + "tokens_seen": 318701568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562688064192578, + "loss": 1.6626, + "theoretical_loss": 4.145617986999348, + "tokens_seen": 318767104 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456258776328987, + "loss": 1.7602, + "theoretical_loss": 4.145519663019309, + "tokens_seen": 318832640 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045624874623871614, + "loss": 1.7263, + "theoretical_loss": 4.145421364905282, + "tokens_seen": 318898176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562387161484454, + "loss": 1.8426, + "theoretical_loss": 4.145323092645146, + "tokens_seen": 318963712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562286860581745, + "loss": 1.6604, + "theoretical_loss": 4.145224846226793, + "tokens_seen": 319029248 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045621865596790374, + "loss": 1.7357, + "theoretical_loss": 4.14512662563812, + "tokens_seen": 319094784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004562086258776329, + "loss": 1.6934, + "theoretical_loss": 4.145028430867035, + "tokens_seen": 319160320 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561985957873621, + "loss": 1.729, + "theoretical_loss": 4.144930261901451, + "tokens_seen": 319225856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561885656970913, + "loss": 1.5794, + "theoretical_loss": 4.144832118729291, + "tokens_seen": 319291392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561785356068205, + "loss": 1.711, + "theoretical_loss": 4.144734001338486, + "tokens_seen": 319356928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045616850551654964, + "loss": 1.7424, + "theoretical_loss": 4.144635909716976, + "tokens_seen": 319422464 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561584754262789, + "loss": 1.7061, + "theoretical_loss": 4.144537843852705, + "tokens_seen": 319488000 + }, + { + "epoch": 0.1, + "learning_rate": 0.000456148445336008, + "loss": 1.8684, + "theoretical_loss": 4.144439803733631, + "tokens_seen": 319553536 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045613841524573724, + "loss": 1.7268, + "theoretical_loss": 4.144341789347715, + "tokens_seen": 319619072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561283851554664, + "loss": 1.8256, + "theoretical_loss": 4.144243800682931, + "tokens_seen": 319684608 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561183550651956, + "loss": 1.5942, + "theoretical_loss": 4.1441458377272555, + "tokens_seen": 319750144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004561083249749248, + "loss": 1.6215, + "theoretical_loss": 4.144047900468678, + "tokens_seen": 319815680 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045609829488465396, + "loss": 1.8891, + "theoretical_loss": 4.143949988895193, + "tokens_seen": 319881216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045608826479438314, + "loss": 1.7131, + "theoretical_loss": 4.1438521029948046, + "tokens_seen": 319946752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560782347041124, + "loss": 1.7357, + "theoretical_loss": 4.1437542427555245, + "tokens_seen": 320012288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560682046138415, + "loss": 1.7275, + "theoretical_loss": 4.143656408165373, + "tokens_seen": 320077824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045605817452357074, + "loss": 1.8187, + "theoretical_loss": 4.143558599212376, + "tokens_seen": 320143360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045604814443329987, + "loss": 1.8226, + "theoretical_loss": 4.143460815884571, + "tokens_seen": 320208896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560381143430291, + "loss": 1.8315, + "theoretical_loss": 4.143363058170001, + "tokens_seen": 320274432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004560280842527583, + "loss": 1.6252, + "theoretical_loss": 4.143265326056719, + "tokens_seen": 320339968 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045601805416248747, + "loss": 1.6917, + "theoretical_loss": 4.1431676195327825, + "tokens_seen": 320405504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045600802407221665, + "loss": 1.6254, + "theoretical_loss": 4.143069938586263, + "tokens_seen": 320471040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559979939819459, + "loss": 1.6829, + "theoretical_loss": 4.142972283205233, + "tokens_seen": 320536576 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455987963891675, + "loss": 1.7689, + "theoretical_loss": 4.142874653377778, + "tokens_seen": 320602112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045597793380140425, + "loss": 1.8006, + "theoretical_loss": 4.14277704909199, + "tokens_seen": 320667648 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045596790371113337, + "loss": 1.7255, + "theoretical_loss": 4.142679470335969, + "tokens_seen": 320733184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559578736208626, + "loss": 1.6849, + "theoretical_loss": 4.142581917097821, + "tokens_seen": 320798720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559478435305918, + "loss": 1.6475, + "theoretical_loss": 4.142484389365664, + "tokens_seen": 320864256 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045593781344032097, + "loss": 1.69, + "theoretical_loss": 4.14238688712762, + "tokens_seen": 320929792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045592778335005015, + "loss": 1.6209, + "theoretical_loss": 4.142289410371823, + "tokens_seen": 320995328 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045591775325977933, + "loss": 1.6413, + "theoretical_loss": 4.14219195908641, + "tokens_seen": 321060864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 123582, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7173292636871338, + "objective/train/theoretical_loss": 4.14209453325953, + "objective/train/tokens_used": 341586400, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004559077231695085, + "loss": 1.7329, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045589769307923775, + "loss": 1.7201, + "theoretical_loss": 4.1419971328793395, + "tokens_seen": 321191936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558876629889669, + "loss": 1.5017, + "theoretical_loss": 4.141899757934, + "tokens_seen": 321257472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558776328986961, + "loss": 1.5995, + "theoretical_loss": 4.141802408411683, + "tokens_seen": 321323008 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045586760280842524, + "loss": 1.6794, + "theoretical_loss": 4.14170508430057, + "tokens_seen": 321388544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558575727181545, + "loss": 1.7757, + "theoretical_loss": 4.141607785588846, + "tokens_seen": 321454080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045584754262788365, + "loss": 1.7082, + "theoretical_loss": 4.141510512264707, + "tokens_seen": 321519616 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045583751253761284, + "loss": 1.5862, + "theoretical_loss": 4.141413264316356, + "tokens_seen": 321585152 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455827482447342, + "loss": 1.7405, + "theoretical_loss": 4.141316041732004, + "tokens_seen": 321650688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045581745235707125, + "loss": 1.7558, + "theoretical_loss": 4.141218844499869, + "tokens_seen": 321716224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004558074222668004, + "loss": 1.7445, + "theoretical_loss": 4.141121672608178, + "tokens_seen": 321781760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557973921765296, + "loss": 1.6734, + "theoretical_loss": 4.141024526045166, + "tokens_seen": 321847296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045578736208625874, + "loss": 1.607, + "theoretical_loss": 4.140927404799075, + "tokens_seen": 321912832 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455777331995988, + "loss": 1.7078, + "theoretical_loss": 4.140830308858154, + "tokens_seen": 321978368 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045576730190571716, + "loss": 1.8387, + "theoretical_loss": 4.140733238210664, + "tokens_seen": 322043904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045575727181544634, + "loss": 1.7089, + "theoretical_loss": 4.140636192844867, + "tokens_seen": 322109440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557472417251755, + "loss": 1.6483, + "theoretical_loss": 4.1405391727490395, + "tokens_seen": 322174976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557372116349047, + "loss": 1.7164, + "theoretical_loss": 4.140442177911462, + "tokens_seen": 322240512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557271815446339, + "loss": 1.8218, + "theoretical_loss": 4.140345208320425, + "tokens_seen": 322306048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004557171514543631, + "loss": 1.5315, + "theoretical_loss": 4.1402482639642235, + "tokens_seen": 322371584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045570712136409224, + "loss": 1.7184, + "theoretical_loss": 4.140151344831164, + "tokens_seen": 322437120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556970912738215, + "loss": 1.6715, + "theoretical_loss": 4.14005445090956, + "tokens_seen": 322502656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556870611835507, + "loss": 1.741, + "theoretical_loss": 4.13995758218773, + "tokens_seen": 322568192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045567703109327984, + "loss": 1.756, + "theoretical_loss": 4.1398607386540025, + "tokens_seen": 322633728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556670010030091, + "loss": 1.6785, + "theoretical_loss": 4.139763920296716, + "tokens_seen": 322699264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556569709127382, + "loss": 1.8291, + "theoretical_loss": 4.139667127104211, + "tokens_seen": 322764800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045564694082246744, + "loss": 1.6839, + "theoretical_loss": 4.139570359064842, + "tokens_seen": 322830336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556369107321966, + "loss": 1.6805, + "theoretical_loss": 4.139473616166967, + "tokens_seen": 322895872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004556268806419258, + "loss": 1.6012, + "theoretical_loss": 4.139376898398953, + "tokens_seen": 322961408 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455616850551655, + "loss": 1.6654, + "theoretical_loss": 4.139280205749175, + "tokens_seen": 323026944 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045560682046138416, + "loss": 1.6241, + "theoretical_loss": 4.139183538206017, + "tokens_seen": 323092480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045559679037111334, + "loss": 1.6418, + "theoretical_loss": 4.139086895757868, + "tokens_seen": 323158016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555867602808426, + "loss": 1.7281, + "theoretical_loss": 4.138990278393125, + "tokens_seen": 323223552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555767301905717, + "loss": 1.606, + "theoretical_loss": 4.138893686100196, + "tokens_seen": 323289088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045556670010030094, + "loss": 1.8253, + "theoretical_loss": 4.138797118867493, + "tokens_seen": 323354624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045555667001003007, + "loss": 1.5913, + "theoretical_loss": 4.138700576683437, + "tokens_seen": 323420160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555466399197593, + "loss": 1.6741, + "theoretical_loss": 4.138604059536458, + "tokens_seen": 323485696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555366098294885, + "loss": 1.6781, + "theoretical_loss": 4.138507567414993, + "tokens_seen": 323551232 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045552657973921767, + "loss": 1.7926, + "theoretical_loss": 4.138411100307485, + "tokens_seen": 323616768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045551654964894685, + "loss": 1.542, + "theoretical_loss": 4.138314658202385, + "tokens_seen": 323682304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004555065195586761, + "loss": 1.7132, + "theoretical_loss": 4.138218241088156, + "tokens_seen": 323747840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554964894684052, + "loss": 1.5793, + "theoretical_loss": 4.138121848953261, + "tokens_seen": 323813376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045548645937813445, + "loss": 1.6525, + "theoretical_loss": 4.138025481786178, + "tokens_seen": 323878912 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045547642928786357, + "loss": 1.6685, + "theoretical_loss": 4.137929139575387, + "tokens_seen": 323944448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554663991975928, + "loss": 1.6841, + "theoretical_loss": 4.137832822309382, + "tokens_seen": 324009984 + }, + { + "epoch": 0.1, + "learning_rate": 0.000455456369107322, + "loss": 1.7289, + "theoretical_loss": 4.1377365299766575, + "tokens_seen": 324075520 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045544633901705117, + "loss": 1.575, + "theoretical_loss": 4.13764026256572, + "tokens_seen": 324141056 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045543630892678035, + "loss": 1.6708, + "theoretical_loss": 4.137544020065082, + "tokens_seen": 324206592 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045542627883650953, + "loss": 1.5911, + "theoretical_loss": 4.1374478024632655, + "tokens_seen": 324272128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004554162487462387, + "loss": 1.6946, + "theoretical_loss": 4.137351609748798, + "tokens_seen": 324337664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 124782, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.620722770690918, + "objective/train/theoretical_loss": 4.137255441910215, + "objective/train/tokens_used": 344863200, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045540621865596795, + "loss": 1.6595, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553961885656971, + "loss": 1.6558, + "theoretical_loss": 4.13715929893606, + "tokens_seen": 324468736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553861584754263, + "loss": 1.7144, + "theoretical_loss": 4.137063180814886, + "tokens_seen": 324534272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045537612838515544, + "loss": 1.7521, + "theoretical_loss": 4.136967087535249, + "tokens_seen": 324599808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553660982948847, + "loss": 1.6373, + "theoretical_loss": 4.136871019085717, + "tokens_seen": 324665344 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045535606820461385, + "loss": 1.6654, + "theoretical_loss": 4.136774975454863, + "tokens_seen": 324730880 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045534603811434304, + "loss": 1.8092, + "theoretical_loss": 4.136678956631269, + "tokens_seen": 324796416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553360080240722, + "loss": 1.6764, + "theoretical_loss": 4.136582962603523, + "tokens_seen": 324861952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045532597793380145, + "loss": 1.6458, + "theoretical_loss": 4.136486993360222, + "tokens_seen": 324927488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553159478435306, + "loss": 1.7131, + "theoretical_loss": 4.136391048889971, + "tokens_seen": 324993024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004553059177532598, + "loss": 1.6305, + "theoretical_loss": 4.136295129181381, + "tokens_seen": 325058560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045529588766298894, + "loss": 1.5984, + "theoretical_loss": 4.136199234223069, + "tokens_seen": 325124096 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552858575727182, + "loss": 1.7603, + "theoretical_loss": 4.136103364003665, + "tokens_seen": 325189632 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045527582748244736, + "loss": 1.6444, + "theoretical_loss": 4.1360075185118, + "tokens_seen": 325255168 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045526579739217654, + "loss": 1.7176, + "theoretical_loss": 4.135911697736118, + "tokens_seen": 325320704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552557673019057, + "loss": 1.7028, + "theoretical_loss": 4.135815901665267, + "tokens_seen": 325386240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552457372116349, + "loss": 1.649, + "theoretical_loss": 4.135720130287904, + "tokens_seen": 325451776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552357071213641, + "loss": 1.703, + "theoretical_loss": 4.135624383592694, + "tokens_seen": 325517312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552256770310933, + "loss": 1.6367, + "theoretical_loss": 4.135528661568307, + "tokens_seen": 325582848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045521564694082244, + "loss": 1.7659, + "theoretical_loss": 4.135432964203422, + "tokens_seen": 325648384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004552056168505517, + "loss": 1.6576, + "theoretical_loss": 4.135337291486727, + "tokens_seen": 325713920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551955867602808, + "loss": 1.6824, + "theoretical_loss": 4.135241643406915, + "tokens_seen": 325779456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045518555667001004, + "loss": 1.7325, + "theoretical_loss": 4.135146019952688, + "tokens_seen": 325844992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551755265797392, + "loss": 1.6134, + "theoretical_loss": 4.135050421112755, + "tokens_seen": 325910528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551654964894684, + "loss": 1.7458, + "theoretical_loss": 4.134954846875832, + "tokens_seen": 325976064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551554663991976, + "loss": 1.7831, + "theoretical_loss": 4.134859297230644, + "tokens_seen": 326041600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551454363089268, + "loss": 1.6003, + "theoretical_loss": 4.13476377216592, + "tokens_seen": 326107136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045513540621865595, + "loss": 1.7399, + "theoretical_loss": 4.134668271670401, + "tokens_seen": 326172672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551253761283852, + "loss": 1.6814, + "theoretical_loss": 4.134572795732832, + "tokens_seen": 326238208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004551153460381143, + "loss": 1.6, + "theoretical_loss": 4.134477344341967, + "tokens_seen": 326303744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045510531594784354, + "loss": 1.7193, + "theoretical_loss": 4.134381917486565, + "tokens_seen": 326369280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550952858575727, + "loss": 1.7188, + "theoretical_loss": 4.134286515155398, + "tokens_seen": 326434816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550852557673019, + "loss": 1.6023, + "theoretical_loss": 4.134191137337238, + "tokens_seen": 326500352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550752256770311, + "loss": 1.7086, + "theoretical_loss": 4.134095784020871, + "tokens_seen": 326565888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045506519558676027, + "loss": 1.7665, + "theoretical_loss": 4.1340004551950855, + "tokens_seen": 326631424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045505516549648945, + "loss": 1.5859, + "theoretical_loss": 4.133905150848681, + "tokens_seen": 326696960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550451354062187, + "loss": 1.6368, + "theoretical_loss": 4.133809870970461, + "tokens_seen": 326762496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550351053159478, + "loss": 1.6353, + "theoretical_loss": 4.1337146155492395, + "tokens_seen": 326828032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045502507522567705, + "loss": 1.5662, + "theoretical_loss": 4.133619384573836, + "tokens_seen": 326893568 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045501504513540623, + "loss": 1.5735, + "theoretical_loss": 4.1335241780330785, + "tokens_seen": 326959104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004550050150451354, + "loss": 1.6659, + "theoretical_loss": 4.1334289959158, + "tokens_seen": 327024640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549949849548646, + "loss": 1.6738, + "theoretical_loss": 4.133333838210844, + "tokens_seen": 327090176 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045498495486459377, + "loss": 1.7786, + "theoretical_loss": 4.133238704907059, + "tokens_seen": 327155712 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045497492477432295, + "loss": 1.9242, + "theoretical_loss": 4.133143595993303, + "tokens_seen": 327221248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549648946840522, + "loss": 1.719, + "theoretical_loss": 4.133048511458438, + "tokens_seen": 327286784 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045495486459378137, + "loss": 1.7246, + "theoretical_loss": 4.132953451291337, + "tokens_seen": 327352320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045494483450351055, + "loss": 1.7151, + "theoretical_loss": 4.132858415480877, + "tokens_seen": 327417856 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045493480441323973, + "loss": 1.7997, + "theoretical_loss": 4.132763404015946, + "tokens_seen": 327483392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549247743229689, + "loss": 1.6786, + "theoretical_loss": 4.132668416885435, + "tokens_seen": 327548928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045491474423269815, + "loss": 1.7069, + "theoretical_loss": 4.132573454078246, + "tokens_seen": 327614464 + }, + { + "debugging/distinct-1-grams": 0.7246169962973871, + "debugging/entropy-1-grams": 5.13190351820591, + "debugging/length": 472.8, + "debugging/num_segments": 10, + "debugging/score": 0.0033104506818385365, + "debugging/score_std": 0.0017105448100298692, + "epoch": 0.1, + "objective/train/docs_used": 126013, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7680082321166992, + "objective/train/theoretical_loss": 4.132478515583286, + "objective/train/tokens_used": 348140000, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004549047141424273, + "loss": 1.7889, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548946840521565, + "loss": 1.6904, + "theoretical_loss": 4.132383601389471, + "tokens_seen": 327745536 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045488465396188564, + "loss": 1.7808, + "theoretical_loss": 4.132288711485721, + "tokens_seen": 327811072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548746238716149, + "loss": 1.7723, + "theoretical_loss": 4.132193845860969, + "tokens_seen": 327876608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045486459378134405, + "loss": 1.6767, + "theoretical_loss": 4.132099004504149, + "tokens_seen": 327942144 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045485456369107324, + "loss": 1.8017, + "theoretical_loss": 4.132004187404206, + "tokens_seen": 328007680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548445336008024, + "loss": 1.6662, + "theoretical_loss": 4.131909394550092, + "tokens_seen": 328073216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045483450351053165, + "loss": 1.7874, + "theoretical_loss": 4.131814625930764, + "tokens_seen": 328138752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004548244734202608, + "loss": 1.6567, + "theoretical_loss": 4.13171988153519, + "tokens_seen": 328204288 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045481444332999, + "loss": 1.5981, + "theoretical_loss": 4.131625161352342, + "tokens_seen": 328269824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045480441323971914, + "loss": 1.9143, + "theoretical_loss": 4.131530465371199, + "tokens_seen": 328335360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547943831494484, + "loss": 1.6907, + "theoretical_loss": 4.131435793580749, + "tokens_seen": 328400896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045478435305917756, + "loss": 1.6019, + "theoretical_loss": 4.131341145969989, + "tokens_seen": 328466432 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045477432296890674, + "loss": 1.5531, + "theoretical_loss": 4.131246522527918, + "tokens_seen": 328531968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547642928786359, + "loss": 1.7228, + "theoretical_loss": 4.1311519232435465, + "tokens_seen": 328597504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547542627883651, + "loss": 1.8466, + "theoretical_loss": 4.131057348105889, + "tokens_seen": 328663040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547442326980943, + "loss": 1.5476, + "theoretical_loss": 4.1309627971039715, + "tokens_seen": 328728576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547342026078235, + "loss": 1.7047, + "theoretical_loss": 4.130868270226824, + "tokens_seen": 328794112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045472417251755264, + "loss": 1.6019, + "theoretical_loss": 4.130773767463483, + "tokens_seen": 328859648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004547141424272819, + "loss": 1.6013, + "theoretical_loss": 4.130679288802993, + "tokens_seen": 328925184 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454704112337011, + "loss": 1.697, + "theoretical_loss": 4.1305848342344085, + "tokens_seen": 328990720 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045469408224674024, + "loss": 1.7675, + "theoretical_loss": 4.130490403746787, + "tokens_seen": 329056256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546840521564694, + "loss": 1.6565, + "theoretical_loss": 4.130395997329196, + "tokens_seen": 329121792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546740220661986, + "loss": 1.7141, + "theoretical_loss": 4.1303016149707075, + "tokens_seen": 329187328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546639919759278, + "loss": 1.6539, + "theoretical_loss": 4.130207256660404, + "tokens_seen": 329252864 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454653961885657, + "loss": 1.7045, + "theoretical_loss": 4.130112922387372, + "tokens_seen": 329318400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045464393179538615, + "loss": 1.6776, + "theoretical_loss": 4.130018612140707, + "tokens_seen": 329383936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546339017051154, + "loss": 1.6742, + "theoretical_loss": 4.12992432590951, + "tokens_seen": 329449472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546238716148445, + "loss": 1.6644, + "theoretical_loss": 4.129830063682892, + "tokens_seen": 329515008 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045461384152457374, + "loss": 1.6316, + "theoretical_loss": 4.129735825449968, + "tokens_seen": 329580544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004546038114343029, + "loss": 1.6462, + "theoretical_loss": 4.129641611199862, + "tokens_seen": 329646080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545937813440321, + "loss": 1.8083, + "theoretical_loss": 4.1295474209217025, + "tokens_seen": 329711616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545837512537613, + "loss": 1.6412, + "theoretical_loss": 4.12945325460463, + "tokens_seen": 329777152 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045457372116349047, + "loss": 1.7004, + "theoretical_loss": 4.129359112237787, + "tokens_seen": 329842688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045456369107321965, + "loss": 1.665, + "theoretical_loss": 4.1292649938103265, + "tokens_seen": 329908224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545536609829489, + "loss": 1.6949, + "theoretical_loss": 4.129170899311406, + "tokens_seen": 329973760 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454543630892678, + "loss": 1.8215, + "theoretical_loss": 4.129076828730192, + "tokens_seen": 330039296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045453360080240725, + "loss": 1.7156, + "theoretical_loss": 4.128982782055858, + "tokens_seen": 330104832 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045452357071213643, + "loss": 1.6469, + "theoretical_loss": 4.128888759277583, + "tokens_seen": 330170368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545135406218656, + "loss": 1.6958, + "theoretical_loss": 4.1287947603845545, + "tokens_seen": 330235904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004545035105315948, + "loss": 1.7618, + "theoretical_loss": 4.128700785365965, + "tokens_seen": 330301440 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045449348044132397, + "loss": 1.7636, + "theoretical_loss": 4.128606834211018, + "tokens_seen": 330366976 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045448345035105315, + "loss": 1.602, + "theoretical_loss": 4.128512906908919, + "tokens_seen": 330432512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544734202607824, + "loss": 1.6435, + "theoretical_loss": 4.128419003448886, + "tokens_seen": 330498048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544633901705115, + "loss": 1.7587, + "theoretical_loss": 4.128325123820138, + "tokens_seen": 330563584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045445336008024075, + "loss": 1.7618, + "theoretical_loss": 4.1282312680119055, + "tokens_seen": 330629120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544433299899699, + "loss": 1.7534, + "theoretical_loss": 4.128137436013425, + "tokens_seen": 330694656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544332998996991, + "loss": 1.6879, + "theoretical_loss": 4.128043627813939, + "tokens_seen": 330760192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544232698094283, + "loss": 1.6608, + "theoretical_loss": 4.127949843402699, + "tokens_seen": 330825728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004544132397191575, + "loss": 1.8183, + "theoretical_loss": 4.12785608276896, + "tokens_seen": 330891264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 127130, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7497310638427734, + "objective/train/theoretical_loss": 4.127762345901987, + "objective/train/tokens_used": 351416800, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045440320962888666, + "loss": 1.7772, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045439317953861584, + "loss": 1.6473, + "theoretical_loss": 4.1276686327910515, + "tokens_seen": 331022336 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454383149448345, + "loss": 1.5993, + "theoretical_loss": 4.127574943425429, + "tokens_seen": 331087872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045437311935807425, + "loss": 1.8481, + "theoretical_loss": 4.127481277794408, + "tokens_seen": 331153408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543630892678034, + "loss": 1.7728, + "theoretical_loss": 4.1273876358872785, + "tokens_seen": 331218944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543530591775326, + "loss": 1.6421, + "theoretical_loss": 4.127294017693339, + "tokens_seen": 331284480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543430290872618, + "loss": 1.7672, + "theoretical_loss": 4.127200423201897, + "tokens_seen": 331350016 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454332998996991, + "loss": 1.6732, + "theoretical_loss": 4.127106852402264, + "tokens_seen": 331415552 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045432296890672016, + "loss": 1.6589, + "theoretical_loss": 4.127013305283759, + "tokens_seen": 331481088 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045431293881644934, + "loss": 1.6366, + "theoretical_loss": 4.12691978183571, + "tokens_seen": 331546624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004543029087261785, + "loss": 1.5818, + "theoretical_loss": 4.126826282047449, + "tokens_seen": 331612160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045429287863590776, + "loss": 1.696, + "theoretical_loss": 4.1267328059083175, + "tokens_seen": 331677696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542828485456369, + "loss": 1.6801, + "theoretical_loss": 4.126639353407663, + "tokens_seen": 331743232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542728184553661, + "loss": 1.6839, + "theoretical_loss": 4.12654592453484, + "tokens_seen": 331808768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045426278836509525, + "loss": 1.7261, + "theoretical_loss": 4.126452519279208, + "tokens_seen": 331874304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542527582748245, + "loss": 1.6749, + "theoretical_loss": 4.126359137630136, + "tokens_seen": 331939840 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045424272818455366, + "loss": 1.7973, + "theoretical_loss": 4.126265779576999, + "tokens_seen": 332005376 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045423269809428284, + "loss": 1.7139, + "theoretical_loss": 4.126172445109178, + "tokens_seen": 332070912 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454222668004012, + "loss": 1.7772, + "theoretical_loss": 4.126079134216063, + "tokens_seen": 332136448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004542126379137412, + "loss": 1.6267, + "theoretical_loss": 4.125985846887049, + "tokens_seen": 332201984 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045420260782347044, + "loss": 1.5861, + "theoretical_loss": 4.125892583111538, + "tokens_seen": 332267520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541925777331996, + "loss": 1.5698, + "theoretical_loss": 4.125799342878939, + "tokens_seen": 332333056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541825476429288, + "loss": 1.6362, + "theoretical_loss": 4.1257061261786685, + "tokens_seen": 332398592 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454172517552658, + "loss": 1.6018, + "theoretical_loss": 4.12561293300015, + "tokens_seen": 332464128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541624874623872, + "loss": 1.6136, + "theoretical_loss": 4.125519763332813, + "tokens_seen": 332529664 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045415245737211635, + "loss": 1.7913, + "theoretical_loss": 4.1254266171660925, + "tokens_seen": 332595200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541424272818456, + "loss": 1.7378, + "theoretical_loss": 4.125333494489434, + "tokens_seen": 332660736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541323971915747, + "loss": 1.7596, + "theoretical_loss": 4.125240395292286, + "tokens_seen": 332726272 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045412236710130395, + "loss": 1.7508, + "theoretical_loss": 4.125147319564107, + "tokens_seen": 332791808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541123370110331, + "loss": 1.7784, + "theoretical_loss": 4.12505426729436, + "tokens_seen": 332857344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004541023069207623, + "loss": 1.5771, + "theoretical_loss": 4.124961238472517, + "tokens_seen": 332922880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540922768304915, + "loss": 1.6675, + "theoretical_loss": 4.124868233088054, + "tokens_seen": 332988416 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045408224674022067, + "loss": 1.6863, + "theoretical_loss": 4.124775251130455, + "tokens_seen": 333053952 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045407221664994985, + "loss": 1.6826, + "theoretical_loss": 4.124682292589212, + "tokens_seen": 333119488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540621865596791, + "loss": 1.5558, + "theoretical_loss": 4.124589357453823, + "tokens_seen": 333185024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540521564694082, + "loss": 1.7201, + "theoretical_loss": 4.124496445713792, + "tokens_seen": 333250560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045404212637913745, + "loss": 1.6368, + "theoretical_loss": 4.12440355735863, + "tokens_seen": 333316096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045403209628886663, + "loss": 1.7515, + "theoretical_loss": 4.124310692377857, + "tokens_seen": 333381632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004540220661985958, + "loss": 1.7493, + "theoretical_loss": 4.124217850760996, + "tokens_seen": 333447168 + }, + { + "epoch": 0.1, + "learning_rate": 0.000454012036108325, + "loss": 1.6467, + "theoretical_loss": 4.124125032497578, + "tokens_seen": 333512704 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045400200601805417, + "loss": 1.6071, + "theoretical_loss": 4.124032237577143, + "tokens_seen": 333578240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045399197592778335, + "loss": 1.5819, + "theoretical_loss": 4.1239394659892366, + "tokens_seen": 333643776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539819458375126, + "loss": 1.7367, + "theoretical_loss": 4.12384671772341, + "tokens_seen": 333709312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539719157472417, + "loss": 1.6724, + "theoretical_loss": 4.12375399276922, + "tokens_seen": 333774848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045396188565697095, + "loss": 1.7496, + "theoretical_loss": 4.1236612911162345, + "tokens_seen": 333840384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539518555667001, + "loss": 1.6332, + "theoretical_loss": 4.123568612754024, + "tokens_seen": 333905920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539418254764293, + "loss": 1.5932, + "theoretical_loss": 4.123475957672168, + "tokens_seen": 333971456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539317953861585, + "loss": 1.7926, + "theoretical_loss": 4.123383325860252, + "tokens_seen": 334036992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004539217652958877, + "loss": 1.7697, + "theoretical_loss": 4.123290717307867, + "tokens_seen": 334102528 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045391173520561686, + "loss": 1.6972, + "theoretical_loss": 4.123198132004614, + "tokens_seen": 334168064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 128373, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8566737174987793, + "objective/train/theoretical_loss": 4.123105569940097, + "objective/train/tokens_used": 354693600, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045390170511534604, + "loss": 1.7508, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538916750250752, + "loss": 1.5712, + "theoretical_loss": 4.123013031103927, + "tokens_seen": 334299136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045388164493480445, + "loss": 1.7375, + "theoretical_loss": 4.122920515485726, + "tokens_seen": 334364672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538716148445336, + "loss": 1.6134, + "theoretical_loss": 4.122828023075117, + "tokens_seen": 334430208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538615847542628, + "loss": 1.61, + "theoretical_loss": 4.122735553861733, + "tokens_seen": 334495744 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453851554663992, + "loss": 1.7724, + "theoretical_loss": 4.122643107835213, + "tokens_seen": 334561280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538415245737212, + "loss": 1.6075, + "theoretical_loss": 4.122550684985203, + "tokens_seen": 334626816 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045383149448345036, + "loss": 1.7488, + "theoretical_loss": 4.122458285301355, + "tokens_seen": 334692352 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045382146439317954, + "loss": 1.7984, + "theoretical_loss": 4.122365908773327, + "tokens_seen": 334757888 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004538114343029087, + "loss": 1.6907, + "theoretical_loss": 4.1222735553907865, + "tokens_seen": 334823424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045380140421263796, + "loss": 1.6472, + "theoretical_loss": 4.122181225143404, + "tokens_seen": 334888960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537913741223671, + "loss": 1.8189, + "theoretical_loss": 4.122088918020857, + "tokens_seen": 334954496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537813440320963, + "loss": 1.7468, + "theoretical_loss": 4.121996634012834, + "tokens_seen": 335020032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045377131394182545, + "loss": 1.631, + "theoretical_loss": 4.121904373109024, + "tokens_seen": 335085568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537612838515547, + "loss": 1.8132, + "theoretical_loss": 4.121812135299129, + "tokens_seen": 335151104 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045375125376128386, + "loss": 1.6505, + "theoretical_loss": 4.1217199205728505, + "tokens_seen": 335216640 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045374122367101304, + "loss": 1.7154, + "theoretical_loss": 4.121627728919903, + "tokens_seen": 335282176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537311935807422, + "loss": 1.7704, + "theoretical_loss": 4.121535560330004, + "tokens_seen": 335347712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537211634904714, + "loss": 1.608, + "theoretical_loss": 4.1214434147928785, + "tokens_seen": 335413248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537111334002006, + "loss": 1.6938, + "theoretical_loss": 4.121351292298257, + "tokens_seen": 335478784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004537011033099298, + "loss": 1.7281, + "theoretical_loss": 4.12125919283588, + "tokens_seen": 335544320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045369107321965895, + "loss": 1.7027, + "theoretical_loss": 4.12116711639549, + "tokens_seen": 335609856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536810431293882, + "loss": 1.7405, + "theoretical_loss": 4.12107506296684, + "tokens_seen": 335675392 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045367101303911737, + "loss": 1.5998, + "theoretical_loss": 4.120983032539687, + "tokens_seen": 335740928 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045366098294884655, + "loss": 1.7249, + "theoretical_loss": 4.120891025103796, + "tokens_seen": 335806464 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045365095285857573, + "loss": 1.6693, + "theoretical_loss": 4.1207990406489365, + "tokens_seen": 335872000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536409227683049, + "loss": 1.7536, + "theoretical_loss": 4.120707079164887, + "tokens_seen": 335937536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536308926780341, + "loss": 1.6888, + "theoretical_loss": 4.120615140641433, + "tokens_seen": 336003072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536208625877633, + "loss": 1.6464, + "theoretical_loss": 4.1205232250683625, + "tokens_seen": 336068608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045361083249749245, + "loss": 1.6793, + "theoretical_loss": 4.120431332435475, + "tokens_seen": 336134144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004536008024072217, + "loss": 1.6343, + "theoretical_loss": 4.120339462732572, + "tokens_seen": 336199680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535907723169508, + "loss": 1.7151, + "theoretical_loss": 4.120247615949465, + "tokens_seen": 336265216 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045358074222668005, + "loss": 1.7207, + "theoretical_loss": 4.12015579207597, + "tokens_seen": 336330752 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045357071213640923, + "loss": 1.7207, + "theoretical_loss": 4.12006399110191, + "tokens_seen": 336396288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535606820461384, + "loss": 1.7815, + "theoretical_loss": 4.119972213017116, + "tokens_seen": 336461824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535506519558676, + "loss": 1.6364, + "theoretical_loss": 4.119880457811424, + "tokens_seen": 336527360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045354062186559683, + "loss": 1.6407, + "theoretical_loss": 4.119788725474675, + "tokens_seen": 336592896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045353059177532596, + "loss": 1.7715, + "theoretical_loss": 4.119697015996719, + "tokens_seen": 336658432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535205616850552, + "loss": 1.6939, + "theoretical_loss": 4.119605329367413, + "tokens_seen": 336723968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004535105315947843, + "loss": 1.6814, + "theoretical_loss": 4.119513665576617, + "tokens_seen": 336789504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045350050150451355, + "loss": 1.7228, + "theoretical_loss": 4.119422024614201, + "tokens_seen": 336855040 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045349047141424274, + "loss": 1.6603, + "theoretical_loss": 4.11933040647004, + "tokens_seen": 336920576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534804413239719, + "loss": 1.6734, + "theoretical_loss": 4.119238811134014, + "tokens_seen": 336986112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534704112337011, + "loss": 1.6424, + "theoretical_loss": 4.119147238596014, + "tokens_seen": 337051648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534603811434303, + "loss": 1.6177, + "theoretical_loss": 4.119055688845931, + "tokens_seen": 337117184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534503510531595, + "loss": 1.7783, + "theoretical_loss": 4.118964161873667, + "tokens_seen": 337182720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534403209628887, + "loss": 1.6775, + "theoretical_loss": 4.118872657669131, + "tokens_seen": 337248256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534302908726179, + "loss": 1.7548, + "theoretical_loss": 4.118781176222235, + "tokens_seen": 337313792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045342026078234706, + "loss": 1.5833, + "theoretical_loss": 4.1186897175229, + "tokens_seen": 337379328 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045341023069207624, + "loss": 1.7754, + "theoretical_loss": 4.1185982815610505, + "tokens_seen": 337444864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 129675, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6130269765853882, + "objective/train/theoretical_loss": 4.118506868326623, + "objective/train/tokens_used": 357970400, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004534002006018054, + "loss": 1.585, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045339017051153465, + "loss": 1.7112, + "theoretical_loss": 4.118415477809553, + "tokens_seen": 337575936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533801404212638, + "loss": 1.6035, + "theoretical_loss": 4.11832410999979, + "tokens_seen": 337641472 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453370110330993, + "loss": 1.742, + "theoretical_loss": 4.118232764887285, + "tokens_seen": 337707008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533600802407222, + "loss": 1.582, + "theoretical_loss": 4.118141442461995, + "tokens_seen": 337772544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533500501504514, + "loss": 1.6182, + "theoretical_loss": 4.118050142713886, + "tokens_seen": 337838080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045334002006018056, + "loss": 1.6775, + "theoretical_loss": 4.1179588656329305, + "tokens_seen": 337903616 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045332998996990974, + "loss": 1.7568, + "theoretical_loss": 4.117867611209104, + "tokens_seen": 337969152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004533199598796389, + "loss": 1.7228, + "theoretical_loss": 4.117776379432392, + "tokens_seen": 338034688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045330992978936816, + "loss": 1.5967, + "theoretical_loss": 4.117685170292786, + "tokens_seen": 338100224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532998996990973, + "loss": 1.7382, + "theoretical_loss": 4.1175939837802815, + "tokens_seen": 338165760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532898696088265, + "loss": 1.6048, + "theoretical_loss": 4.117502819884882, + "tokens_seen": 338231296 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045327983951855565, + "loss": 1.7153, + "theoretical_loss": 4.117411678596596, + "tokens_seen": 338296832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532698094282849, + "loss": 1.8324, + "theoretical_loss": 4.117320559905441, + "tokens_seen": 338362368 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045325977933801406, + "loss": 1.7808, + "theoretical_loss": 4.117229463801439, + "tokens_seen": 338427904 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045324974924774324, + "loss": 1.8107, + "theoretical_loss": 4.117138390274619, + "tokens_seen": 338493440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532397191574724, + "loss": 1.6799, + "theoretical_loss": 4.117047339315014, + "tokens_seen": 338558976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532296890672016, + "loss": 1.821, + "theoretical_loss": 4.116956310912666, + "tokens_seen": 338624512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004532196589769308, + "loss": 1.864, + "theoretical_loss": 4.116865305057625, + "tokens_seen": 338690048 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045320962888666, + "loss": 1.7792, + "theoretical_loss": 4.116774321739942, + "tokens_seen": 338755584 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045319959879638915, + "loss": 1.6384, + "theoretical_loss": 4.116683360949678, + "tokens_seen": 338821120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531895687061184, + "loss": 1.6449, + "theoretical_loss": 4.1165924226769, + "tokens_seen": 338886656 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045317953861584757, + "loss": 1.7527, + "theoretical_loss": 4.11650150691168, + "tokens_seen": 338952192 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045316950852557675, + "loss": 1.71, + "theoretical_loss": 4.116410613644098, + "tokens_seen": 339017728 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045315947843530593, + "loss": 1.6638, + "theoretical_loss": 4.116319742864239, + "tokens_seen": 339083264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531494483450351, + "loss": 1.7007, + "theoretical_loss": 4.1162288945621945, + "tokens_seen": 339148800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531394182547643, + "loss": 1.7578, + "theoretical_loss": 4.116138068728063, + "tokens_seen": 339214336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531293881644935, + "loss": 1.611, + "theoretical_loss": 4.1160472653519475, + "tokens_seen": 339279872 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045311935807422265, + "loss": 1.5057, + "theoretical_loss": 4.11595648442396, + "tokens_seen": 339345408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004531093279839519, + "loss": 1.5733, + "theoretical_loss": 4.115865725934217, + "tokens_seen": 339410944 + }, + { + "epoch": 0.1, + "learning_rate": 0.000453099297893681, + "loss": 1.7109, + "theoretical_loss": 4.11577498987284, + "tokens_seen": 339476480 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045308926780341025, + "loss": 1.7305, + "theoretical_loss": 4.11568427622996, + "tokens_seen": 339542016 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045307923771313943, + "loss": 1.6255, + "theoretical_loss": 4.115593584995712, + "tokens_seen": 339607552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530692076228686, + "loss": 1.612, + "theoretical_loss": 4.115502916160237, + "tokens_seen": 339673088 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530591775325978, + "loss": 1.6622, + "theoretical_loss": 4.115412269713684, + "tokens_seen": 339738624 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045304914744232703, + "loss": 1.8309, + "theoretical_loss": 4.115321645646207, + "tokens_seen": 339804160 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045303911735205616, + "loss": 1.6304, + "theoretical_loss": 4.115231043947968, + "tokens_seen": 339869696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530290872617854, + "loss": 1.6154, + "theoretical_loss": 4.11514046460913, + "tokens_seen": 339935232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004530190571715145, + "loss": 1.6693, + "theoretical_loss": 4.115049907619868, + "tokens_seen": 340000768 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045300902708124375, + "loss": 1.7455, + "theoretical_loss": 4.114959372970362, + "tokens_seen": 340066304 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045299899699097294, + "loss": 1.6322, + "theoretical_loss": 4.114868860650796, + "tokens_seen": 340131840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529889669007021, + "loss": 1.8013, + "theoretical_loss": 4.114778370651361, + "tokens_seen": 340197376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529789368104313, + "loss": 1.6755, + "theoretical_loss": 4.114687902962258, + "tokens_seen": 340262912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529689067201605, + "loss": 1.6852, + "theoretical_loss": 4.114597457573687, + "tokens_seen": 340328448 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045295887662988966, + "loss": 1.7232, + "theoretical_loss": 4.11450703447586, + "tokens_seen": 340393984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529488465396189, + "loss": 1.6696, + "theoretical_loss": 4.114416633658992, + "tokens_seen": 340459520 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452938816449348, + "loss": 1.7264, + "theoretical_loss": 4.114326255113308, + "tokens_seen": 340525056 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045292878635907726, + "loss": 1.5486, + "theoretical_loss": 4.114235898829034, + "tokens_seen": 340590592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529187562688064, + "loss": 1.749, + "theoretical_loss": 4.114145564796406, + "tokens_seen": 340656128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004529087261785356, + "loss": 1.7311, + "theoretical_loss": 4.114055253005665, + "tokens_seen": 340721664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 130755, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.740278720855713, + "objective/train/theoretical_loss": 4.113964963447058, + "objective/train/tokens_used": 361247200, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528986960882648, + "loss": 1.6466, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452888665997994, + "loss": 1.6383, + "theoretical_loss": 4.113874696110838, + "tokens_seen": 340852736 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045287863590772316, + "loss": 1.7663, + "theoretical_loss": 4.113784450987264, + "tokens_seen": 340918272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528686058174524, + "loss": 1.5324, + "theoretical_loss": 4.113694228066604, + "tokens_seen": 340983808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528585757271815, + "loss": 1.7603, + "theoretical_loss": 4.113604027339126, + "tokens_seen": 341049344 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045284854563691076, + "loss": 1.6387, + "theoretical_loss": 4.11351384879511, + "tokens_seen": 341114880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528385155466399, + "loss": 1.5884, + "theoretical_loss": 4.11342369242484, + "tokens_seen": 341180416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528284854563691, + "loss": 1.797, + "theoretical_loss": 4.1133335582186055, + "tokens_seen": 341245952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528184553660983, + "loss": 1.6206, + "theoretical_loss": 4.1132434461667025, + "tokens_seen": 341311488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004528084252758275, + "loss": 1.8154, + "theoretical_loss": 4.113153356259433, + "tokens_seen": 341377024 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045279839518555667, + "loss": 1.7358, + "theoretical_loss": 4.1130632884871074, + "tokens_seen": 341442560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045278836509528585, + "loss": 1.5965, + "theoretical_loss": 4.112973242840038, + "tokens_seen": 341508096 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045277833500501503, + "loss": 1.6813, + "theoretical_loss": 4.1128832193085465, + "tokens_seen": 341573632 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045276830491474426, + "loss": 1.6939, + "theoretical_loss": 4.11279321788296, + "tokens_seen": 341639168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004527582748244734, + "loss": 1.6409, + "theoretical_loss": 4.112703238553609, + "tokens_seen": 341704704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004527482447342026, + "loss": 1.6945, + "theoretical_loss": 4.112613281310835, + "tokens_seen": 341770240 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045273821464393175, + "loss": 1.577, + "theoretical_loss": 4.1125233461449815, + "tokens_seen": 341835776 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452728184553661, + "loss": 1.7081, + "theoretical_loss": 4.1124334330464, + "tokens_seen": 341901312 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045271815446339017, + "loss": 1.7201, + "theoretical_loss": 4.112343542005447, + "tokens_seen": 341966848 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045270812437311935, + "loss": 1.4961, + "theoretical_loss": 4.112253673012486, + "tokens_seen": 342032384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526980942828486, + "loss": 1.6681, + "theoretical_loss": 4.112163826057886, + "tokens_seen": 342097920 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045268806419257777, + "loss": 1.6334, + "theoretical_loss": 4.112074001132023, + "tokens_seen": 342163456 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045267803410230695, + "loss": 1.7123, + "theoretical_loss": 4.111984198225278, + "tokens_seen": 342228992 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045266800401203613, + "loss": 1.725, + "theoretical_loss": 4.1118944173280365, + "tokens_seen": 342294528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526579739217653, + "loss": 1.6704, + "theoretical_loss": 4.111804658430694, + "tokens_seen": 342360064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526479438314945, + "loss": 1.5385, + "theoretical_loss": 4.11171492152365, + "tokens_seen": 342425600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526379137412237, + "loss": 1.6648, + "theoretical_loss": 4.111625206597309, + "tokens_seen": 342491136 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045262788365095285, + "loss": 1.751, + "theoretical_loss": 4.111535513642082, + "tokens_seen": 342556672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526178535606821, + "loss": 1.6102, + "theoretical_loss": 4.111445842648387, + "tokens_seen": 342622208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004526078234704112, + "loss": 1.6703, + "theoretical_loss": 4.111356193606648, + "tokens_seen": 342687744 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045259779338014045, + "loss": 1.6735, + "theoretical_loss": 4.111266566507294, + "tokens_seen": 342753280 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045258776328986963, + "loss": 1.6127, + "theoretical_loss": 4.11117696134076, + "tokens_seen": 342818816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525777331995988, + "loss": 1.5712, + "theoretical_loss": 4.111087378097488, + "tokens_seen": 342884352 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452567703109328, + "loss": 1.7529, + "theoretical_loss": 4.1109978167679255, + "tokens_seen": 342949888 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045255767301905723, + "loss": 1.6433, + "theoretical_loss": 4.110908277342526, + "tokens_seen": 343015424 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045254764292878636, + "loss": 1.7044, + "theoretical_loss": 4.110818759811748, + "tokens_seen": 343080960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525376128385156, + "loss": 1.6067, + "theoretical_loss": 4.110729264166058, + "tokens_seen": 343146496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004525275827482447, + "loss": 1.7963, + "theoretical_loss": 4.110639790395926, + "tokens_seen": 343212032 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045251755265797395, + "loss": 1.6116, + "theoretical_loss": 4.110550338491832, + "tokens_seen": 343277568 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045250752256770314, + "loss": 1.8024, + "theoretical_loss": 4.1104609084442565, + "tokens_seen": 343343104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524974924774323, + "loss": 1.6516, + "theoretical_loss": 4.11037150024369, + "tokens_seen": 343408640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524874623871615, + "loss": 1.6873, + "theoretical_loss": 4.110282113880627, + "tokens_seen": 343474176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524774322968907, + "loss": 1.644, + "theoretical_loss": 4.11019274934557, + "tokens_seen": 343539712 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045246740220661986, + "loss": 1.6906, + "theoretical_loss": 4.110103406629026, + "tokens_seen": 343605248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524573721163491, + "loss": 1.5997, + "theoretical_loss": 4.110014085721506, + "tokens_seen": 343670784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524473420260782, + "loss": 1.7353, + "theoretical_loss": 4.109924786613531, + "tokens_seen": 343736320 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045243731193580746, + "loss": 1.6546, + "theoretical_loss": 4.109835509295625, + "tokens_seen": 343801856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524272818455366, + "loss": 1.6248, + "theoretical_loss": 4.10974625375832, + "tokens_seen": 343867392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004524172517552658, + "loss": 1.6875, + "theoretical_loss": 4.1096570199921505, + "tokens_seen": 343932928 + }, + { + "epoch": 0.1, + "learning_rate": 0.000452407221664995, + "loss": 1.6553, + "theoretical_loss": 4.109567807987661, + "tokens_seen": 343998464 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 131936, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7221022844314575, + "objective/train/theoretical_loss": 4.109478617735399, + "objective/train/tokens_used": 364524000, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523971915747242, + "loss": 1.8338, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045238716148445336, + "loss": 1.5997, + "theoretical_loss": 4.10938944922592, + "tokens_seen": 344129536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523771313941826, + "loss": 1.6757, + "theoretical_loss": 4.109300302449784, + "tokens_seen": 344195072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523671013039117, + "loss": 1.5631, + "theoretical_loss": 4.109211177397556, + "tokens_seen": 344260608 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045235707121364096, + "loss": 1.5899, + "theoretical_loss": 4.109122074059811, + "tokens_seen": 344326144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523470411233701, + "loss": 1.5993, + "theoretical_loss": 4.1090329924271245, + "tokens_seen": 344391680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523370110330993, + "loss": 1.5991, + "theoretical_loss": 4.108943932490082, + "tokens_seen": 344457216 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523269809428285, + "loss": 1.7351, + "theoretical_loss": 4.108854894239271, + "tokens_seen": 344522752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004523169508525577, + "loss": 1.6427, + "theoretical_loss": 4.10876587766529, + "tokens_seen": 344588288 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045230692076228687, + "loss": 1.6315, + "theoretical_loss": 4.108676882758738, + "tokens_seen": 344653824 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045229689067201605, + "loss": 1.6746, + "theoretical_loss": 4.1085879095102245, + "tokens_seen": 344719360 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045228686058174523, + "loss": 1.7928, + "theoretical_loss": 4.108498957910362, + "tokens_seen": 344784896 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045227683049147446, + "loss": 1.7028, + "theoretical_loss": 4.108410027949769, + "tokens_seen": 344850432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522668004012036, + "loss": 1.6821, + "theoretical_loss": 4.10832111961907, + "tokens_seen": 344915968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522567703109328, + "loss": 1.713, + "theoretical_loss": 4.1082322329088985, + "tokens_seen": 344981504 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045224674022066195, + "loss": 1.8237, + "theoretical_loss": 4.108143367809888, + "tokens_seen": 345047040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004522367101303912, + "loss": 1.6244, + "theoretical_loss": 4.108054524312683, + "tokens_seen": 345112576 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045222668004012037, + "loss": 1.6231, + "theoretical_loss": 4.107965702407931, + "tokens_seen": 345178112 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045221664994984955, + "loss": 1.6322, + "theoretical_loss": 4.107876902086286, + "tokens_seen": 345243648 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045220661985957873, + "loss": 1.6831, + "theoretical_loss": 4.107788123338408, + "tokens_seen": 345309184 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045219658976930797, + "loss": 1.7115, + "theoretical_loss": 4.1076993661549634, + "tokens_seen": 345374720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521865596790371, + "loss": 1.7352, + "theoretical_loss": 4.107610630526622, + "tokens_seen": 345440256 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045217652958876633, + "loss": 1.906, + "theoretical_loss": 4.107521916444064, + "tokens_seen": 345505792 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045216649949849546, + "loss": 1.6024, + "theoretical_loss": 4.10743322389797, + "tokens_seen": 345571328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521564694082247, + "loss": 1.5894, + "theoretical_loss": 4.107344552879029, + "tokens_seen": 345636864 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045214643931795387, + "loss": 1.7057, + "theoretical_loss": 4.107255903377938, + "tokens_seen": 345702400 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045213640922768305, + "loss": 1.7444, + "theoretical_loss": 4.107167275385395, + "tokens_seen": 345767936 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045212637913741223, + "loss": 1.7886, + "theoretical_loss": 4.107078668892108, + "tokens_seen": 345833472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521163490471414, + "loss": 1.6521, + "theoretical_loss": 4.106990083888789, + "tokens_seen": 345899008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004521063189568706, + "loss": 1.7272, + "theoretical_loss": 4.106901520366154, + "tokens_seen": 345964544 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045209628886659983, + "loss": 1.6611, + "theoretical_loss": 4.106812978314929, + "tokens_seen": 346030080 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045208625877632896, + "loss": 1.7207, + "theoretical_loss": 4.106724457725842, + "tokens_seen": 346095616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520762286860582, + "loss": 1.75, + "theoretical_loss": 4.106635958589628, + "tokens_seen": 346161152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520661985957873, + "loss": 1.6431, + "theoretical_loss": 4.106547480897029, + "tokens_seen": 346226688 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045205616850551656, + "loss": 1.5729, + "theoretical_loss": 4.106459024638791, + "tokens_seen": 346292224 + }, + { + "epoch": 0.1, + "learning_rate": 0.00045204613841524574, + "loss": 1.6131, + "theoretical_loss": 4.106370589805666, + "tokens_seen": 346357760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520361083249749, + "loss": 1.7663, + "theoretical_loss": 4.1062821763884125, + "tokens_seen": 346423296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0004520260782347041, + "loss": 1.8047, + "theoretical_loss": 4.106193784377794, + "tokens_seen": 346488832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045201604814443334, + "loss": 1.699, + "theoretical_loss": 4.106105413764581, + "tokens_seen": 346554368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045200601805416246, + "loss": 1.582, + "theoretical_loss": 4.106017064539548, + "tokens_seen": 346619904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519959879638917, + "loss": 1.8022, + "theoretical_loss": 4.1059287366934765, + "tokens_seen": 346685440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519859578736208, + "loss": 1.6313, + "theoretical_loss": 4.105840430217153, + "tokens_seen": 346750976 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045197592778335006, + "loss": 1.7099, + "theoretical_loss": 4.10575214510137, + "tokens_seen": 346816512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045196589769307924, + "loss": 1.614, + "theoretical_loss": 4.105663881336925, + "tokens_seen": 346882048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519558676028084, + "loss": 1.6481, + "theoretical_loss": 4.105575638914623, + "tokens_seen": 346947584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045194583751253766, + "loss": 1.7314, + "theoretical_loss": 4.105487417825273, + "tokens_seen": 347013120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519358074222668, + "loss": 1.6945, + "theoretical_loss": 4.10539921805969, + "tokens_seen": 347078656 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451925777331996, + "loss": 1.562, + "theoretical_loss": 4.105311039608695, + "tokens_seen": 347144192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519157472417252, + "loss": 1.5661, + "theoretical_loss": 4.105222882463115, + "tokens_seen": 347209728 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004519057171514544, + "loss": 1.673, + "theoretical_loss": 4.105134746613783, + "tokens_seen": 347275264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 133174, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6553549766540527, + "objective/train/theoretical_loss": 4.105046632051535, + "objective/train/tokens_used": 367800800, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045189568706118356, + "loss": 1.601, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518856569709128, + "loss": 1.644, + "theoretical_loss": 4.1049585387672165, + "tokens_seen": 347406336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518756268806419, + "loss": 1.6819, + "theoretical_loss": 4.1048704667516756, + "tokens_seen": 347471872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045186559679037116, + "loss": 1.6626, + "theoretical_loss": 4.104782415995768, + "tokens_seen": 347537408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518555667001003, + "loss": 1.7109, + "theoretical_loss": 4.104694386490354, + "tokens_seen": 347602944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518455366098295, + "loss": 1.6245, + "theoretical_loss": 4.1046063782263, + "tokens_seen": 347668480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518355065195587, + "loss": 1.7237, + "theoretical_loss": 4.104518391194477, + "tokens_seen": 347734016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004518254764292879, + "loss": 1.6072, + "theoretical_loss": 4.104430425385765, + "tokens_seen": 347799552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045181544633901707, + "loss": 1.6399, + "theoretical_loss": 4.104342480791044, + "tokens_seen": 347865088 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045180541624874625, + "loss": 1.6001, + "theoretical_loss": 4.104254557401205, + "tokens_seen": 347930624 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045179538615847543, + "loss": 1.6404, + "theoretical_loss": 4.104166655207142, + "tokens_seen": 347996160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045178535606820466, + "loss": 1.6943, + "theoretical_loss": 4.104078774199754, + "tokens_seen": 348061696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004517753259779338, + "loss": 1.7232, + "theoretical_loss": 4.10399091436995, + "tokens_seen": 348127232 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451765295887663, + "loss": 1.6666, + "theoretical_loss": 4.103903075708637, + "tokens_seen": 348192768 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045175526579739215, + "loss": 1.5773, + "theoretical_loss": 4.103815258206735, + "tokens_seen": 348258304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004517452357071214, + "loss": 1.6509, + "theoretical_loss": 4.103727461855165, + "tokens_seen": 348323840 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045173520561685057, + "loss": 1.5344, + "theoretical_loss": 4.103639686644856, + "tokens_seen": 348389376 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045172517552657975, + "loss": 1.7409, + "theoretical_loss": 4.103551932566741, + "tokens_seen": 348454912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045171514543630893, + "loss": 1.7462, + "theoretical_loss": 4.103464199611759, + "tokens_seen": 348520448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045170511534603817, + "loss": 1.6456, + "theoretical_loss": 4.1033764877708565, + "tokens_seen": 348585984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516950852557673, + "loss": 1.895, + "theoretical_loss": 4.103288797034983, + "tokens_seen": 348651520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045168505516549653, + "loss": 1.77, + "theoretical_loss": 4.103201127395095, + "tokens_seen": 348717056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045167502507522566, + "loss": 1.6687, + "theoretical_loss": 4.103113478842153, + "tokens_seen": 348782592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516649949849549, + "loss": 1.5756, + "theoretical_loss": 4.103025851367125, + "tokens_seen": 348848128 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045165496489468407, + "loss": 1.651, + "theoretical_loss": 4.102938244960985, + "tokens_seen": 348913664 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045164493480441325, + "loss": 1.5804, + "theoretical_loss": 4.102850659614709, + "tokens_seen": 348979200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045163490471414244, + "loss": 1.6462, + "theoretical_loss": 4.102763095319282, + "tokens_seen": 349044736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516248746238716, + "loss": 1.7101, + "theoretical_loss": 4.102675552065694, + "tokens_seen": 349110272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004516148445336008, + "loss": 1.6752, + "theoretical_loss": 4.10258802984494, + "tokens_seen": 349175808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045160481444333003, + "loss": 1.5943, + "theoretical_loss": 4.102500528648019, + "tokens_seen": 349241344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045159478435305916, + "loss": 1.7134, + "theoretical_loss": 4.102413048465938, + "tokens_seen": 349306880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515847542627884, + "loss": 1.6029, + "theoretical_loss": 4.10232558928971, + "tokens_seen": 349372416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515747241725175, + "loss": 1.5948, + "theoretical_loss": 4.10223815111035, + "tokens_seen": 349437952 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045156469408224676, + "loss": 1.66, + "theoretical_loss": 4.102150733918882, + "tokens_seen": 349503488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045155466399197594, + "loss": 1.8228, + "theoretical_loss": 4.102063337706333, + "tokens_seen": 349569024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515446339017051, + "loss": 1.6093, + "theoretical_loss": 4.101975962463738, + "tokens_seen": 349634560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515346038114343, + "loss": 1.6708, + "theoretical_loss": 4.101888608182136, + "tokens_seen": 349700096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045152457372116354, + "loss": 1.7003, + "theoretical_loss": 4.101801274852571, + "tokens_seen": 349765632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045151454363089266, + "loss": 1.6574, + "theoretical_loss": 4.1017139624660945, + "tokens_seen": 349831168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004515045135406219, + "loss": 1.4988, + "theoretical_loss": 4.1016266710137606, + "tokens_seen": 349896704 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451494483450351, + "loss": 1.6377, + "theoretical_loss": 4.101539400486631, + "tokens_seen": 349962240 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045148445336008026, + "loss": 1.5685, + "theoretical_loss": 4.101452150875773, + "tokens_seen": 350027776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045147442326980944, + "loss": 1.7385, + "theoretical_loss": 4.101364922172259, + "tokens_seen": 350093312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514643931795386, + "loss": 1.6382, + "theoretical_loss": 4.101277714367166, + "tokens_seen": 350158848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514543630892678, + "loss": 1.5895, + "theoretical_loss": 4.1011905274515765, + "tokens_seen": 350224384 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451444332998997, + "loss": 1.7266, + "theoretical_loss": 4.101103361416581, + "tokens_seen": 350289920 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045143430290872617, + "loss": 1.5495, + "theoretical_loss": 4.101016216253273, + "tokens_seen": 350355456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004514242728184554, + "loss": 1.6692, + "theoretical_loss": 4.10092909195275, + "tokens_seen": 350420992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045141424272818453, + "loss": 1.6779, + "theoretical_loss": 4.10084198850612, + "tokens_seen": 350486528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045140421263791376, + "loss": 1.6896, + "theoretical_loss": 4.100754905904491, + "tokens_seen": 350552064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 134382, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7152607440948486, + "objective/train/theoretical_loss": 4.100667844138981, + "objective/train/tokens_used": 371077600, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045139418254764294, + "loss": 1.7256, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513841524573721, + "loss": 1.6444, + "theoretical_loss": 4.10058080320071, + "tokens_seen": 350683136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513741223671013, + "loss": 1.7162, + "theoretical_loss": 4.100493783080806, + "tokens_seen": 350748672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513640922768305, + "loss": 1.8119, + "theoretical_loss": 4.1004067837704, + "tokens_seen": 350814208 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045135406218655967, + "loss": 1.6851, + "theoretical_loss": 4.1003198052606304, + "tokens_seen": 350879744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513440320962889, + "loss": 1.6686, + "theoretical_loss": 4.10023284754264, + "tokens_seen": 350945280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045133400200601803, + "loss": 1.6042, + "theoretical_loss": 4.100145910607576, + "tokens_seen": 351010816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045132397191574727, + "loss": 1.6159, + "theoretical_loss": 4.100058994446596, + "tokens_seen": 351076352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004513139418254764, + "loss": 1.7213, + "theoretical_loss": 4.099972099050856, + "tokens_seen": 351141888 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045130391173520563, + "loss": 1.7689, + "theoretical_loss": 4.099885224411522, + "tokens_seen": 351207424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512938816449348, + "loss": 1.5754, + "theoretical_loss": 4.099798370519764, + "tokens_seen": 351272960 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451283851554664, + "loss": 1.615, + "theoretical_loss": 4.099711537366758, + "tokens_seen": 351338496 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045127382146439317, + "loss": 1.6889, + "theoretical_loss": 4.099624724943685, + "tokens_seen": 351404032 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045126379137412235, + "loss": 1.6289, + "theoretical_loss": 4.099537933241731, + "tokens_seen": 351469568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045125376128385153, + "loss": 1.685, + "theoretical_loss": 4.0994511622520875, + "tokens_seen": 351535104 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045124373119358077, + "loss": 1.5913, + "theoretical_loss": 4.099364411965952, + "tokens_seen": 351600640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512337011033099, + "loss": 1.7937, + "theoretical_loss": 4.0992776823745265, + "tokens_seen": 351666176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045122367101303913, + "loss": 1.6254, + "theoretical_loss": 4.09919097346902, + "tokens_seen": 351731712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512136409227683, + "loss": 1.656, + "theoretical_loss": 4.099104285240646, + "tokens_seen": 351797248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004512036108324975, + "loss": 1.7311, + "theoretical_loss": 4.099017617680621, + "tokens_seen": 351862784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045119358074222673, + "loss": 1.703, + "theoretical_loss": 4.098930970780172, + "tokens_seen": 351928320 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045118355065195586, + "loss": 1.6626, + "theoretical_loss": 4.098844344530526, + "tokens_seen": 351993856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511735205616851, + "loss": 1.7201, + "theoretical_loss": 4.098757738922918, + "tokens_seen": 352059392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511634904714143, + "loss": 1.6801, + "theoretical_loss": 4.09867115394859, + "tokens_seen": 352124928 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045115346038114345, + "loss": 1.6376, + "theoretical_loss": 4.098584589598786, + "tokens_seen": 352190464 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045114343029087264, + "loss": 1.5858, + "theoretical_loss": 4.098498045864756, + "tokens_seen": 352256000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004511334002006018, + "loss": 1.677, + "theoretical_loss": 4.0984115227377576, + "tokens_seen": 352321536 + }, + { + "epoch": 0.11, + "learning_rate": 0.000451123370110331, + "loss": 1.6293, + "theoretical_loss": 4.098325020209051, + "tokens_seen": 352387072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045111334002006023, + "loss": 1.7528, + "theoretical_loss": 4.098238538269904, + "tokens_seen": 352452608 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045110330992978936, + "loss": 1.6195, + "theoretical_loss": 4.098152076911588, + "tokens_seen": 352518144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510932798395186, + "loss": 1.6711, + "theoretical_loss": 4.098065636125381, + "tokens_seen": 352583680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510832497492477, + "loss": 1.6556, + "theoretical_loss": 4.097979215902566, + "tokens_seen": 352649216 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045107321965897696, + "loss": 1.6927, + "theoretical_loss": 4.097892816234429, + "tokens_seen": 352714752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045106318956870614, + "loss": 1.5485, + "theoretical_loss": 4.097806437112267, + "tokens_seen": 352780288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510531594784353, + "loss": 1.5632, + "theoretical_loss": 4.0977200785273755, + "tokens_seen": 352845824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510431293881645, + "loss": 1.5491, + "theoretical_loss": 4.09763374047106, + "tokens_seen": 352911360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045103309929789374, + "loss": 1.7043, + "theoretical_loss": 4.097547422934629, + "tokens_seen": 352976896 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045102306920762286, + "loss": 1.6266, + "theoretical_loss": 4.097461125909398, + "tokens_seen": 353042432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510130391173521, + "loss": 1.7166, + "theoretical_loss": 4.097374849386685, + "tokens_seen": 353107968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004510030090270812, + "loss": 1.7341, + "theoretical_loss": 4.097288593357817, + "tokens_seen": 353173504 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045099297893681046, + "loss": 1.458, + "theoretical_loss": 4.097202357814124, + "tokens_seen": 353239040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045098294884653964, + "loss": 1.7514, + "theoretical_loss": 4.097116142746941, + "tokens_seen": 353304576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509729187562688, + "loss": 1.66, + "theoretical_loss": 4.09702994814761, + "tokens_seen": 353370112 + }, + { + "epoch": 0.11, + "learning_rate": 0.000450962888665998, + "loss": 1.7042, + "theoretical_loss": 4.096943774007476, + "tokens_seen": 353435648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509528585757272, + "loss": 1.7403, + "theoretical_loss": 4.096857620317892, + "tokens_seen": 353501184 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045094282848545637, + "loss": 1.6433, + "theoretical_loss": 4.096771487070214, + "tokens_seen": 353566720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004509327983951856, + "loss": 1.6389, + "theoretical_loss": 4.096685374255803, + "tokens_seen": 353632256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045092276830491473, + "loss": 1.6556, + "theoretical_loss": 4.096599281866028, + "tokens_seen": 353697792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045091273821464396, + "loss": 1.733, + "theoretical_loss": 4.09651320989226, + "tokens_seen": 353763328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045090270812437314, + "loss": 1.6227, + "theoretical_loss": 4.096427158325877, + "tokens_seen": 353828864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 135614, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.442173957824707, + "objective/train/theoretical_loss": 4.096341127158263, + "objective/train/tokens_used": 374354400, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508926780341023, + "loss": 1.6023, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508826479438315, + "loss": 1.6929, + "theoretical_loss": 4.096255116380806, + "tokens_seen": 353959936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508726178535607, + "loss": 1.6823, + "theoretical_loss": 4.096169125984899, + "tokens_seen": 354025472 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045086258776328987, + "loss": 1.6651, + "theoretical_loss": 4.0960831559619395, + "tokens_seen": 354091008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508525576730191, + "loss": 1.6833, + "theoretical_loss": 4.095997206303333, + "tokens_seen": 354156544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045084252758274823, + "loss": 1.6554, + "theoretical_loss": 4.095911277000488, + "tokens_seen": 354222080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045083249749247747, + "loss": 1.6462, + "theoretical_loss": 4.095825368044819, + "tokens_seen": 354287616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004508224674022066, + "loss": 1.7107, + "theoretical_loss": 4.0957394794277455, + "tokens_seen": 354353152 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045081243731193583, + "loss": 1.6273, + "theoretical_loss": 4.095653611140692, + "tokens_seen": 354418688 + }, + { + "epoch": 0.11, + "learning_rate": 0.000450802407221665, + "loss": 1.7058, + "theoretical_loss": 4.09556776317509, + "tokens_seen": 354484224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507923771313942, + "loss": 1.7045, + "theoretical_loss": 4.095481935522372, + "tokens_seen": 354549760 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045078234704112337, + "loss": 1.6472, + "theoretical_loss": 4.09539612817398, + "tokens_seen": 354615296 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045077231695085255, + "loss": 1.6791, + "theoretical_loss": 4.095310341121358, + "tokens_seen": 354680832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045076228686058173, + "loss": 1.717, + "theoretical_loss": 4.095224574355958, + "tokens_seen": 354746368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045075225677031097, + "loss": 1.7316, + "theoretical_loss": 4.095138827869237, + "tokens_seen": 354811904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507422266800401, + "loss": 1.6634, + "theoretical_loss": 4.0950531016526535, + "tokens_seen": 354877440 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045073219658976933, + "loss": 1.6386, + "theoretical_loss": 4.094967395697675, + "tokens_seen": 354942976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507221664994985, + "loss": 1.6482, + "theoretical_loss": 4.094881709995773, + "tokens_seen": 355008512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507121364092277, + "loss": 1.7429, + "theoretical_loss": 4.094796044538423, + "tokens_seen": 355074048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004507021063189569, + "loss": 1.6153, + "theoretical_loss": 4.094710399317109, + "tokens_seen": 355139584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045069207622868606, + "loss": 1.6764, + "theoretical_loss": 4.094624774323315, + "tokens_seen": 355205120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045068204613841524, + "loss": 1.6681, + "theoretical_loss": 4.094539169548535, + "tokens_seen": 355270656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506720160481445, + "loss": 1.6062, + "theoretical_loss": 4.094453584984265, + "tokens_seen": 355336192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506619859578736, + "loss": 1.6331, + "theoretical_loss": 4.094368020622008, + "tokens_seen": 355401728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045065195586760284, + "loss": 1.7501, + "theoretical_loss": 4.094282476453272, + "tokens_seen": 355467264 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045064192577733196, + "loss": 1.7012, + "theoretical_loss": 4.094196952469568, + "tokens_seen": 355532800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506318956870612, + "loss": 1.7209, + "theoretical_loss": 4.0941114486624155, + "tokens_seen": 355598336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004506218655967904, + "loss": 1.6688, + "theoretical_loss": 4.094025965023335, + "tokens_seen": 355663872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045061183550651956, + "loss": 1.5561, + "theoretical_loss": 4.093940501543857, + "tokens_seen": 355729408 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045060180541624874, + "loss": 1.5723, + "theoretical_loss": 4.093855058215513, + "tokens_seen": 355794944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505917753259779, + "loss": 1.7321, + "theoretical_loss": 4.093769635029841, + "tokens_seen": 355860480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505817452357071, + "loss": 1.6754, + "theoretical_loss": 4.0936842319783855, + "tokens_seen": 355926016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045057171514543634, + "loss": 1.5963, + "theoretical_loss": 4.093598849052694, + "tokens_seen": 355991552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045056168505516547, + "loss": 1.7297, + "theoretical_loss": 4.093513486244321, + "tokens_seen": 356057088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505516549648947, + "loss": 1.5834, + "theoretical_loss": 4.093428143544823, + "tokens_seen": 356122624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505416248746239, + "loss": 1.6973, + "theoretical_loss": 4.093342820945766, + "tokens_seen": 356188160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045053159478435306, + "loss": 1.5895, + "theoretical_loss": 4.093257518438718, + "tokens_seen": 356253696 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045052156469408224, + "loss": 1.6446, + "theoretical_loss": 4.093172236015253, + "tokens_seen": 356319232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505115346038114, + "loss": 1.6416, + "theoretical_loss": 4.093086973666949, + "tokens_seen": 356384768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004505015045135406, + "loss": 1.6473, + "theoretical_loss": 4.0930017313853915, + "tokens_seen": 356450304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045049147442326984, + "loss": 1.5908, + "theoretical_loss": 4.092916509162169, + "tokens_seen": 356515840 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045048144433299897, + "loss": 1.7073, + "theoretical_loss": 4.092831306988875, + "tokens_seen": 356581376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504714142427282, + "loss": 1.648, + "theoretical_loss": 4.0927461248571095, + "tokens_seen": 356646912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045046138415245733, + "loss": 1.6368, + "theoretical_loss": 4.092660962758477, + "tokens_seen": 356712448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045045135406218657, + "loss": 1.6458, + "theoretical_loss": 4.092575820684586, + "tokens_seen": 356777984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504413239719158, + "loss": 1.5487, + "theoretical_loss": 4.092490698627052, + "tokens_seen": 356843520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045043129388164493, + "loss": 1.6383, + "theoretical_loss": 4.0924055965774935, + "tokens_seen": 356909056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045042126379137416, + "loss": 1.7061, + "theoretical_loss": 4.092320514527535, + "tokens_seen": 356974592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045041123370110334, + "loss": 1.6218, + "theoretical_loss": 4.092235452468808, + "tokens_seen": 357040128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004504012036108325, + "loss": 1.6361, + "theoretical_loss": 4.092150410392945, + "tokens_seen": 357105664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 136331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7456684112548828, + "objective/train/theoretical_loss": 4.092065388291585, + "objective/train/tokens_used": 377631200, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503911735205617, + "loss": 1.7439, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503811434302909, + "loss": 1.7393, + "theoretical_loss": 4.091980386156376, + "tokens_seen": 357236736 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045037111334002007, + "loss": 1.7351, + "theoretical_loss": 4.091895403978965, + "tokens_seen": 357302272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503610832497493, + "loss": 1.6556, + "theoretical_loss": 4.091810441751007, + "tokens_seen": 357367808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045035105315947843, + "loss": 1.7615, + "theoretical_loss": 4.091725499464163, + "tokens_seen": 357433344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045034102306920767, + "loss": 1.7099, + "theoretical_loss": 4.091640577110096, + "tokens_seen": 357498880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503309929789368, + "loss": 1.6057, + "theoretical_loss": 4.091555674680476, + "tokens_seen": 357564416 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045032096288866603, + "loss": 1.5952, + "theoretical_loss": 4.09147079216698, + "tokens_seen": 357629952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503109327983952, + "loss": 1.4527, + "theoretical_loss": 4.0913859295612856, + "tokens_seen": 357695488 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004503009027081244, + "loss": 1.6429, + "theoretical_loss": 4.091301086855078, + "tokens_seen": 357761024 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045029087261785357, + "loss": 1.6535, + "theoretical_loss": 4.091216264040048, + "tokens_seen": 357826560 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045028084252758275, + "loss": 1.4885, + "theoretical_loss": 4.091131461107889, + "tokens_seen": 357892096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045027081243731193, + "loss": 1.5906, + "theoretical_loss": 4.091046678050301, + "tokens_seen": 357957632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045026078234704117, + "loss": 1.6244, + "theoretical_loss": 4.090961914858989, + "tokens_seen": 358023168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502507522567703, + "loss": 1.5192, + "theoretical_loss": 4.090877171525663, + "tokens_seen": 358088704 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045024072216649953, + "loss": 1.7578, + "theoretical_loss": 4.090792448042038, + "tokens_seen": 358154240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502306920762287, + "loss": 1.6154, + "theoretical_loss": 4.0907077443998325, + "tokens_seen": 358219776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502206619859579, + "loss": 1.7491, + "theoretical_loss": 4.090623060590772, + "tokens_seen": 358285312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004502106318956871, + "loss": 1.6818, + "theoretical_loss": 4.090538396606585, + "tokens_seen": 358350848 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045020060180541626, + "loss": 1.626, + "theoretical_loss": 4.090453752439008, + "tokens_seen": 358416384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045019057171514544, + "loss": 1.6677, + "theoretical_loss": 4.09036912807978, + "tokens_seen": 358481920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501805416248747, + "loss": 1.6347, + "theoretical_loss": 4.090284523520644, + "tokens_seen": 358547456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501705115346038, + "loss": 1.5167, + "theoretical_loss": 4.09019993875335, + "tokens_seen": 358612992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045016048144433304, + "loss": 1.672, + "theoretical_loss": 4.090115373769653, + "tokens_seen": 358678528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045015045135406216, + "loss": 1.5783, + "theoretical_loss": 4.0900308285613125, + "tokens_seen": 358744064 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501404212637914, + "loss": 1.741, + "theoretical_loss": 4.089946303120092, + "tokens_seen": 358809600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501303911735206, + "loss": 1.6098, + "theoretical_loss": 4.089861797437761, + "tokens_seen": 358875136 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045012036108324976, + "loss": 1.6767, + "theoretical_loss": 4.089777311506094, + "tokens_seen": 358940672 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045011033099297894, + "loss": 1.6863, + "theoretical_loss": 4.089692845316868, + "tokens_seen": 359006208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004501003009027081, + "loss": 1.5681, + "theoretical_loss": 4.08960839886187, + "tokens_seen": 359071744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500902708124373, + "loss": 1.6299, + "theoretical_loss": 4.089523972132886, + "tokens_seen": 359137280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045008024072216654, + "loss": 1.7198, + "theoretical_loss": 4.089439565121712, + "tokens_seen": 359202816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045007021063189567, + "loss": 1.7331, + "theoretical_loss": 4.089355177820146, + "tokens_seen": 359268352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500601805416249, + "loss": 1.6026, + "theoretical_loss": 4.0892708102199915, + "tokens_seen": 359333888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500501504513541, + "loss": 1.7053, + "theoretical_loss": 4.089186462313056, + "tokens_seen": 359399424 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045004012036108326, + "loss": 1.6023, + "theoretical_loss": 4.089102134091155, + "tokens_seen": 359464960 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045003009027081244, + "loss": 1.6543, + "theoretical_loss": 4.0890178255461045, + "tokens_seen": 359530496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500200601805416, + "loss": 1.6748, + "theoretical_loss": 4.08893353666973, + "tokens_seen": 359596032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004500100300902708, + "loss": 1.6049, + "theoretical_loss": 4.088849267453858, + "tokens_seen": 359661568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00045000000000000004, + "loss": 1.5952, + "theoretical_loss": 4.088765017890322, + "tokens_seen": 359727104 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044998996990972917, + "loss": 1.6042, + "theoretical_loss": 4.08868078797096, + "tokens_seen": 359792640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004499799398194584, + "loss": 1.5668, + "theoretical_loss": 4.088596577687614, + "tokens_seen": 359858176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044996990972918753, + "loss": 1.7639, + "theoretical_loss": 4.088512387032132, + "tokens_seen": 359923712 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044995987963891677, + "loss": 1.5918, + "theoretical_loss": 4.088428215996366, + "tokens_seen": 359989248 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044994984954864595, + "loss": 1.6669, + "theoretical_loss": 4.0883440645721745, + "tokens_seen": 360054784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044993981945837513, + "loss": 1.6968, + "theoretical_loss": 4.088259932751419, + "tokens_seen": 360120320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004499297893681043, + "loss": 1.5588, + "theoretical_loss": 4.088175820525966, + "tokens_seen": 360185856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044991975927783355, + "loss": 1.6575, + "theoretical_loss": 4.088091727887688, + "tokens_seen": 360251392 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044990972918756267, + "loss": 1.7509, + "theoretical_loss": 4.088007654828463, + "tokens_seen": 360316928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498996990972919, + "loss": 1.7003, + "theoretical_loss": 4.087923601340169, + "tokens_seen": 360382464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 137426, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7957924604415894, + "objective/train/theoretical_loss": 4.0878395674146955, + "objective/train/tokens_used": 380908000, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044988966900702103, + "loss": 1.7803, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044987963891675027, + "loss": 1.6556, + "theoretical_loss": 4.087755553043934, + "tokens_seen": 360513536 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044986960882647945, + "loss": 1.5828, + "theoretical_loss": 4.087671558219778, + "tokens_seen": 360579072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044985957873620863, + "loss": 1.546, + "theoretical_loss": 4.087587582934131, + "tokens_seen": 360644608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498495486459378, + "loss": 1.7312, + "theoretical_loss": 4.087503627178897, + "tokens_seen": 360710144 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449839518555667, + "loss": 1.7252, + "theoretical_loss": 4.087419690945988, + "tokens_seen": 360775680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498294884653962, + "loss": 1.6041, + "theoretical_loss": 4.087335774227318, + "tokens_seen": 360841216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004498194583751254, + "loss": 1.5158, + "theoretical_loss": 4.0872518770148085, + "tokens_seen": 360906752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044980942828485454, + "loss": 1.6188, + "theoretical_loss": 4.087167999300384, + "tokens_seen": 360972288 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044979939819458377, + "loss": 1.6867, + "theoretical_loss": 4.087084141075974, + "tokens_seen": 361037824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497893681043129, + "loss": 1.7047, + "theoretical_loss": 4.087000302333513, + "tokens_seen": 361103360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044977933801404213, + "loss": 1.7727, + "theoretical_loss": 4.086916483064941, + "tokens_seen": 361168896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497693079237713, + "loss": 1.6796, + "theoretical_loss": 4.086832683262203, + "tokens_seen": 361234432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497592778335005, + "loss": 1.6655, + "theoretical_loss": 4.086748902917246, + "tokens_seen": 361299968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497492477432297, + "loss": 1.618, + "theoretical_loss": 4.086665142022025, + "tokens_seen": 361365504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497392176529589, + "loss": 1.6472, + "theoretical_loss": 4.086581400568499, + "tokens_seen": 361431040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044972918756268804, + "loss": 1.6371, + "theoretical_loss": 4.086497678548631, + "tokens_seen": 361496576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497191574724173, + "loss": 1.7909, + "theoretical_loss": 4.086413975954389, + "tokens_seen": 361562112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004497091273821464, + "loss": 1.7459, + "theoretical_loss": 4.086330292777745, + "tokens_seen": 361627648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044969909729187564, + "loss": 1.6408, + "theoretical_loss": 4.086246629010679, + "tokens_seen": 361693184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496890672016049, + "loss": 1.6858, + "theoretical_loss": 4.086162984645172, + "tokens_seen": 361758720 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449679037111334, + "loss": 1.5698, + "theoretical_loss": 4.086079359673212, + "tokens_seen": 361824256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044966900702106324, + "loss": 1.5779, + "theoretical_loss": 4.085995754086789, + "tokens_seen": 361889792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044965897693079236, + "loss": 1.5135, + "theoretical_loss": 4.085912167877902, + "tokens_seen": 361955328 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496489468405216, + "loss": 1.6681, + "theoretical_loss": 4.085828601038552, + "tokens_seen": 362020864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496389167502508, + "loss": 1.6377, + "theoretical_loss": 4.0857450535607445, + "tokens_seen": 362086400 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044962888665997996, + "loss": 1.7305, + "theoretical_loss": 4.085661525436492, + "tokens_seen": 362151936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044961885656970914, + "loss": 1.7635, + "theoretical_loss": 4.0855780166578075, + "tokens_seen": 362217472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004496088264794383, + "loss": 1.6549, + "theoretical_loss": 4.0854945272167145, + "tokens_seen": 362283008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495987963891675, + "loss": 1.5983, + "theoretical_loss": 4.085411057105237, + "tokens_seen": 362348544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044958876629889674, + "loss": 1.7265, + "theoretical_loss": 4.0853276063154045, + "tokens_seen": 362414080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044957873620862587, + "loss": 1.6337, + "theoretical_loss": 4.085244174839252, + "tokens_seen": 362479616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495687061183551, + "loss": 1.6405, + "theoretical_loss": 4.085160762668819, + "tokens_seen": 362545152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495586760280843, + "loss": 1.6631, + "theoretical_loss": 4.085077369796149, + "tokens_seen": 362610688 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044954864593781346, + "loss": 1.6339, + "theoretical_loss": 4.084993996213292, + "tokens_seen": 362676224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044953861584754264, + "loss": 1.6938, + "theoretical_loss": 4.084910641912301, + "tokens_seen": 362741760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004495285857572718, + "loss": 1.6961, + "theoretical_loss": 4.084827306885234, + "tokens_seen": 362807296 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449518555667001, + "loss": 1.7178, + "theoretical_loss": 4.084743991124155, + "tokens_seen": 362872832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044950852557673024, + "loss": 1.6507, + "theoretical_loss": 4.08466069462113, + "tokens_seen": 362938368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044949849548645937, + "loss": 1.8126, + "theoretical_loss": 4.084577417368232, + "tokens_seen": 363003904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494884653961886, + "loss": 1.7103, + "theoretical_loss": 4.084494159357538, + "tokens_seen": 363069440 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044947843530591773, + "loss": 1.7322, + "theoretical_loss": 4.0844109205811305, + "tokens_seen": 363134976 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044946840521564697, + "loss": 1.7087, + "theoretical_loss": 4.084327701031095, + "tokens_seen": 363200512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044945837512537615, + "loss": 1.6378, + "theoretical_loss": 4.084244500699524, + "tokens_seen": 363266048 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044944834503510533, + "loss": 1.6703, + "theoretical_loss": 4.084161319578511, + "tokens_seen": 363331584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494383149448345, + "loss": 1.6991, + "theoretical_loss": 4.084078157660159, + "tokens_seen": 363397120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044942828485456375, + "loss": 1.5839, + "theoretical_loss": 4.083995014936571, + "tokens_seen": 363462656 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044941825476429287, + "loss": 1.7186, + "theoretical_loss": 4.0839118913998576, + "tokens_seen": 363528192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004494082246740221, + "loss": 1.6219, + "theoretical_loss": 4.083828787042134, + "tokens_seen": 363593728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044939819458375123, + "loss": 1.7147, + "theoretical_loss": 4.083745701855518, + "tokens_seen": 363659264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 138686, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.718767523765564, + "objective/train/theoretical_loss": 4.083662635832135, + "objective/train/tokens_used": 384184800, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044938816449348047, + "loss": 1.7056, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044937813440320965, + "loss": 1.7226, + "theoretical_loss": 4.0835795889641116, + "tokens_seen": 363790336 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044936810431293883, + "loss": 1.6298, + "theoretical_loss": 4.0834965612435825, + "tokens_seen": 363855872 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449358074222668, + "loss": 1.7515, + "theoretical_loss": 4.083413552662685, + "tokens_seen": 363921408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493480441323972, + "loss": 1.6951, + "theoretical_loss": 4.08333056321356, + "tokens_seen": 363986944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493380140421264, + "loss": 1.619, + "theoretical_loss": 4.083247592888357, + "tokens_seen": 364052480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004493279839518556, + "loss": 1.6939, + "theoretical_loss": 4.083164641679225, + "tokens_seen": 364118016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044931795386158474, + "loss": 1.6744, + "theoretical_loss": 4.083081709578323, + "tokens_seen": 364183552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044930792377131397, + "loss": 1.7509, + "theoretical_loss": 4.082998796577811, + "tokens_seen": 364249088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492978936810431, + "loss": 1.6185, + "theoretical_loss": 4.082915902669853, + "tokens_seen": 364314624 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044928786359077234, + "loss": 1.7257, + "theoretical_loss": 4.082833027846622, + "tokens_seen": 364380160 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492778335005015, + "loss": 1.7661, + "theoretical_loss": 4.08275017210029, + "tokens_seen": 364445696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492678034102307, + "loss": 1.6133, + "theoretical_loss": 4.082667335423038, + "tokens_seen": 364511232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492577733199599, + "loss": 1.6884, + "theoretical_loss": 4.0825845178070495, + "tokens_seen": 364576768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492477432296891, + "loss": 1.6118, + "theoretical_loss": 4.0825017192445126, + "tokens_seen": 364642304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044923771313941824, + "loss": 1.6426, + "theoretical_loss": 4.082418939727622, + "tokens_seen": 364707840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492276830491475, + "loss": 1.6247, + "theoretical_loss": 4.082336179248574, + "tokens_seen": 364773376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004492176529588766, + "loss": 1.6112, + "theoretical_loss": 4.082253437799572, + "tokens_seen": 364838912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044920762286860584, + "loss": 1.6993, + "theoretical_loss": 4.082170715372823, + "tokens_seen": 364904448 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449197592778335, + "loss": 1.7887, + "theoretical_loss": 4.082088011960538, + "tokens_seen": 364969984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491875626880642, + "loss": 1.7086, + "theoretical_loss": 4.082005327554934, + "tokens_seen": 365035520 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491775325977934, + "loss": 1.5872, + "theoretical_loss": 4.081922662148231, + "tokens_seen": 365101056 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044916750250752256, + "loss": 1.7102, + "theoretical_loss": 4.081840015732654, + "tokens_seen": 365166592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044915747241725174, + "loss": 1.651, + "theoretical_loss": 4.081757388300435, + "tokens_seen": 365232128 + }, + { + "epoch": 0.11, + "learning_rate": 0.000449147442326981, + "loss": 1.5952, + "theoretical_loss": 4.081674779843806, + "tokens_seen": 365297664 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491374122367101, + "loss": 1.6407, + "theoretical_loss": 4.081592190355008, + "tokens_seen": 365363200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044912738214643934, + "loss": 1.5665, + "theoretical_loss": 4.081509619826283, + "tokens_seen": 365428736 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044911735205616847, + "loss": 1.5865, + "theoretical_loss": 4.081427068249881, + "tokens_seen": 365494272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004491073219658977, + "loss": 1.7228, + "theoretical_loss": 4.081344535618053, + "tokens_seen": 365559808 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490972918756269, + "loss": 1.6367, + "theoretical_loss": 4.081262021923058, + "tokens_seen": 365625344 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044908726178535607, + "loss": 1.7218, + "theoretical_loss": 4.081179527157156, + "tokens_seen": 365690880 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044907723169508525, + "loss": 1.6274, + "theoretical_loss": 4.081097051312614, + "tokens_seen": 365756416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490672016048145, + "loss": 1.6204, + "theoretical_loss": 4.081014594381705, + "tokens_seen": 365821952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490571715145436, + "loss": 1.6433, + "theoretical_loss": 4.0809321563567025, + "tokens_seen": 365887488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044904714142427284, + "loss": 1.5711, + "theoretical_loss": 4.080849737229887, + "tokens_seen": 365953024 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044903711133400197, + "loss": 1.6621, + "theoretical_loss": 4.080767336993542, + "tokens_seen": 366018560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490270812437312, + "loss": 1.6494, + "theoretical_loss": 4.080684955639958, + "tokens_seen": 366084096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004490170511534604, + "loss": 1.7265, + "theoretical_loss": 4.080602593161428, + "tokens_seen": 366149632 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044900702106318957, + "loss": 1.6368, + "theoretical_loss": 4.0805202495502515, + "tokens_seen": 366215168 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044899699097291875, + "loss": 1.659, + "theoretical_loss": 4.080437924798729, + "tokens_seen": 366280704 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044898696088264793, + "loss": 1.5808, + "theoretical_loss": 4.080355618899169, + "tokens_seen": 366346240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489769307923771, + "loss": 1.65, + "theoretical_loss": 4.080273331843884, + "tokens_seen": 366411776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044896690070210635, + "loss": 1.7413, + "theoretical_loss": 4.080191063625188, + "tokens_seen": 366477312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489568706118355, + "loss": 1.597, + "theoretical_loss": 4.080108814235404, + "tokens_seen": 366542848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489468405215647, + "loss": 1.6505, + "theoretical_loss": 4.080026583666855, + "tokens_seen": 366608384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044893681043129395, + "loss": 1.6478, + "theoretical_loss": 4.079944371911871, + "tokens_seen": 366673920 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044892678034102307, + "loss": 1.7445, + "theoretical_loss": 4.07986217896279, + "tokens_seen": 366739456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004489167502507523, + "loss": 1.78, + "theoretical_loss": 4.079780004811946, + "tokens_seen": 366804992 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044890672016048143, + "loss": 1.6541, + "theoretical_loss": 4.079697849451684, + "tokens_seen": 366870528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044889669007021067, + "loss": 1.8983, + "theoretical_loss": 4.079615712874352, + "tokens_seen": 366936064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 139849, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.735669732093811, + "objective/train/theoretical_loss": 4.079533595072301, + "objective/train/tokens_used": 387461600, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044888665997993985, + "loss": 1.7021, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044887662988966903, + "loss": 1.7184, + "theoretical_loss": 4.0794514960378905, + "tokens_seen": 367067136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488665997993982, + "loss": 1.713, + "theoretical_loss": 4.079369415763479, + "tokens_seen": 367132672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488565697091274, + "loss": 1.6383, + "theoretical_loss": 4.079287354241432, + "tokens_seen": 367198208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488465396188566, + "loss": 1.6876, + "theoretical_loss": 4.07920531146412, + "tokens_seen": 367263744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488365095285858, + "loss": 1.6776, + "theoretical_loss": 4.0791232874239185, + "tokens_seen": 367329280 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044882647943831494, + "loss": 1.6454, + "theoretical_loss": 4.079041282113206, + "tokens_seen": 367394816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488164493480442, + "loss": 1.7785, + "theoretical_loss": 4.078959295524365, + "tokens_seen": 367460352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004488064192577733, + "loss": 1.6863, + "theoretical_loss": 4.078877327649785, + "tokens_seen": 367525888 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044879638916750254, + "loss": 1.5702, + "theoretical_loss": 4.078795378481858, + "tokens_seen": 367591424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487863590772317, + "loss": 1.5905, + "theoretical_loss": 4.07871344801298, + "tokens_seen": 367656960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487763289869609, + "loss": 1.7459, + "theoretical_loss": 4.0786315362355525, + "tokens_seen": 367722496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487662988966901, + "loss": 1.6666, + "theoretical_loss": 4.078549643141981, + "tokens_seen": 367788032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487562688064193, + "loss": 1.6385, + "theoretical_loss": 4.078467768724677, + "tokens_seen": 367853568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044874623871614844, + "loss": 1.5945, + "theoretical_loss": 4.078385912976053, + "tokens_seen": 367919104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487362086258777, + "loss": 1.67, + "theoretical_loss": 4.078304075888529, + "tokens_seen": 367984640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487261785356068, + "loss": 1.7311, + "theoretical_loss": 4.078222257454529, + "tokens_seen": 368050176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044871614844533604, + "loss": 1.6936, + "theoretical_loss": 4.07814045766648, + "tokens_seen": 368115712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004487061183550652, + "loss": 1.6735, + "theoretical_loss": 4.0780586765168145, + "tokens_seen": 368181248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486960882647944, + "loss": 1.5526, + "theoretical_loss": 4.07797691399797, + "tokens_seen": 368246784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486860581745236, + "loss": 1.7424, + "theoretical_loss": 4.077895170102386, + "tokens_seen": 368312320 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044867602808425276, + "loss": 1.6952, + "theoretical_loss": 4.077813444822509, + "tokens_seen": 368377856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044866599799398194, + "loss": 1.5788, + "theoretical_loss": 4.077731738150789, + "tokens_seen": 368443392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486559679037112, + "loss": 1.6384, + "theoretical_loss": 4.077650050079681, + "tokens_seen": 368508928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486459378134403, + "loss": 1.5942, + "theoretical_loss": 4.077568380601642, + "tokens_seen": 368574464 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044863590772316954, + "loss": 1.5493, + "theoretical_loss": 4.077486729709136, + "tokens_seen": 368640000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044862587763289867, + "loss": 1.6088, + "theoretical_loss": 4.077405097394631, + "tokens_seen": 368705536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486158475426279, + "loss": 1.7427, + "theoretical_loss": 4.077323483650598, + "tokens_seen": 368771072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004486058174523571, + "loss": 1.7378, + "theoretical_loss": 4.077241888469514, + "tokens_seen": 368836608 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044859578736208627, + "loss": 1.6634, + "theoretical_loss": 4.077160311843858, + "tokens_seen": 368902144 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044858575727181545, + "loss": 1.6583, + "theoretical_loss": 4.077078753766118, + "tokens_seen": 368967680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485757271815447, + "loss": 1.8296, + "theoretical_loss": 4.076997214228782, + "tokens_seen": 369033216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485656970912738, + "loss": 1.6546, + "theoretical_loss": 4.076915693224343, + "tokens_seen": 369098752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044855566700100304, + "loss": 1.7523, + "theoretical_loss": 4.0768341907453, + "tokens_seen": 369164288 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044854563691073217, + "loss": 1.6487, + "theoretical_loss": 4.0767527067841565, + "tokens_seen": 369229824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485356068204614, + "loss": 1.597, + "theoretical_loss": 4.076671241333417, + "tokens_seen": 369295360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004485255767301906, + "loss": 1.6335, + "theoretical_loss": 4.076589794385596, + "tokens_seen": 369360896 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044851554663991977, + "loss": 1.5925, + "theoretical_loss": 4.076508365933206, + "tokens_seen": 369426432 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044850551654964895, + "loss": 1.7029, + "theoretical_loss": 4.076426955968769, + "tokens_seen": 369491968 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044849548645937813, + "loss": 1.6834, + "theoretical_loss": 4.076345564484807, + "tokens_seen": 369557504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484854563691073, + "loss": 1.613, + "theoretical_loss": 4.076264191473852, + "tokens_seen": 369623040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044847542627883655, + "loss": 1.7288, + "theoretical_loss": 4.076182836928435, + "tokens_seen": 369688576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484653961885657, + "loss": 1.7521, + "theoretical_loss": 4.0761015008410935, + "tokens_seen": 369754112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484553660982949, + "loss": 1.697, + "theoretical_loss": 4.07602018320437, + "tokens_seen": 369819648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044844533600802404, + "loss": 1.5918, + "theoretical_loss": 4.075938884010809, + "tokens_seen": 369885184 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044843530591775327, + "loss": 1.6178, + "theoretical_loss": 4.075857603252963, + "tokens_seen": 369950720 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044842527582748245, + "loss": 1.6233, + "theoretical_loss": 4.075776340923385, + "tokens_seen": 370016256 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044841524573721163, + "loss": 1.6959, + "theoretical_loss": 4.075695097014635, + "tokens_seen": 370081792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004484052156469408, + "loss": 1.6713, + "theoretical_loss": 4.075613871519276, + "tokens_seen": 370147328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044839518555667005, + "loss": 1.7204, + "theoretical_loss": 4.0755326644298755, + "tokens_seen": 370212864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 141078, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6299824714660645, + "objective/train/theoretical_loss": 4.075451475739007, + "objective/train/tokens_used": 390738400, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483851554663992, + "loss": 1.5816, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483751253761284, + "loss": 1.6564, + "theoretical_loss": 4.075370305439243, + "tokens_seen": 370343936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044836509528585754, + "loss": 1.6017, + "theoretical_loss": 4.075289153523168, + "tokens_seen": 370409472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483550651955868, + "loss": 1.6342, + "theoretical_loss": 4.075208019983366, + "tokens_seen": 370475008 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044834503510531596, + "loss": 1.7463, + "theoretical_loss": 4.075126904812425, + "tokens_seen": 370540544 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044833500501504514, + "loss": 1.5822, + "theoretical_loss": 4.0750458080029395, + "tokens_seen": 370606080 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483249749247743, + "loss": 1.6845, + "theoretical_loss": 4.074964729547506, + "tokens_seen": 370671616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483149448345035, + "loss": 1.5762, + "theoretical_loss": 4.074883669438729, + "tokens_seen": 370737152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004483049147442327, + "loss": 1.6618, + "theoretical_loss": 4.074802627669213, + "tokens_seen": 370802688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482948846539619, + "loss": 1.6649, + "theoretical_loss": 4.07472160423157, + "tokens_seen": 370868224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044828485456369104, + "loss": 1.6257, + "theoretical_loss": 4.074640599118413, + "tokens_seen": 370933760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482748244734203, + "loss": 1.639, + "theoretical_loss": 4.074559612322363, + "tokens_seen": 370999296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482647943831494, + "loss": 1.6974, + "theoretical_loss": 4.074478643836043, + "tokens_seen": 371064832 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044825476429287864, + "loss": 1.679, + "theoretical_loss": 4.074397693652081, + "tokens_seen": 371130368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482447342026078, + "loss": 1.6789, + "theoretical_loss": 4.074316761763107, + "tokens_seen": 371195904 + }, + { + "epoch": 0.11, + "learning_rate": 0.000448234704112337, + "loss": 1.7182, + "theoretical_loss": 4.074235848161761, + "tokens_seen": 371261440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482246740220662, + "loss": 1.7002, + "theoretical_loss": 4.074154952840681, + "tokens_seen": 371326976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004482146439317954, + "loss": 1.6468, + "theoretical_loss": 4.074074075792513, + "tokens_seen": 371392512 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044820461384152455, + "loss": 1.7202, + "theoretical_loss": 4.073993217009905, + "tokens_seen": 371458048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481945837512538, + "loss": 1.5677, + "theoretical_loss": 4.073912376485511, + "tokens_seen": 371523584 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044818455366098296, + "loss": 1.64, + "theoretical_loss": 4.073831554211989, + "tokens_seen": 371589120 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044817452357071214, + "loss": 1.6041, + "theoretical_loss": 4.0737507501820005, + "tokens_seen": 371654656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481644934804414, + "loss": 1.6233, + "theoretical_loss": 4.073669964388212, + "tokens_seen": 371720192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481544633901705, + "loss": 1.8027, + "theoretical_loss": 4.073589196823293, + "tokens_seen": 371785728 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044814443329989974, + "loss": 1.7451, + "theoretical_loss": 4.073508447479918, + "tokens_seen": 371851264 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044813440320962887, + "loss": 1.7033, + "theoretical_loss": 4.073427716350767, + "tokens_seen": 371916800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481243731193581, + "loss": 1.8209, + "theoretical_loss": 4.0733470034285215, + "tokens_seen": 371982336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004481143430290873, + "loss": 1.6047, + "theoretical_loss": 4.07326630870587, + "tokens_seen": 372047872 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044810431293881647, + "loss": 1.6105, + "theoretical_loss": 4.073185632175504, + "tokens_seen": 372113408 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044809428284854565, + "loss": 1.7156, + "theoretical_loss": 4.073104973830118, + "tokens_seen": 372178944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480842527582749, + "loss": 1.754, + "theoretical_loss": 4.073024333662413, + "tokens_seen": 372244480 + }, + { + "epoch": 0.11, + "learning_rate": 0.000448074222668004, + "loss": 1.675, + "theoretical_loss": 4.072943711665093, + "tokens_seen": 372310016 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044806419257773324, + "loss": 1.5033, + "theoretical_loss": 4.072863107830866, + "tokens_seen": 372375552 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044805416248746237, + "loss": 1.6776, + "theoretical_loss": 4.072782522152444, + "tokens_seen": 372441088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480441323971916, + "loss": 1.6454, + "theoretical_loss": 4.072701954622546, + "tokens_seen": 372506624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004480341023069208, + "loss": 1.7426, + "theoretical_loss": 4.072621405233891, + "tokens_seen": 372572160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044802407221664997, + "loss": 1.7059, + "theoretical_loss": 4.072540873979205, + "tokens_seen": 372637696 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044801404212637915, + "loss": 1.6289, + "theoretical_loss": 4.072460360851217, + "tokens_seen": 372703232 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044800401203610833, + "loss": 1.6561, + "theoretical_loss": 4.07237986584266, + "tokens_seen": 372768768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479939819458375, + "loss": 1.6155, + "theoretical_loss": 4.0722993889462735, + "tokens_seen": 372834304 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044798395185556675, + "loss": 1.6307, + "theoretical_loss": 4.072218930154797, + "tokens_seen": 372899840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479739217652959, + "loss": 1.5433, + "theoretical_loss": 4.072138489460979, + "tokens_seen": 372965376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004479638916750251, + "loss": 1.6248, + "theoretical_loss": 4.0720580668575685, + "tokens_seen": 373030912 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044795386158475424, + "loss": 1.7447, + "theoretical_loss": 4.0719776623373205, + "tokens_seen": 373096448 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044794383149448347, + "loss": 1.6369, + "theoretical_loss": 4.071897275892994, + "tokens_seen": 373161984 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044793380140421265, + "loss": 1.6291, + "theoretical_loss": 4.071816907517349, + "tokens_seen": 373227520 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044792377131394183, + "loss": 1.6005, + "theoretical_loss": 4.071736557203157, + "tokens_seen": 373293056 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447913741223671, + "loss": 1.6345, + "theoretical_loss": 4.071656224943186, + "tokens_seen": 373358592 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044790371113340025, + "loss": 1.6291, + "theoretical_loss": 4.071575910730212, + "tokens_seen": 373424128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478936810431294, + "loss": 1.4936, + "theoretical_loss": 4.071495614557016, + "tokens_seen": 373489664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 142333, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7273192405700684, + "objective/train/theoretical_loss": 4.071415336416378, + "objective/train/tokens_used": 394015200, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478836509528586, + "loss": 1.7275, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044787362086258774, + "loss": 1.5365, + "theoretical_loss": 4.071335076301089, + "tokens_seen": 373620736 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447863590772317, + "loss": 1.7124, + "theoretical_loss": 4.071254834203941, + "tokens_seen": 373686272 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044785356068204616, + "loss": 1.7428, + "theoretical_loss": 4.071174610117729, + "tokens_seen": 373751808 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044784353059177534, + "loss": 1.6615, + "theoretical_loss": 4.071094404035253, + "tokens_seen": 373817344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478335005015045, + "loss": 1.6531, + "theoretical_loss": 4.071014215949317, + "tokens_seen": 373882880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478234704112337, + "loss": 1.5146, + "theoretical_loss": 4.070934045852731, + "tokens_seen": 373948416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478134403209629, + "loss": 1.6749, + "theoretical_loss": 4.070853893738307, + "tokens_seen": 374013952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004478034102306921, + "loss": 1.725, + "theoretical_loss": 4.070773759598861, + "tokens_seen": 374079488 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044779338014042124, + "loss": 1.6104, + "theoretical_loss": 4.070693643427215, + "tokens_seen": 374145024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477833500501505, + "loss": 1.5756, + "theoretical_loss": 4.070613545216193, + "tokens_seen": 374210560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477733199598796, + "loss": 1.6504, + "theoretical_loss": 4.070533464958626, + "tokens_seen": 374276096 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044776328986960884, + "loss": 1.611, + "theoretical_loss": 4.070453402647344, + "tokens_seen": 374341632 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447753259779338, + "loss": 1.7204, + "theoretical_loss": 4.070373358275187, + "tokens_seen": 374407168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477432296890672, + "loss": 1.5861, + "theoretical_loss": 4.070293331834996, + "tokens_seen": 374472704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477331995987964, + "loss": 1.6854, + "theoretical_loss": 4.070213323319616, + "tokens_seen": 374538240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004477231695085256, + "loss": 1.813, + "theoretical_loss": 4.070133332721897, + "tokens_seen": 374603776 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044771313941825475, + "loss": 1.676, + "theoretical_loss": 4.070053360034693, + "tokens_seen": 374669312 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447703109327984, + "loss": 1.6422, + "theoretical_loss": 4.069973405250862, + "tokens_seen": 374734848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476930792377131, + "loss": 1.6704, + "theoretical_loss": 4.069893468363265, + "tokens_seen": 374800384 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044768304914744234, + "loss": 1.6413, + "theoretical_loss": 4.06981354936477, + "tokens_seen": 374865920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476730190571715, + "loss": 1.6324, + "theoretical_loss": 4.069733648248246, + "tokens_seen": 374931456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476629889669007, + "loss": 1.6259, + "theoretical_loss": 4.069653765006565, + "tokens_seen": 374996992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476529588766299, + "loss": 1.6049, + "theoretical_loss": 4.06957389963261, + "tokens_seen": 375062528 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044764292878635907, + "loss": 1.678, + "theoretical_loss": 4.069494052119261, + "tokens_seen": 375128064 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044763289869608825, + "loss": 1.6757, + "theoretical_loss": 4.069414222459403, + "tokens_seen": 375193600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476228686058175, + "loss": 1.7365, + "theoretical_loss": 4.069334410645928, + "tokens_seen": 375259136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004476128385155466, + "loss": 1.6158, + "theoretical_loss": 4.069254616671732, + "tokens_seen": 375324672 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044760280842527585, + "loss": 1.7278, + "theoretical_loss": 4.069174840529712, + "tokens_seen": 375390208 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044759277833500503, + "loss": 1.7554, + "theoretical_loss": 4.069095082212772, + "tokens_seen": 375455744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475827482447342, + "loss": 1.6708, + "theoretical_loss": 4.0690153417138175, + "tokens_seen": 375521280 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475727181544634, + "loss": 1.8008, + "theoretical_loss": 4.06893561902576, + "tokens_seen": 375586816 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044756268806419257, + "loss": 1.7223, + "theoretical_loss": 4.068855914141515, + "tokens_seen": 375652352 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044755265797392175, + "loss": 1.7592, + "theoretical_loss": 4.068776227054, + "tokens_seen": 375717888 + }, + { + "epoch": 0.11, + "learning_rate": 0.000447542627883651, + "loss": 1.5971, + "theoretical_loss": 4.06869655775614, + "tokens_seen": 375783424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475325977933801, + "loss": 1.6664, + "theoretical_loss": 4.068616906240861, + "tokens_seen": 375848960 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044752256770310935, + "loss": 1.6484, + "theoretical_loss": 4.068537272501095, + "tokens_seen": 375914496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475125376128385, + "loss": 1.5955, + "theoretical_loss": 4.068457656529775, + "tokens_seen": 375980032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004475025075225677, + "loss": 1.7916, + "theoretical_loss": 4.068378058319843, + "tokens_seen": 376045568 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474924774322969, + "loss": 1.8013, + "theoretical_loss": 4.068298477864241, + "tokens_seen": 376111104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474824473420261, + "loss": 1.6106, + "theoretical_loss": 4.068218915155917, + "tokens_seen": 376176640 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044747241725175526, + "loss": 1.6469, + "theoretical_loss": 4.068139370187821, + "tokens_seen": 376242176 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044746238716148444, + "loss": 1.6793, + "theoretical_loss": 4.06805984295291, + "tokens_seen": 376307712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474523570712136, + "loss": 1.6089, + "theoretical_loss": 4.067980333444142, + "tokens_seen": 376373248 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044744232698094285, + "loss": 1.6983, + "theoretical_loss": 4.067900841654481, + "tokens_seen": 376438784 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044743229689067204, + "loss": 1.7223, + "theoretical_loss": 4.067821367576894, + "tokens_seen": 376504320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474222668004012, + "loss": 1.6331, + "theoretical_loss": 4.067741911204354, + "tokens_seen": 376569856 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044741223671013045, + "loss": 1.6057, + "theoretical_loss": 4.067662472529834, + "tokens_seen": 376635392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004474022066198596, + "loss": 1.8353, + "theoretical_loss": 4.067583051546316, + "tokens_seen": 376700928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473921765295888, + "loss": 1.6773, + "theoretical_loss": 4.067503648246781, + "tokens_seen": 376766464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 143470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7137348651885986, + "objective/train/theoretical_loss": 4.067424262624218, + "objective/train/tokens_used": 397292000, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044738214643931794, + "loss": 1.7084, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473721163490472, + "loss": 1.7616, + "theoretical_loss": 4.06734489467162, + "tokens_seen": 376897536 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044736208625877636, + "loss": 1.6258, + "theoretical_loss": 4.0672655443819785, + "tokens_seen": 376963072 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044735205616850554, + "loss": 1.6207, + "theoretical_loss": 4.067186211748297, + "tokens_seen": 377028608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473420260782347, + "loss": 1.6956, + "theoretical_loss": 4.067106896763575, + "tokens_seen": 377094144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473319959879639, + "loss": 1.6111, + "theoretical_loss": 4.067027599420824, + "tokens_seen": 377159680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473219658976931, + "loss": 1.6148, + "theoretical_loss": 4.0669483197130525, + "tokens_seen": 377225216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004473119358074223, + "loss": 1.6007, + "theoretical_loss": 4.066869057633277, + "tokens_seen": 377290752 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044730190571715144, + "loss": 1.5993, + "theoretical_loss": 4.066789813174517, + "tokens_seen": 377356288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472918756268807, + "loss": 1.6182, + "theoretical_loss": 4.066710586329796, + "tokens_seen": 377421824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472818455366098, + "loss": 1.6573, + "theoretical_loss": 4.066631377092141, + "tokens_seen": 377487360 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044727181544633904, + "loss": 1.6588, + "theoretical_loss": 4.066552185454582, + "tokens_seen": 377552896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472617853560682, + "loss": 1.5274, + "theoretical_loss": 4.066473011410157, + "tokens_seen": 377618432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472517552657974, + "loss": 1.5824, + "theoretical_loss": 4.066393854951903, + "tokens_seen": 377683968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472417251755266, + "loss": 1.6732, + "theoretical_loss": 4.066314716072864, + "tokens_seen": 377749504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472316950852558, + "loss": 1.5328, + "theoretical_loss": 4.066235594766088, + "tokens_seen": 377815040 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044722166499498495, + "loss": 1.5899, + "theoretical_loss": 4.066156491024625, + "tokens_seen": 377880576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472116349047142, + "loss": 1.5668, + "theoretical_loss": 4.06607740484153, + "tokens_seen": 377946112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004472016048144433, + "loss": 1.7918, + "theoretical_loss": 4.065998336209863, + "tokens_seen": 378011648 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044719157472417254, + "loss": 1.7001, + "theoretical_loss": 4.065919285122686, + "tokens_seen": 378077184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471815446339017, + "loss": 1.63, + "theoretical_loss": 4.065840251573066, + "tokens_seen": 378142720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471715145436309, + "loss": 1.8203, + "theoretical_loss": 4.065761235554075, + "tokens_seen": 378208256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471614844533601, + "loss": 1.5455, + "theoretical_loss": 4.065682237058787, + "tokens_seen": 378273792 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044715145436308927, + "loss": 1.6451, + "theoretical_loss": 4.06560325608028, + "tokens_seen": 378339328 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044714142427281845, + "loss": 1.5554, + "theoretical_loss": 4.065524292611637, + "tokens_seen": 378404864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471313941825477, + "loss": 1.619, + "theoretical_loss": 4.065445346645945, + "tokens_seen": 378470400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004471213640922768, + "loss": 1.6264, + "theoretical_loss": 4.0653664181762945, + "tokens_seen": 378535936 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044711133400200605, + "loss": 1.7386, + "theoretical_loss": 4.06528750719578, + "tokens_seen": 378601472 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044710130391173523, + "loss": 1.7237, + "theoretical_loss": 4.0652086136975, + "tokens_seen": 378667008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470912738214644, + "loss": 1.6014, + "theoretical_loss": 4.065129737674556, + "tokens_seen": 378732544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470812437311936, + "loss": 1.5179, + "theoretical_loss": 4.065050879120054, + "tokens_seen": 378798080 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044707121364092277, + "loss": 1.5846, + "theoretical_loss": 4.064972038027106, + "tokens_seen": 378863616 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044706118355065195, + "loss": 1.6482, + "theoretical_loss": 4.064893214388825, + "tokens_seen": 378929152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470511534603812, + "loss": 1.5966, + "theoretical_loss": 4.064814408198327, + "tokens_seen": 378994688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470411233701103, + "loss": 1.6863, + "theoretical_loss": 4.0647356194487365, + "tokens_seen": 379060224 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044703109327983955, + "loss": 1.5541, + "theoretical_loss": 4.064656848133178, + "tokens_seen": 379125760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470210631895687, + "loss": 1.6307, + "theoretical_loss": 4.064578094244782, + "tokens_seen": 379191296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470110330992979, + "loss": 1.5638, + "theoretical_loss": 4.06449935777668, + "tokens_seen": 379256832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004470010030090271, + "loss": 1.6254, + "theoretical_loss": 4.0644206387220105, + "tokens_seen": 379322368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0004469909729187563, + "loss": 1.6226, + "theoretical_loss": 4.064341937073916, + "tokens_seen": 379387904 + }, + { + "epoch": 0.11, + "learning_rate": 0.00044698094282848546, + "loss": 1.7268, + "theoretical_loss": 4.0642632528255405, + "tokens_seen": 379453440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044697091273821464, + "loss": 1.8032, + "theoretical_loss": 4.064184585970033, + "tokens_seen": 379518976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469608826479438, + "loss": 1.5651, + "theoretical_loss": 4.064105936500547, + "tokens_seen": 379584512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044695085255767305, + "loss": 1.5664, + "theoretical_loss": 4.064027304410239, + "tokens_seen": 379650048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469408224674022, + "loss": 1.6479, + "theoretical_loss": 4.063948689692269, + "tokens_seen": 379715584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469307923771314, + "loss": 1.7367, + "theoretical_loss": 4.063870092339803, + "tokens_seen": 379781120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469207622868606, + "loss": 1.6245, + "theoretical_loss": 4.0637915123460076, + "tokens_seen": 379846656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004469107321965898, + "loss": 1.6379, + "theoretical_loss": 4.0637129497040565, + "tokens_seen": 379912192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044690070210631896, + "loss": 1.5858, + "theoretical_loss": 4.063634404407126, + "tokens_seen": 379977728 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044689067201604814, + "loss": 1.657, + "theoretical_loss": 4.063555876448395, + "tokens_seen": 380043264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 144698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5739954710006714, + "objective/train/theoretical_loss": 4.063477365821048, + "objective/train/tokens_used": 400568800, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468806419257773, + "loss": 1.5462, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044687061183550656, + "loss": 1.6747, + "theoretical_loss": 4.063398872518273, + "tokens_seen": 380174336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468605817452357, + "loss": 1.6194, + "theoretical_loss": 4.0633203965332605, + "tokens_seen": 380239872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468505516549649, + "loss": 1.6226, + "theoretical_loss": 4.063241937859207, + "tokens_seen": 380305408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044684052156469405, + "loss": 1.7031, + "theoretical_loss": 4.063163496489312, + "tokens_seen": 380370944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468304914744233, + "loss": 1.6801, + "theoretical_loss": 4.0630850724167775, + "tokens_seen": 380436480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044682046138415246, + "loss": 1.6311, + "theoretical_loss": 4.063006665634811, + "tokens_seen": 380502016 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044681043129388164, + "loss": 1.7017, + "theoretical_loss": 4.062928276136624, + "tokens_seen": 380567552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004468004012036108, + "loss": 1.5853, + "theoretical_loss": 4.0628499039154296, + "tokens_seen": 380633088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044679037111334, + "loss": 1.5261, + "theoretical_loss": 4.062771548964447, + "tokens_seen": 380698624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467803410230692, + "loss": 1.5691, + "theoretical_loss": 4.062693211276899, + "tokens_seen": 380764160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467703109327984, + "loss": 1.6739, + "theoretical_loss": 4.062614890846012, + "tokens_seen": 380829696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044676028084252755, + "loss": 1.6716, + "theoretical_loss": 4.062536587665014, + "tokens_seen": 380895232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467502507522568, + "loss": 1.5838, + "theoretical_loss": 4.0624583017271405, + "tokens_seen": 380960768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044674022066198597, + "loss": 1.7975, + "theoretical_loss": 4.062380033025629, + "tokens_seen": 381026304 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044673019057171515, + "loss": 1.6897, + "theoretical_loss": 4.062301781553719, + "tokens_seen": 381091840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044672016048144433, + "loss": 1.5415, + "theoretical_loss": 4.062223547304658, + "tokens_seen": 381157376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467101303911735, + "loss": 1.6191, + "theoretical_loss": 4.0621453302716946, + "tokens_seen": 381222912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004467001003009027, + "loss": 1.575, + "theoretical_loss": 4.062067130448081, + "tokens_seen": 381288448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466900702106319, + "loss": 1.605, + "theoretical_loss": 4.061988947827073, + "tokens_seen": 381353984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466800401203611, + "loss": 1.654, + "theoretical_loss": 4.061910782401933, + "tokens_seen": 381419520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466700100300903, + "loss": 1.7782, + "theoretical_loss": 4.061832634165923, + "tokens_seen": 381485056 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044665997993981947, + "loss": 1.625, + "theoretical_loss": 4.061754503112313, + "tokens_seen": 381550592 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044664994984954865, + "loss": 1.5312, + "theoretical_loss": 4.061676389234375, + "tokens_seen": 381616128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004466399197592779, + "loss": 1.5762, + "theoretical_loss": 4.061598292525382, + "tokens_seen": 381681664 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446629889669007, + "loss": 1.7269, + "theoretical_loss": 4.061520212978616, + "tokens_seen": 381747200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044661985957873625, + "loss": 1.6354, + "theoretical_loss": 4.061442150587358, + "tokens_seen": 381812736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044660982948846543, + "loss": 1.6599, + "theoretical_loss": 4.0613641053448974, + "tokens_seen": 381878272 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465997993981946, + "loss": 1.7435, + "theoretical_loss": 4.061286077244523, + "tokens_seen": 381943808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465897693079238, + "loss": 1.764, + "theoretical_loss": 4.06120806627953, + "tokens_seen": 382009344 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044657973921765297, + "loss": 1.679, + "theoretical_loss": 4.061130072443216, + "tokens_seen": 382074880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044656970912738215, + "loss": 1.6121, + "theoretical_loss": 4.0610520957288845, + "tokens_seen": 382140416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465596790371114, + "loss": 1.6431, + "theoretical_loss": 4.06097413612984, + "tokens_seen": 382205952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465496489468405, + "loss": 1.6122, + "theoretical_loss": 4.060896193639392, + "tokens_seen": 382271488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044653961885656975, + "loss": 1.7333, + "theoretical_loss": 4.060818268250855, + "tokens_seen": 382337024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465295887662989, + "loss": 1.7015, + "theoretical_loss": 4.060740359957546, + "tokens_seen": 382402560 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465195586760281, + "loss": 1.6176, + "theoretical_loss": 4.060662468752784, + "tokens_seen": 382468096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004465095285857573, + "loss": 1.7565, + "theoretical_loss": 4.060584594629895, + "tokens_seen": 382533632 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464994984954865, + "loss": 1.6022, + "theoretical_loss": 4.060506737582208, + "tokens_seen": 382599168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044648946840521566, + "loss": 1.7163, + "theoretical_loss": 4.060428897603054, + "tokens_seen": 382664704 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044647943831494484, + "loss": 1.6812, + "theoretical_loss": 4.060351074685769, + "tokens_seen": 382730240 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446469408224674, + "loss": 1.5504, + "theoretical_loss": 4.060273268823693, + "tokens_seen": 382795776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044645937813440325, + "loss": 1.6748, + "theoretical_loss": 4.060195480010169, + "tokens_seen": 382861312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464493480441324, + "loss": 1.5722, + "theoretical_loss": 4.060117708238544, + "tokens_seen": 382926848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464393179538616, + "loss": 1.7189, + "theoretical_loss": 4.060039953502169, + "tokens_seen": 382992384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004464292878635908, + "loss": 1.578, + "theoretical_loss": 4.059962215794399, + "tokens_seen": 383057920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044641925777332, + "loss": 1.5914, + "theoretical_loss": 4.059884495108591, + "tokens_seen": 383123456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044640922768304916, + "loss": 1.7133, + "theoretical_loss": 4.059806791438108, + "tokens_seen": 383188992 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044639919759277834, + "loss": 1.5448, + "theoretical_loss": 4.059729104776316, + "tokens_seen": 383254528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463891675025075, + "loss": 1.6561, + "theoretical_loss": 4.059651435116583, + "tokens_seen": 383320064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 145827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.651728630065918, + "objective/train/theoretical_loss": 4.059573782452284, + "objective/train/tokens_used": 403845600, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044637913741223676, + "loss": 1.6277, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463691073219659, + "loss": 1.6103, + "theoretical_loss": 4.0594961467767945, + "tokens_seen": 383451136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463590772316951, + "loss": 1.5868, + "theoretical_loss": 4.059418528083495, + "tokens_seen": 383516672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044634904714142425, + "loss": 1.676, + "theoretical_loss": 4.0593409263657705, + "tokens_seen": 383582208 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004463390170511535, + "loss": 1.6348, + "theoretical_loss": 4.059263341617009, + "tokens_seen": 383647744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044632898696088266, + "loss": 1.5238, + "theoretical_loss": 4.059185773830602, + "tokens_seen": 383713280 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044631895687061184, + "loss": 1.7146, + "theoretical_loss": 4.059108222999944, + "tokens_seen": 383778816 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446308926780341, + "loss": 1.6459, + "theoretical_loss": 4.059030689118436, + "tokens_seen": 383844352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462988966900702, + "loss": 1.5756, + "theoretical_loss": 4.058953172179479, + "tokens_seen": 383909888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462888665997994, + "loss": 1.7022, + "theoretical_loss": 4.05887567217648, + "tokens_seen": 383975424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462788365095286, + "loss": 1.6037, + "theoretical_loss": 4.05879818910285, + "tokens_seen": 384040960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044626880641925775, + "loss": 1.7308, + "theoretical_loss": 4.058720722952002, + "tokens_seen": 384106496 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446258776328987, + "loss": 1.6338, + "theoretical_loss": 4.058643273717355, + "tokens_seen": 384172032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044624874623871617, + "loss": 1.5726, + "theoretical_loss": 4.058565841392328, + "tokens_seen": 384237568 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044623871614844535, + "loss": 1.6806, + "theoretical_loss": 4.058488425970347, + "tokens_seen": 384303104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044622868605817453, + "loss": 1.5524, + "theoretical_loss": 4.058411027444841, + "tokens_seen": 384368640 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462186559679037, + "loss": 1.6557, + "theoretical_loss": 4.05833364580924, + "tokens_seen": 384434176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004462086258776329, + "loss": 1.6317, + "theoretical_loss": 4.058256281056984, + "tokens_seen": 384499712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461985957873621, + "loss": 1.6462, + "theoretical_loss": 4.058178933181509, + "tokens_seen": 384565248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044618856569709125, + "loss": 1.6151, + "theoretical_loss": 4.05810160217626, + "tokens_seen": 384630784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461785356068205, + "loss": 1.6453, + "theoretical_loss": 4.058024288034684, + "tokens_seen": 384696320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461685055165496, + "loss": 1.5582, + "theoretical_loss": 4.057946990750231, + "tokens_seen": 384761856 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044615847542627885, + "loss": 1.7469, + "theoretical_loss": 4.057869710316355, + "tokens_seen": 384827392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044614844533600803, + "loss": 1.6498, + "theoretical_loss": 4.0577924467265145, + "tokens_seen": 384892928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461384152457372, + "loss": 1.7663, + "theoretical_loss": 4.057715199974171, + "tokens_seen": 384958464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004461283851554664, + "loss": 1.6327, + "theoretical_loss": 4.05763797005279, + "tokens_seen": 385024000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044611835506519563, + "loss": 1.6294, + "theoretical_loss": 4.05756075695584, + "tokens_seen": 385089536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044610832497492476, + "loss": 1.7097, + "theoretical_loss": 4.0574835606767925, + "tokens_seen": 385155072 + }, + { + "epoch": 0.12, + "learning_rate": 0.000446098294884654, + "loss": 1.6955, + "theoretical_loss": 4.057406381209126, + "tokens_seen": 385220608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460882647943831, + "loss": 1.6649, + "theoretical_loss": 4.057329218546318, + "tokens_seen": 385286144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044607823470411235, + "loss": 1.6661, + "theoretical_loss": 4.0572520726818535, + "tokens_seen": 385351680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044606820461384153, + "loss": 1.6616, + "theoretical_loss": 4.057174943609218, + "tokens_seen": 385417216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460581745235707, + "loss": 1.6431, + "theoretical_loss": 4.057097831321904, + "tokens_seen": 385482752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460481444332999, + "loss": 1.634, + "theoretical_loss": 4.057020735813405, + "tokens_seen": 385548288 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460381143430291, + "loss": 1.6407, + "theoretical_loss": 4.056943657077219, + "tokens_seen": 385613824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044602808425275826, + "loss": 1.5868, + "theoretical_loss": 4.056866595106847, + "tokens_seen": 385679360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460180541624875, + "loss": 1.7129, + "theoretical_loss": 4.056789549895795, + "tokens_seen": 385744896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004460080240722166, + "loss": 1.6326, + "theoretical_loss": 4.056712521437571, + "tokens_seen": 385810432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044599799398194586, + "loss": 1.6288, + "theoretical_loss": 4.056635509725689, + "tokens_seen": 385875968 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445987963891675, + "loss": 1.7551, + "theoretical_loss": 4.056558514753664, + "tokens_seen": 385941504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459779338014042, + "loss": 1.5533, + "theoretical_loss": 4.056481536515014, + "tokens_seen": 386007040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459679037111334, + "loss": 1.6081, + "theoretical_loss": 4.056404575003266, + "tokens_seen": 386072576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459578736208626, + "loss": 1.6079, + "theoretical_loss": 4.0563276302119435, + "tokens_seen": 386138112 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044594784353059176, + "loss": 1.579, + "theoretical_loss": 4.056250702134578, + "tokens_seen": 386203648 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445937813440321, + "loss": 1.6539, + "theoretical_loss": 4.0561737907647055, + "tokens_seen": 386269184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004459277833500502, + "loss": 1.6419, + "theoretical_loss": 4.056096896095861, + "tokens_seen": 386334720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044591775325977936, + "loss": 1.5792, + "theoretical_loss": 4.056020018121587, + "tokens_seen": 386400256 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044590772316950854, + "loss": 1.5475, + "theoretical_loss": 4.055943156835428, + "tokens_seen": 386465792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458976930792377, + "loss": 1.6022, + "theoretical_loss": 4.055866312230933, + "tokens_seen": 386531328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044588766298896696, + "loss": 1.641, + "theoretical_loss": 4.055789484301653, + "tokens_seen": 386596864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 146981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.766300082206726, + "objective/train/theoretical_loss": 4.055712673041144, + "objective/train/tokens_used": 407122400, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458776328986961, + "loss": 1.7718, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458676028084253, + "loss": 1.6839, + "theoretical_loss": 4.055635878442967, + "tokens_seen": 386727936 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044585757271815445, + "loss": 1.6192, + "theoretical_loss": 4.055559100500682, + "tokens_seen": 386793472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458475426278837, + "loss": 1.6339, + "theoretical_loss": 4.055482339207858, + "tokens_seen": 386859008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044583751253761286, + "loss": 1.511, + "theoretical_loss": 4.055405594558062, + "tokens_seen": 386924544 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044582748244734204, + "loss": 1.6065, + "theoretical_loss": 4.05532886654487, + "tokens_seen": 386990080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458174523570712, + "loss": 1.7144, + "theoretical_loss": 4.055252155161858, + "tokens_seen": 387055616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004458074222668004, + "loss": 1.723, + "theoretical_loss": 4.055175460402607, + "tokens_seen": 387121152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457973921765296, + "loss": 1.7377, + "theoretical_loss": 4.055098782260701, + "tokens_seen": 387186688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457873620862588, + "loss": 1.5674, + "theoretical_loss": 4.055022120729728, + "tokens_seen": 387252224 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044577733199598795, + "loss": 1.572, + "theoretical_loss": 4.054945475803278, + "tokens_seen": 387317760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457673019057172, + "loss": 1.7204, + "theoretical_loss": 4.054868847474948, + "tokens_seen": 387383296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044575727181544637, + "loss": 1.5541, + "theoretical_loss": 4.054792235738335, + "tokens_seen": 387448832 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044574724172517555, + "loss": 1.5576, + "theoretical_loss": 4.054715640587042, + "tokens_seen": 387514368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044573721163490473, + "loss": 1.6493, + "theoretical_loss": 4.054639062014673, + "tokens_seen": 387579904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457271815446339, + "loss": 1.6331, + "theoretical_loss": 4.054562500014839, + "tokens_seen": 387645440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457171514543631, + "loss": 1.6614, + "theoretical_loss": 4.054485954581152, + "tokens_seen": 387710976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004457071213640923, + "loss": 1.7314, + "theoretical_loss": 4.054409425707227, + "tokens_seen": 387776512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044569709127382145, + "loss": 1.6372, + "theoretical_loss": 4.054332913386686, + "tokens_seen": 387842048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456870611835507, + "loss": 1.5999, + "theoretical_loss": 4.054256417613149, + "tokens_seen": 387907584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456770310932798, + "loss": 1.6345, + "theoretical_loss": 4.054179938380246, + "tokens_seen": 387973120 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044566700100300905, + "loss": 1.5837, + "theoretical_loss": 4.054103475681605, + "tokens_seen": 388038656 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044565697091273823, + "loss": 1.4546, + "theoretical_loss": 4.054027029510861, + "tokens_seen": 388104192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456469408224674, + "loss": 1.5157, + "theoretical_loss": 4.053950599861651, + "tokens_seen": 388169728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456369107321966, + "loss": 1.6111, + "theoretical_loss": 4.053874186727617, + "tokens_seen": 388235264 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044562688064192583, + "loss": 1.7551, + "theoretical_loss": 4.053797790102402, + "tokens_seen": 388300800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044561685055165496, + "loss": 1.659, + "theoretical_loss": 4.053721409979654, + "tokens_seen": 388366336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004456068204613842, + "loss": 1.6899, + "theoretical_loss": 4.053645046353025, + "tokens_seen": 388431872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455967903711133, + "loss": 1.5093, + "theoretical_loss": 4.05356869921617, + "tokens_seen": 388497408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044558676028084255, + "loss": 1.6176, + "theoretical_loss": 4.053492368562747, + "tokens_seen": 388562944 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044557673019057173, + "loss": 1.7148, + "theoretical_loss": 4.0534160543864175, + "tokens_seen": 388628480 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455667001003009, + "loss": 1.65, + "theoretical_loss": 4.053339756680849, + "tokens_seen": 388694016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455566700100301, + "loss": 1.6395, + "theoretical_loss": 4.053263475439707, + "tokens_seen": 388759552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455466399197593, + "loss": 1.5527, + "theoretical_loss": 4.053187210656668, + "tokens_seen": 388825088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044553660982948846, + "loss": 1.6992, + "theoretical_loss": 4.053110962325406, + "tokens_seen": 388890624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455265797392177, + "loss": 1.5218, + "theoretical_loss": 4.053034730439601, + "tokens_seen": 388956160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004455165496489468, + "loss": 1.6331, + "theoretical_loss": 4.052958514992934, + "tokens_seen": 389021696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044550651955867606, + "loss": 1.6635, + "theoretical_loss": 4.052882315979094, + "tokens_seen": 389087232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454964894684052, + "loss": 1.4846, + "theoretical_loss": 4.052806133391769, + "tokens_seen": 389152768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454864593781344, + "loss": 1.6703, + "theoretical_loss": 4.052729967224654, + "tokens_seen": 389218304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454764292878636, + "loss": 1.5628, + "theoretical_loss": 4.0526538174714455, + "tokens_seen": 389283840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454663991975928, + "loss": 1.6734, + "theoretical_loss": 4.052577684125843, + "tokens_seen": 389349376 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044545636910732196, + "loss": 1.7154, + "theoretical_loss": 4.052501567181552, + "tokens_seen": 389414912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454463390170512, + "loss": 1.6772, + "theoretical_loss": 4.052425466632278, + "tokens_seen": 389480448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454363089267803, + "loss": 1.64, + "theoretical_loss": 4.052349382471734, + "tokens_seen": 389545984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044542627883650956, + "loss": 1.4726, + "theoretical_loss": 4.052273314693632, + "tokens_seen": 389611520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454162487462387, + "loss": 1.6242, + "theoretical_loss": 4.05219726329169, + "tokens_seen": 389677056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004454062186559679, + "loss": 1.7203, + "theoretical_loss": 4.052121228259631, + "tokens_seen": 389742592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453961885656971, + "loss": 1.6618, + "theoretical_loss": 4.052045209591179, + "tokens_seen": 389808128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453861584754263, + "loss": 1.6137, + "theoretical_loss": 4.05196920728006, + "tokens_seen": 389873664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 148229, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7491153478622437, + "objective/train/theoretical_loss": 4.051893221320009, + "objective/train/tokens_used": 410399200, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044537612838515547, + "loss": 1.7324, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044536609829488465, + "loss": 1.6581, + "theoretical_loss": 4.05181725170476, + "tokens_seen": 390004736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044535606820461383, + "loss": 1.6557, + "theoretical_loss": 4.051741298428049, + "tokens_seen": 390070272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044534603811434306, + "loss": 1.6371, + "theoretical_loss": 4.0516653614836216, + "tokens_seen": 390135808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453360080240722, + "loss": 1.6195, + "theoretical_loss": 4.05158944086522, + "tokens_seen": 390201344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453259779338014, + "loss": 1.6273, + "theoretical_loss": 4.0515135365665955, + "tokens_seen": 390266880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044531594784353055, + "loss": 1.5038, + "theoretical_loss": 4.0514376485815, + "tokens_seen": 390332416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004453059177532598, + "loss": 1.615, + "theoretical_loss": 4.051361776903688, + "tokens_seen": 390397952 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044529588766298897, + "loss": 1.6801, + "theoretical_loss": 4.051285921526919, + "tokens_seen": 390463488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044528585757271815, + "loss": 1.6663, + "theoretical_loss": 4.051210082444956, + "tokens_seen": 390529024 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044527582748244733, + "loss": 1.5642, + "theoretical_loss": 4.051134259651565, + "tokens_seen": 390594560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044526579739217657, + "loss": 1.7364, + "theoretical_loss": 4.0510584531405165, + "tokens_seen": 390660096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004452557673019057, + "loss": 1.5721, + "theoretical_loss": 4.050982662905581, + "tokens_seen": 390725632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044524573721163493, + "loss": 1.6163, + "theoretical_loss": 4.0509068889405375, + "tokens_seen": 390791168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044523570712136406, + "loss": 1.7385, + "theoretical_loss": 4.050831131239164, + "tokens_seen": 390856704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004452256770310933, + "loss": 1.7252, + "theoretical_loss": 4.050755389795244, + "tokens_seen": 390922240 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044521564694082247, + "loss": 1.683, + "theoretical_loss": 4.050679664602564, + "tokens_seen": 390987776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044520561685055165, + "loss": 1.6829, + "theoretical_loss": 4.050603955654914, + "tokens_seen": 391053312 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044519558676028083, + "loss": 1.6138, + "theoretical_loss": 4.050528262946088, + "tokens_seen": 391118848 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044518555667001, + "loss": 1.6244, + "theoretical_loss": 4.050452586469882, + "tokens_seen": 391184384 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044517552657973925, + "loss": 1.4319, + "theoretical_loss": 4.050376926220098, + "tokens_seen": 391249920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044516549648946843, + "loss": 1.5461, + "theoretical_loss": 4.050301282190537, + "tokens_seen": 391315456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451554663991976, + "loss": 1.6967, + "theoretical_loss": 4.050225654375008, + "tokens_seen": 391380992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451454363089268, + "loss": 1.7262, + "theoretical_loss": 4.050150042767321, + "tokens_seen": 391446528 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044513540621865603, + "loss": 1.5765, + "theoretical_loss": 4.050074447361289, + "tokens_seen": 391512064 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044512537612838516, + "loss": 1.5929, + "theoretical_loss": 4.04999886815073, + "tokens_seen": 391577600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451153460381144, + "loss": 1.6146, + "theoretical_loss": 4.0499233051294645, + "tokens_seen": 391643136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004451053159478435, + "loss": 1.6706, + "theoretical_loss": 4.049847758291317, + "tokens_seen": 391708672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044509528585757275, + "loss": 1.633, + "theoretical_loss": 4.049772227630114, + "tokens_seen": 391774208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044508525576730194, + "loss": 1.6285, + "theoretical_loss": 4.049696713139687, + "tokens_seen": 391839744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450752256770311, + "loss": 1.6542, + "theoretical_loss": 4.049621214813869, + "tokens_seen": 391905280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450651955867603, + "loss": 1.6003, + "theoretical_loss": 4.0495457326465, + "tokens_seen": 391970816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450551654964895, + "loss": 1.4902, + "theoretical_loss": 4.049470266631419, + "tokens_seen": 392036352 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044504513540621866, + "loss": 1.6949, + "theoretical_loss": 4.04939481676247, + "tokens_seen": 392101888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450351053159479, + "loss": 1.7049, + "theoretical_loss": 4.0493193830335015, + "tokens_seen": 392167424 + }, + { + "epoch": 0.12, + "learning_rate": 0.000445025075225677, + "loss": 1.5919, + "theoretical_loss": 4.0492439654383645, + "tokens_seen": 392232960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044501504513540626, + "loss": 1.7086, + "theoretical_loss": 4.049168563970914, + "tokens_seen": 392298496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004450050150451354, + "loss": 1.6493, + "theoretical_loss": 4.049093178625007, + "tokens_seen": 392364032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449949849548646, + "loss": 1.5826, + "theoretical_loss": 4.049017809394506, + "tokens_seen": 392429568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449849548645938, + "loss": 1.5196, + "theoretical_loss": 4.048942456273274, + "tokens_seen": 392495104 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444974924774323, + "loss": 1.5659, + "theoretical_loss": 4.048867119255178, + "tokens_seen": 392560640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044496489468405216, + "loss": 1.6535, + "theoretical_loss": 4.048791798334092, + "tokens_seen": 392626176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449548645937814, + "loss": 1.6789, + "theoretical_loss": 4.048716493503888, + "tokens_seen": 392691712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449448345035105, + "loss": 1.7316, + "theoretical_loss": 4.048641204758446, + "tokens_seen": 392757248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044493480441323976, + "loss": 1.6505, + "theoretical_loss": 4.048565932091646, + "tokens_seen": 392822784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449247743229689, + "loss": 1.6901, + "theoretical_loss": 4.048490675497373, + "tokens_seen": 392888320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449147442326981, + "loss": 1.6574, + "theoretical_loss": 4.048415434969516, + "tokens_seen": 392953856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004449047141424273, + "loss": 1.6081, + "theoretical_loss": 4.048340210501965, + "tokens_seen": 393019392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448946840521565, + "loss": 1.6501, + "theoretical_loss": 4.048265002088614, + "tokens_seen": 393084928 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044488465396188567, + "loss": 1.5724, + "theoretical_loss": 4.048189809723364, + "tokens_seen": 393150464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 149327, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7710505723953247, + "objective/train/theoretical_loss": 4.048114633400113, + "objective/train/tokens_used": 413676000, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044487462387161485, + "loss": 1.6523, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044486459378134403, + "loss": 1.6695, + "theoretical_loss": 4.048039473112768, + "tokens_seen": 393281536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044485456369107326, + "loss": 1.5928, + "theoretical_loss": 4.047964328855237, + "tokens_seen": 393347072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448445336008024, + "loss": 1.6453, + "theoretical_loss": 4.047889200621428, + "tokens_seen": 393412608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004448345035105316, + "loss": 1.7137, + "theoretical_loss": 4.047814088405261, + "tokens_seen": 393478144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044482447342026075, + "loss": 1.729, + "theoretical_loss": 4.0477389922006495, + "tokens_seen": 393543680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044481444332999, + "loss": 1.662, + "theoretical_loss": 4.047663912001518, + "tokens_seen": 393609216 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044480441323971917, + "loss": 1.6649, + "theoretical_loss": 4.047588847801788, + "tokens_seen": 393674752 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044479438314944835, + "loss": 1.6863, + "theoretical_loss": 4.04751379959539, + "tokens_seen": 393740288 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044478435305917753, + "loss": 1.6529, + "theoretical_loss": 4.047438767376255, + "tokens_seen": 393805824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044477432296890677, + "loss": 1.6983, + "theoretical_loss": 4.047363751138316, + "tokens_seen": 393871360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004447642928786359, + "loss": 1.5913, + "theoretical_loss": 4.047288750875512, + "tokens_seen": 393936896 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044475426278836513, + "loss": 1.6629, + "theoretical_loss": 4.047213766581785, + "tokens_seen": 394002432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044474423269809426, + "loss": 1.4895, + "theoretical_loss": 4.047138798251078, + "tokens_seen": 394067968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004447342026078235, + "loss": 1.6533, + "theoretical_loss": 4.047063845877339, + "tokens_seen": 394133504 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044472417251755267, + "loss": 1.6216, + "theoretical_loss": 4.04698890945452, + "tokens_seen": 394199040 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044471414242728185, + "loss": 1.7196, + "theoretical_loss": 4.046913988976575, + "tokens_seen": 394264576 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044470411233701103, + "loss": 1.6393, + "theoretical_loss": 4.046839084437463, + "tokens_seen": 394330112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446940822467402, + "loss": 1.6775, + "theoretical_loss": 4.046764195831142, + "tokens_seen": 394395648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446840521564694, + "loss": 1.7146, + "theoretical_loss": 4.046689323151578, + "tokens_seen": 394461184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044467402206619863, + "loss": 1.6242, + "theoretical_loss": 4.0466144663927395, + "tokens_seen": 394526720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044466399197592776, + "loss": 1.7348, + "theoretical_loss": 4.046539625548595, + "tokens_seen": 394592256 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444653961885657, + "loss": 1.731, + "theoretical_loss": 4.046464800613121, + "tokens_seen": 394657792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446439317953861, + "loss": 1.5386, + "theoretical_loss": 4.046389991580294, + "tokens_seen": 394723328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044463390170511536, + "loss": 1.5501, + "theoretical_loss": 4.046315198444094, + "tokens_seen": 394788864 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044462387161484454, + "loss": 1.6102, + "theoretical_loss": 4.046240421198506, + "tokens_seen": 394854400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446138415245737, + "loss": 1.6931, + "theoretical_loss": 4.046165659837517, + "tokens_seen": 394919936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004446038114343029, + "loss": 1.6748, + "theoretical_loss": 4.046090914355117, + "tokens_seen": 394985472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044459378134403214, + "loss": 1.6545, + "theoretical_loss": 4.046016184745302, + "tokens_seen": 395051008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044458375125376126, + "loss": 1.6698, + "theoretical_loss": 4.045941471002066, + "tokens_seen": 395116544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445737211634905, + "loss": 1.688, + "theoretical_loss": 4.0458667731194105, + "tokens_seen": 395182080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445636910732196, + "loss": 1.6492, + "theoretical_loss": 4.045792091091339, + "tokens_seen": 395247616 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044455366098294886, + "loss": 1.681, + "theoretical_loss": 4.04571742491186, + "tokens_seen": 395313152 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044454363089267804, + "loss": 1.5878, + "theoretical_loss": 4.045642774574981, + "tokens_seen": 395378688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445336008024072, + "loss": 1.658, + "theoretical_loss": 4.045568140074717, + "tokens_seen": 395444224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445235707121364, + "loss": 1.5583, + "theoretical_loss": 4.045493521405085, + "tokens_seen": 395509760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004445135406218656, + "loss": 1.742, + "theoretical_loss": 4.045418918560104, + "tokens_seen": 395575296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044450351053159477, + "loss": 1.567, + "theoretical_loss": 4.045344331533796, + "tokens_seen": 395640832 + }, + { + "epoch": 0.12, + "learning_rate": 0.000444493480441324, + "loss": 1.7221, + "theoretical_loss": 4.04526976032019, + "tokens_seen": 395706368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044448345035105313, + "loss": 1.6933, + "theoretical_loss": 4.0451952049133135, + "tokens_seen": 395771904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044447342026078236, + "loss": 1.7121, + "theoretical_loss": 4.0451206653072, + "tokens_seen": 395837440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444633901705115, + "loss": 1.5836, + "theoretical_loss": 4.045046141495887, + "tokens_seen": 395902976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444533600802407, + "loss": 1.6472, + "theoretical_loss": 4.044971633473412, + "tokens_seen": 395968512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044444332998996996, + "loss": 1.5673, + "theoretical_loss": 4.044897141233818, + "tokens_seen": 396034048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444332998996991, + "loss": 1.6416, + "theoretical_loss": 4.044822664771151, + "tokens_seen": 396099584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444232698094283, + "loss": 1.5984, + "theoretical_loss": 4.044748204079459, + "tokens_seen": 396165120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444132397191575, + "loss": 1.6966, + "theoretical_loss": 4.044673759152797, + "tokens_seen": 396230656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004444032096288867, + "loss": 1.7384, + "theoretical_loss": 4.0445993299852185, + "tokens_seen": 396296192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044439317953861587, + "loss": 1.7425, + "theoretical_loss": 4.044524916570782, + "tokens_seen": 396361728 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044438314944834505, + "loss": 1.7457, + "theoretical_loss": 4.04445051890355, + "tokens_seen": 396427264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 150450, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6188480854034424, + "objective/train/theoretical_loss": 4.044376136977588, + "objective/train/tokens_used": 416952800, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044437311935807423, + "loss": 1.5953, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044436308926780346, + "loss": 1.6348, + "theoretical_loss": 4.044301770786964, + "tokens_seen": 396558336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443530591775326, + "loss": 1.6738, + "theoretical_loss": 4.044227420325749, + "tokens_seen": 396623872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443430290872618, + "loss": 1.7662, + "theoretical_loss": 4.044153085588018, + "tokens_seen": 396689408 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044433299899699095, + "loss": 1.5674, + "theoretical_loss": 4.044078766567851, + "tokens_seen": 396754944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004443229689067202, + "loss": 1.6953, + "theoretical_loss": 4.044004463259327, + "tokens_seen": 396820480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044431293881644937, + "loss": 1.5566, + "theoretical_loss": 4.043930175656531, + "tokens_seen": 396886016 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044430290872617855, + "loss": 1.5061, + "theoretical_loss": 4.043855903753551, + "tokens_seen": 396951552 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044429287863590773, + "loss": 1.5186, + "theoretical_loss": 4.043781647544477, + "tokens_seen": 397017088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044428284854563697, + "loss": 1.7074, + "theoretical_loss": 4.043707407023404, + "tokens_seen": 397082624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442728184553661, + "loss": 1.696, + "theoretical_loss": 4.043633182184428, + "tokens_seen": 397148160 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044426278836509533, + "loss": 1.58, + "theoretical_loss": 4.043558973021652, + "tokens_seen": 397213696 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044425275827482446, + "loss": 1.6085, + "theoretical_loss": 4.043484779529176, + "tokens_seen": 397279232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442427281845537, + "loss": 1.7364, + "theoretical_loss": 4.04341060170111, + "tokens_seen": 397344768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044423269809428287, + "loss": 1.7542, + "theoretical_loss": 4.043336439531561, + "tokens_seen": 397410304 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044422266800401205, + "loss": 1.7076, + "theoretical_loss": 4.043262293014646, + "tokens_seen": 397475840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044421263791374123, + "loss": 1.7207, + "theoretical_loss": 4.043188162144478, + "tokens_seen": 397541376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004442026078234704, + "loss": 1.7174, + "theoretical_loss": 4.0431140469151785, + "tokens_seen": 397606912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441925777331996, + "loss": 1.6668, + "theoretical_loss": 4.043039947320869, + "tokens_seen": 397672448 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044418254764292883, + "loss": 1.7832, + "theoretical_loss": 4.0429658633556755, + "tokens_seen": 397737984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044417251755265796, + "loss": 1.7174, + "theoretical_loss": 4.042891795013729, + "tokens_seen": 397803520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441624874623872, + "loss": 1.7039, + "theoretical_loss": 4.042817742289159, + "tokens_seen": 397869056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441524573721163, + "loss": 1.6779, + "theoretical_loss": 4.042743705176102, + "tokens_seen": 397934592 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044414242728184556, + "loss": 1.6394, + "theoretical_loss": 4.042669683668698, + "tokens_seen": 398000128 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044413239719157474, + "loss": 1.6834, + "theoretical_loss": 4.042595677761087, + "tokens_seen": 398065664 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441223671013039, + "loss": 1.534, + "theoretical_loss": 4.0425216874474135, + "tokens_seen": 398131200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004441123370110331, + "loss": 1.6435, + "theoretical_loss": 4.0424477127218275, + "tokens_seen": 398196736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044410230692076234, + "loss": 1.7769, + "theoretical_loss": 4.042373753578479, + "tokens_seen": 398262272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044409227683049146, + "loss": 1.7299, + "theoretical_loss": 4.042299810011523, + "tokens_seen": 398327808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440822467402207, + "loss": 1.7765, + "theoretical_loss": 4.042225882015115, + "tokens_seen": 398393344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440722166499498, + "loss": 1.6358, + "theoretical_loss": 4.0421519695834185, + "tokens_seen": 398458880 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044406218655967906, + "loss": 1.5244, + "theoretical_loss": 4.042078072710597, + "tokens_seen": 398524416 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044405215646940824, + "loss": 1.5587, + "theoretical_loss": 4.042004191390816, + "tokens_seen": 398589952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440421263791374, + "loss": 1.6585, + "theoretical_loss": 4.041930325618246, + "tokens_seen": 398655488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440320962888666, + "loss": 1.5923, + "theoretical_loss": 4.041856475387061, + "tokens_seen": 398721024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440220661985958, + "loss": 1.6462, + "theoretical_loss": 4.041782640691437, + "tokens_seen": 398786560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044401203610832497, + "loss": 1.6909, + "theoretical_loss": 4.041708821525554, + "tokens_seen": 398852096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004440020060180542, + "loss": 1.6816, + "theoretical_loss": 4.0416350178835945, + "tokens_seen": 398917632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044399197592778333, + "loss": 1.5584, + "theoretical_loss": 4.041561229759743, + "tokens_seen": 398983168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044398194583751256, + "loss": 1.6976, + "theoretical_loss": 4.0414874571481905, + "tokens_seen": 399048704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439719157472417, + "loss": 1.6225, + "theoretical_loss": 4.041413700043128, + "tokens_seen": 399114240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439618856569709, + "loss": 1.6947, + "theoretical_loss": 4.041339958438751, + "tokens_seen": 399179776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439518555667001, + "loss": 1.7462, + "theoretical_loss": 4.041266232329257, + "tokens_seen": 399245312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439418254764293, + "loss": 1.7263, + "theoretical_loss": 4.041192521708849, + "tokens_seen": 399310848 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044393179538615847, + "loss": 1.6412, + "theoretical_loss": 4.041118826571731, + "tokens_seen": 399376384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004439217652958877, + "loss": 1.6371, + "theoretical_loss": 4.041045146912111, + "tokens_seen": 399441920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044391173520561683, + "loss": 1.6173, + "theoretical_loss": 4.040971482724199, + "tokens_seen": 399507456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044390170511534607, + "loss": 1.5546, + "theoretical_loss": 4.040897834002209, + "tokens_seen": 399572992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438916750250752, + "loss": 1.7021, + "theoretical_loss": 4.040824200740358, + "tokens_seen": 399638528 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044388164493480443, + "loss": 1.6162, + "theoretical_loss": 4.040750582932868, + "tokens_seen": 399704064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 151634, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6410927772521973, + "objective/train/theoretical_loss": 4.04067698057396, + "objective/train/tokens_used": 420229600, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438716148445336, + "loss": 1.543, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438615847542628, + "loss": 1.5362, + "theoretical_loss": 4.040603393657862, + "tokens_seen": 399835136 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044385155466399197, + "loss": 1.6097, + "theoretical_loss": 4.040529822178802, + "tokens_seen": 399900672 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044384152457372115, + "loss": 1.5814, + "theoretical_loss": 4.040456266131013, + "tokens_seen": 399966208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044383149448345033, + "loss": 1.6214, + "theoretical_loss": 4.0403827255087315, + "tokens_seen": 400031744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044382146439317957, + "loss": 1.5627, + "theoretical_loss": 4.040309200306195, + "tokens_seen": 400097280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004438114343029087, + "loss": 1.6515, + "theoretical_loss": 4.0402356905176475, + "tokens_seen": 400162816 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044380140421263793, + "loss": 1.6988, + "theoretical_loss": 4.040162196137332, + "tokens_seen": 400228352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437913741223671, + "loss": 1.6753, + "theoretical_loss": 4.040088717159497, + "tokens_seen": 400293888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437813440320963, + "loss": 1.5636, + "theoretical_loss": 4.040015253578394, + "tokens_seen": 400359424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437713139418255, + "loss": 1.6002, + "theoretical_loss": 4.039941805388277, + "tokens_seen": 400424960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044376128385155466, + "loss": 1.5943, + "theoretical_loss": 4.0398683725834035, + "tokens_seen": 400490496 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044375125376128384, + "loss": 1.5266, + "theoretical_loss": 4.039794955158033, + "tokens_seen": 400556032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044374122367101307, + "loss": 1.6067, + "theoretical_loss": 4.039721553106431, + "tokens_seen": 400621568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437311935807422, + "loss": 1.7396, + "theoretical_loss": 4.039648166422863, + "tokens_seen": 400687104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044372116349047143, + "loss": 1.6687, + "theoretical_loss": 4.039574795101597, + "tokens_seen": 400752640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044371113340020056, + "loss": 1.6744, + "theoretical_loss": 4.039501439136908, + "tokens_seen": 400818176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004437011033099298, + "loss": 1.7078, + "theoretical_loss": 4.0394280985230715, + "tokens_seen": 400883712 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044369107321965903, + "loss": 1.665, + "theoretical_loss": 4.039354773254365, + "tokens_seen": 400949248 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044368104312938816, + "loss": 1.6012, + "theoretical_loss": 4.0392814633250715, + "tokens_seen": 401014784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436710130391174, + "loss": 1.6501, + "theoretical_loss": 4.039208168729475, + "tokens_seen": 401080320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436609829488465, + "loss": 1.6699, + "theoretical_loss": 4.039134889461866, + "tokens_seen": 401145856 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044365095285857576, + "loss": 1.4842, + "theoretical_loss": 4.039061625516532, + "tokens_seen": 401211392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044364092276830494, + "loss": 1.666, + "theoretical_loss": 4.038988376887771, + "tokens_seen": 401276928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436308926780341, + "loss": 1.6296, + "theoretical_loss": 4.038915143569876, + "tokens_seen": 401342464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004436208625877633, + "loss": 1.597, + "theoretical_loss": 4.038841925557151, + "tokens_seen": 401408000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044361083249749254, + "loss": 1.5931, + "theoretical_loss": 4.038768722843898, + "tokens_seen": 401473536 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044360080240722166, + "loss": 1.6069, + "theoretical_loss": 4.038695535424422, + "tokens_seen": 401539072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435907723169509, + "loss": 1.6504, + "theoretical_loss": 4.038622363293034, + "tokens_seen": 401604608 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044358074222668, + "loss": 1.5632, + "theoretical_loss": 4.0385492064440465, + "tokens_seen": 401670144 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044357071213640926, + "loss": 1.6131, + "theoretical_loss": 4.038476064871775, + "tokens_seen": 401735680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044356068204613844, + "loss": 1.6558, + "theoretical_loss": 4.0384029385705364, + "tokens_seen": 401801216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435506519558676, + "loss": 1.4403, + "theoretical_loss": 4.038329827534654, + "tokens_seen": 401866752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435406218655968, + "loss": 1.6902, + "theoretical_loss": 4.038256731758453, + "tokens_seen": 401932288 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443530591775326, + "loss": 1.5707, + "theoretical_loss": 4.038183651236259, + "tokens_seen": 401997824 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044352056168505517, + "loss": 1.6891, + "theoretical_loss": 4.038110585962404, + "tokens_seen": 402063360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004435105315947844, + "loss": 1.571, + "theoretical_loss": 4.038037535931221, + "tokens_seen": 402128896 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044350050150451353, + "loss": 1.6168, + "theoretical_loss": 4.037964501137048, + "tokens_seen": 402194432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044349047141424276, + "loss": 1.5711, + "theoretical_loss": 4.037891481574224, + "tokens_seen": 402259968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434804413239719, + "loss": 1.7774, + "theoretical_loss": 4.037818477237091, + "tokens_seen": 402325504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434704112337011, + "loss": 1.6057, + "theoretical_loss": 4.037745488119996, + "tokens_seen": 402391040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434603811434303, + "loss": 1.6008, + "theoretical_loss": 4.037672514217288, + "tokens_seen": 402456576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434503510531595, + "loss": 1.4936, + "theoretical_loss": 4.037599555523318, + "tokens_seen": 402522112 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044344032096288867, + "loss": 1.6237, + "theoretical_loss": 4.037526612032441, + "tokens_seen": 402587648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434302908726179, + "loss": 1.6165, + "theoretical_loss": 4.037453683739015, + "tokens_seen": 402653184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044342026078234703, + "loss": 1.6842, + "theoretical_loss": 4.037380770637402, + "tokens_seen": 402718720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044341023069207627, + "loss": 1.6119, + "theoretical_loss": 4.037307872721964, + "tokens_seen": 402784256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004434002006018054, + "loss": 1.6664, + "theoretical_loss": 4.037234989987069, + "tokens_seen": 402849792 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044339017051153463, + "loss": 1.5818, + "theoretical_loss": 4.0371621224270875, + "tokens_seen": 402915328 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004433801404212638, + "loss": 1.7132, + "theoretical_loss": 4.037089270036391, + "tokens_seen": 402980864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 152845, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.577258586883545, + "objective/train/theoretical_loss": 4.037016432809357, + "objective/train/tokens_used": 423506400, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443370110330993, + "loss": 1.6508, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044336008024072217, + "loss": 1.7134, + "theoretical_loss": 4.036943610740362, + "tokens_seen": 403111936 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044335005015045135, + "loss": 1.6213, + "theoretical_loss": 4.036870803823791, + "tokens_seen": 403177472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044334002006018053, + "loss": 1.6432, + "theoretical_loss": 4.036798012054027, + "tokens_seen": 403243008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044332998996990977, + "loss": 1.5584, + "theoretical_loss": 4.036725235425458, + "tokens_seen": 403308544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004433199598796389, + "loss": 1.5997, + "theoretical_loss": 4.0366524739324765, + "tokens_seen": 403374080 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044330992978936813, + "loss": 1.6677, + "theoretical_loss": 4.036579727569475, + "tokens_seen": 403439616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432998996990973, + "loss": 1.6696, + "theoretical_loss": 4.036506996330849, + "tokens_seen": 403505152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432898696088265, + "loss": 1.631, + "theoretical_loss": 4.036434280211001, + "tokens_seen": 403570688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432798395185557, + "loss": 1.4668, + "theoretical_loss": 4.036361579204334, + "tokens_seen": 403636224 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044326980942828486, + "loss": 1.6014, + "theoretical_loss": 4.036288893305251, + "tokens_seen": 403701760 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044325977933801404, + "loss": 1.6339, + "theoretical_loss": 4.036216222508163, + "tokens_seen": 403767296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044324974924774327, + "loss": 1.6488, + "theoretical_loss": 4.036143566807481, + "tokens_seen": 403832832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004432397191574724, + "loss": 1.6737, + "theoretical_loss": 4.03607092619762, + "tokens_seen": 403898368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044322968906720163, + "loss": 1.7029, + "theoretical_loss": 4.035998300672999, + "tokens_seen": 403963904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044321965897693076, + "loss": 1.5397, + "theoretical_loss": 4.035925690228035, + "tokens_seen": 404029440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044320962888666, + "loss": 1.6178, + "theoretical_loss": 4.035853094857156, + "tokens_seen": 404094976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431995987963892, + "loss": 1.5878, + "theoretical_loss": 4.035780514554787, + "tokens_seen": 404160512 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044318956870611836, + "loss": 1.6734, + "theoretical_loss": 4.0357079493153565, + "tokens_seen": 404226048 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044317953861584754, + "loss": 1.6181, + "theoretical_loss": 4.035635399133298, + "tokens_seen": 404291584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431695085255767, + "loss": 1.6106, + "theoretical_loss": 4.035562864003047, + "tokens_seen": 404357120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431594784353059, + "loss": 1.6442, + "theoretical_loss": 4.035490343919042, + "tokens_seen": 404422656 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044314944834503514, + "loss": 1.5713, + "theoretical_loss": 4.035417838875725, + "tokens_seen": 404488192 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044313941825476427, + "loss": 1.6489, + "theoretical_loss": 4.035345348867539, + "tokens_seen": 404553728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431293881644935, + "loss": 1.6129, + "theoretical_loss": 4.035272873888932, + "tokens_seen": 404619264 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004431193580742227, + "loss": 1.7473, + "theoretical_loss": 4.035200413934355, + "tokens_seen": 404684800 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044310932798395186, + "loss": 1.6721, + "theoretical_loss": 4.03512796899826, + "tokens_seen": 404750336 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044309929789368104, + "loss": 1.5062, + "theoretical_loss": 4.035055539075104, + "tokens_seen": 404815872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004430892678034102, + "loss": 1.5756, + "theoretical_loss": 4.034983124159346, + "tokens_seen": 404881408 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004430792377131394, + "loss": 1.6113, + "theoretical_loss": 4.0349107242454485, + "tokens_seen": 404946944 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044306920762286864, + "loss": 1.6541, + "theoretical_loss": 4.034838339327876, + "tokens_seen": 405012480 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044305917753259777, + "loss": 1.607, + "theoretical_loss": 4.034765969401096, + "tokens_seen": 405078016 + }, + { + "epoch": 0.12, + "learning_rate": 0.000443049147442327, + "loss": 1.6281, + "theoretical_loss": 4.03469361445958, + "tokens_seen": 405143552 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044303911735205613, + "loss": 1.5753, + "theoretical_loss": 4.034621274497802, + "tokens_seen": 405209088 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044302908726178537, + "loss": 1.6643, + "theoretical_loss": 4.0345489495102385, + "tokens_seen": 405274624 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044301905717151455, + "loss": 1.7348, + "theoretical_loss": 4.034476639491368, + "tokens_seen": 405340160 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044300902708124373, + "loss": 1.6747, + "theoretical_loss": 4.034404344435676, + "tokens_seen": 405405696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429989969909729, + "loss": 1.6572, + "theoretical_loss": 4.0343320643376455, + "tokens_seen": 405471232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429889669007021, + "loss": 1.8678, + "theoretical_loss": 4.034259799191765, + "tokens_seen": 405536768 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044297893681043127, + "loss": 1.596, + "theoretical_loss": 4.034187548992527, + "tokens_seen": 405602304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429689067201605, + "loss": 1.6351, + "theoretical_loss": 4.034115313734426, + "tokens_seen": 405667840 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044295887662988963, + "loss": 1.6625, + "theoretical_loss": 4.034043093411958, + "tokens_seen": 405733376 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044294884653961887, + "loss": 1.5875, + "theoretical_loss": 4.033970888019624, + "tokens_seen": 405798912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429388164493481, + "loss": 1.556, + "theoretical_loss": 4.033898697551926, + "tokens_seen": 405864448 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044292878635907723, + "loss": 1.5543, + "theoretical_loss": 4.0338265220033716, + "tokens_seen": 405929984 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044291875626880647, + "loss": 1.7317, + "theoretical_loss": 4.0337543613684685, + "tokens_seen": 405995520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004429087261785356, + "loss": 1.5467, + "theoretical_loss": 4.033682215641729, + "tokens_seen": 406061056 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044289869608826483, + "loss": 1.5183, + "theoretical_loss": 4.033610084817667, + "tokens_seen": 406126592 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442888665997994, + "loss": 1.6488, + "theoretical_loss": 4.0335379688908, + "tokens_seen": 406192128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428786359077232, + "loss": 1.6538, + "theoretical_loss": 4.03346586785565, + "tokens_seen": 406257664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 154148, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5283465385437012, + "objective/train/theoretical_loss": 4.03339378170674, + "objective/train/tokens_used": 426783200, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044286860581745237, + "loss": 1.5543, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044285857572718155, + "loss": 1.5219, + "theoretical_loss": 4.033321710438594, + "tokens_seen": 406388736 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044284854563691073, + "loss": 1.5343, + "theoretical_loss": 4.033249654045743, + "tokens_seen": 406454272 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044283851554663997, + "loss": 1.614, + "theoretical_loss": 4.033177612522721, + "tokens_seen": 406519808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428284854563691, + "loss": 1.4171, + "theoretical_loss": 4.0331055858640585, + "tokens_seen": 406585344 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044281845536609833, + "loss": 1.6251, + "theoretical_loss": 4.033033574064297, + "tokens_seen": 406650880 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004428084252758275, + "loss": 1.6454, + "theoretical_loss": 4.032961577117976, + "tokens_seen": 406716416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427983951855567, + "loss": 1.6517, + "theoretical_loss": 4.032889595019638, + "tokens_seen": 406781952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427883650952859, + "loss": 1.7347, + "theoretical_loss": 4.032817627763832, + "tokens_seen": 406847488 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044277833500501506, + "loss": 1.6972, + "theoretical_loss": 4.0327456753451045, + "tokens_seen": 406913024 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044276830491474424, + "loss": 1.6925, + "theoretical_loss": 4.03267373775801, + "tokens_seen": 406978560 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044275827482447347, + "loss": 1.5923, + "theoretical_loss": 4.032601814997102, + "tokens_seen": 407044096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427482447342026, + "loss": 1.6219, + "theoretical_loss": 4.03252990705694, + "tokens_seen": 407109632 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044273821464393184, + "loss": 1.6719, + "theoretical_loss": 4.032458013932084, + "tokens_seen": 407175168 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044272818455366096, + "loss": 1.5741, + "theoretical_loss": 4.032386135617098, + "tokens_seen": 407240704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427181544633902, + "loss": 1.7665, + "theoretical_loss": 4.032314272106548, + "tokens_seen": 407306240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004427081243731194, + "loss": 1.6166, + "theoretical_loss": 4.0322424233950045, + "tokens_seen": 407371776 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044269809428284856, + "loss": 1.6554, + "theoretical_loss": 4.032170589477039, + "tokens_seen": 407437312 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044268806419257774, + "loss": 1.7086, + "theoretical_loss": 4.032098770347228, + "tokens_seen": 407502848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426780341023069, + "loss": 1.556, + "theoretical_loss": 4.0320269660001475, + "tokens_seen": 407568384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426680040120361, + "loss": 1.7033, + "theoretical_loss": 4.0319551764303805, + "tokens_seen": 407633920 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044265797392176534, + "loss": 1.6245, + "theoretical_loss": 4.03188340163251, + "tokens_seen": 407699456 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044264794383149447, + "loss": 1.7111, + "theoretical_loss": 4.031811641601122, + "tokens_seen": 407764992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426379137412237, + "loss": 1.46, + "theoretical_loss": 4.031739896330807, + "tokens_seen": 407830528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004426278836509529, + "loss": 1.6523, + "theoretical_loss": 4.031668165816156, + "tokens_seen": 407896064 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044261785356068206, + "loss": 1.675, + "theoretical_loss": 4.031596450051766, + "tokens_seen": 407961600 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044260782347041124, + "loss": 1.5503, + "theoretical_loss": 4.031524749032234, + "tokens_seen": 408027136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425977933801404, + "loss": 1.6991, + "theoretical_loss": 4.031453062752161, + "tokens_seen": 408092672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425877632898696, + "loss": 1.5233, + "theoretical_loss": 4.031381391206151, + "tokens_seen": 408158208 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044257773319959884, + "loss": 1.5366, + "theoretical_loss": 4.03130973438881, + "tokens_seen": 408223744 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044256770310932797, + "loss": 1.6818, + "theoretical_loss": 4.031238092294748, + "tokens_seen": 408289280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425576730190572, + "loss": 1.7629, + "theoretical_loss": 4.031166464918576, + "tokens_seen": 408354816 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044254764292878633, + "loss": 1.6162, + "theoretical_loss": 4.0310948522549115, + "tokens_seen": 408420352 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044253761283851557, + "loss": 1.7158, + "theoretical_loss": 4.031023254298371, + "tokens_seen": 408485888 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044252758274824475, + "loss": 1.5321, + "theoretical_loss": 4.030951671043574, + "tokens_seen": 408551424 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044251755265797393, + "loss": 1.676, + "theoretical_loss": 4.030880102485147, + "tokens_seen": 408616960 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004425075225677031, + "loss": 1.6047, + "theoretical_loss": 4.030808548617714, + "tokens_seen": 408682496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424974924774323, + "loss": 1.5731, + "theoretical_loss": 4.030737009435906, + "tokens_seen": 408748032 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044248746238716147, + "loss": 1.601, + "theoretical_loss": 4.030665484934353, + "tokens_seen": 408813568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424774322968907, + "loss": 1.6403, + "theoretical_loss": 4.030593975107692, + "tokens_seen": 408879104 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044246740220661983, + "loss": 1.5449, + "theoretical_loss": 4.03052247995056, + "tokens_seen": 408944640 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044245737211634907, + "loss": 1.5965, + "theoretical_loss": 4.030450999457596, + "tokens_seen": 409010176 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044244734202607825, + "loss": 1.6077, + "theoretical_loss": 4.030379533623445, + "tokens_seen": 409075712 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044243731193580743, + "loss": 1.6126, + "theoretical_loss": 4.030308082442753, + "tokens_seen": 409141248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424272818455366, + "loss": 1.5263, + "theoretical_loss": 4.030236645910169, + "tokens_seen": 409206784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004424172517552658, + "loss": 1.6301, + "theoretical_loss": 4.030165224020345, + "tokens_seen": 409272320 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442407221664995, + "loss": 1.581, + "theoretical_loss": 4.030093816767934, + "tokens_seen": 409337856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423971915747242, + "loss": 1.7332, + "theoretical_loss": 4.030022424147595, + "tokens_seen": 409403392 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044238716148445334, + "loss": 1.5804, + "theoretical_loss": 4.029951046153988, + "tokens_seen": 409468928 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044237713139418257, + "loss": 1.4783, + "theoretical_loss": 4.029879682781776, + "tokens_seen": 409534464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 155224, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.74814772605896, + "objective/train/theoretical_loss": 4.029808334025624, + "objective/train/tokens_used": 430060000, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423671013039117, + "loss": 1.6241, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044235707121364093, + "loss": 1.4811, + "theoretical_loss": 4.0297369998802015, + "tokens_seen": 409665536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423470411233701, + "loss": 1.5279, + "theoretical_loss": 4.0296656803401785, + "tokens_seen": 409731072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423370110330993, + "loss": 1.6891, + "theoretical_loss": 4.029594375400231, + "tokens_seen": 409796608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423269809428285, + "loss": 1.4307, + "theoretical_loss": 4.029523085055036, + "tokens_seen": 409862144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004423169508525577, + "loss": 1.5746, + "theoretical_loss": 4.029451809299273, + "tokens_seen": 409927680 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044230692076228684, + "loss": 1.7121, + "theoretical_loss": 4.029380548127622, + "tokens_seen": 409993216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422968906720161, + "loss": 1.6224, + "theoretical_loss": 4.029309301534772, + "tokens_seen": 410058752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422868605817452, + "loss": 1.5504, + "theoretical_loss": 4.02923806951541, + "tokens_seen": 410124288 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044227683049147444, + "loss": 1.5857, + "theoretical_loss": 4.029166852064225, + "tokens_seen": 410189824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422668004012036, + "loss": 1.7122, + "theoretical_loss": 4.029095649175913, + "tokens_seen": 410255360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422567703109328, + "loss": 1.6665, + "theoretical_loss": 4.02902446084517, + "tokens_seen": 410320896 + }, + { + "epoch": 0.12, + "learning_rate": 0.000442246740220662, + "loss": 1.6499, + "theoretical_loss": 4.028953287066695, + "tokens_seen": 410386432 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044223671013039116, + "loss": 1.6128, + "theoretical_loss": 4.02888212783519, + "tokens_seen": 410451968 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044222668004012034, + "loss": 1.6228, + "theoretical_loss": 4.028810983145361, + "tokens_seen": 410517504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422166499498496, + "loss": 1.7193, + "theoretical_loss": 4.0287398529919125, + "tokens_seen": 410583040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004422066198595787, + "loss": 1.4741, + "theoretical_loss": 4.028668737369558, + "tokens_seen": 410648576 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044219658976930794, + "loss": 1.616, + "theoretical_loss": 4.028597636273009, + "tokens_seen": 410714112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421865596790371, + "loss": 1.5962, + "theoretical_loss": 4.028526549696982, + "tokens_seen": 410779648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421765295887663, + "loss": 1.6076, + "theoretical_loss": 4.0284554776361965, + "tokens_seen": 410845184 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044216649949849554, + "loss": 1.5928, + "theoretical_loss": 4.028384420085372, + "tokens_seen": 410910720 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044215646940822467, + "loss": 1.6126, + "theoretical_loss": 4.028313377039234, + "tokens_seen": 410976256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421464393179539, + "loss": 1.6372, + "theoretical_loss": 4.028242348492509, + "tokens_seen": 411041792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421364092276831, + "loss": 1.6126, + "theoretical_loss": 4.028171334439927, + "tokens_seen": 411107328 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044212637913741226, + "loss": 1.5361, + "theoretical_loss": 4.02810033487622, + "tokens_seen": 411172864 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044211634904714144, + "loss": 1.7319, + "theoretical_loss": 4.028029349796125, + "tokens_seen": 411238400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004421063189568706, + "loss": 1.5651, + "theoretical_loss": 4.027958379194376, + "tokens_seen": 411303936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420962888665998, + "loss": 1.6736, + "theoretical_loss": 4.027887423065717, + "tokens_seen": 411369472 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044208625877632904, + "loss": 1.61, + "theoretical_loss": 4.027816481404892, + "tokens_seen": 411435008 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044207622868605817, + "loss": 1.5426, + "theoretical_loss": 4.0277455542066445, + "tokens_seen": 411500544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420661985957874, + "loss": 1.5939, + "theoretical_loss": 4.027674641465725, + "tokens_seen": 411566080 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044205616850551653, + "loss": 1.7483, + "theoretical_loss": 4.027603743176885, + "tokens_seen": 411631616 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044204613841524577, + "loss": 1.5948, + "theoretical_loss": 4.02753285933488, + "tokens_seen": 411697152 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044203610832497495, + "loss": 1.7274, + "theoretical_loss": 4.027461989934466, + "tokens_seen": 411762688 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044202607823470413, + "loss": 1.6019, + "theoretical_loss": 4.027391134970402, + "tokens_seen": 411828224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420160481444333, + "loss": 1.5097, + "theoretical_loss": 4.027320294437452, + "tokens_seen": 411893760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004420060180541625, + "loss": 1.7557, + "theoretical_loss": 4.027249468330381, + "tokens_seen": 411959296 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044199598796389167, + "loss": 1.7636, + "theoretical_loss": 4.027178656643957, + "tokens_seen": 412024832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004419859578736209, + "loss": 1.6488, + "theoretical_loss": 4.0271078593729515, + "tokens_seen": 412090368 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044197592778335003, + "loss": 1.647, + "theoretical_loss": 4.027037076512137, + "tokens_seen": 412155904 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044196589769307927, + "loss": 1.6178, + "theoretical_loss": 4.026966308056291, + "tokens_seen": 412221440 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044195586760280845, + "loss": 1.5561, + "theoretical_loss": 4.026895554000192, + "tokens_seen": 412286976 + }, + { + "epoch": 0.12, + "learning_rate": 0.00044194583751253763, + "loss": 1.6166, + "theoretical_loss": 4.026824814338621, + "tokens_seen": 412352512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0004419358074222668, + "loss": 1.6703, + "theoretical_loss": 4.026754089066364, + "tokens_seen": 412418048 + }, + { + "epoch": 0.12, + "learning_rate": 0.000441925777331996, + "loss": 1.5375, + "theoretical_loss": 4.026683378178206, + "tokens_seen": 412483584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004419157472417252, + "loss": 1.716, + "theoretical_loss": 4.026612681668939, + "tokens_seen": 412549120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004419057171514544, + "loss": 1.6122, + "theoretical_loss": 4.026541999533355, + "tokens_seen": 412614656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044189568706118354, + "loss": 1.5985, + "theoretical_loss": 4.026471331766249, + "tokens_seen": 412680192 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044188565697091277, + "loss": 1.7587, + "theoretical_loss": 4.026400678362418, + "tokens_seen": 412745728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418756268806419, + "loss": 1.5202, + "theoretical_loss": 4.026330039316665, + "tokens_seen": 412811264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 156351, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5218216180801392, + "objective/train/theoretical_loss": 4.026259414623792, + "objective/train/tokens_used": 433336800, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044186559679037113, + "loss": 1.5363, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418555667001003, + "loss": 1.6968, + "theoretical_loss": 4.026188804278605, + "tokens_seen": 412942336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418455366098295, + "loss": 1.5348, + "theoretical_loss": 4.026118208275914, + "tokens_seen": 413007872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418355065195587, + "loss": 1.7234, + "theoretical_loss": 4.02604762661053, + "tokens_seen": 413073408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418254764292879, + "loss": 1.5441, + "theoretical_loss": 4.025977059277267, + "tokens_seen": 413138944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044181544633901704, + "loss": 1.5776, + "theoretical_loss": 4.025906506270941, + "tokens_seen": 413204480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004418054162487463, + "loss": 1.6123, + "theoretical_loss": 4.0258359675863735, + "tokens_seen": 413270016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417953861584754, + "loss": 1.5827, + "theoretical_loss": 4.025765443218386, + "tokens_seen": 413335552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044178535606820464, + "loss": 1.6984, + "theoretical_loss": 4.025694933161804, + "tokens_seen": 413401088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417753259779338, + "loss": 1.5603, + "theoretical_loss": 4.025624437411455, + "tokens_seen": 413466624 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441765295887663, + "loss": 1.5376, + "theoretical_loss": 4.0255539559621685, + "tokens_seen": 413532160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417552657973922, + "loss": 1.5462, + "theoretical_loss": 4.025483488808779, + "tokens_seen": 413597696 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044174523570712136, + "loss": 1.6021, + "theoretical_loss": 4.025413035946121, + "tokens_seen": 413663232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044173520561685054, + "loss": 1.7036, + "theoretical_loss": 4.025342597369034, + "tokens_seen": 413728768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417251755265798, + "loss": 1.5745, + "theoretical_loss": 4.025272173072359, + "tokens_seen": 413794304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004417151454363089, + "loss": 1.5311, + "theoretical_loss": 4.025201763050939, + "tokens_seen": 413859840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044170511534603814, + "loss": 1.6305, + "theoretical_loss": 4.025131367299621, + "tokens_seen": 413925376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044169508525576727, + "loss": 1.4747, + "theoretical_loss": 4.025060985813255, + "tokens_seen": 413990912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416850551654965, + "loss": 1.5602, + "theoretical_loss": 4.024990618586692, + "tokens_seen": 414056448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416750250752257, + "loss": 1.5073, + "theoretical_loss": 4.024920265614787, + "tokens_seen": 414121984 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044166499498495487, + "loss": 1.6599, + "theoretical_loss": 4.024849926892396, + "tokens_seen": 414187520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044165496489468405, + "loss": 1.6532, + "theoretical_loss": 4.02477960241438, + "tokens_seen": 414253056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416449348044133, + "loss": 1.5718, + "theoretical_loss": 4.024709292175602, + "tokens_seen": 414318592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004416349047141424, + "loss": 1.5941, + "theoretical_loss": 4.024638996170927, + "tokens_seen": 414384128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044162487462387164, + "loss": 1.6552, + "theoretical_loss": 4.024568714395221, + "tokens_seen": 414449664 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044161484453360077, + "loss": 1.5027, + "theoretical_loss": 4.0244984468433564, + "tokens_seen": 414515200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044160481444333, + "loss": 1.6151, + "theoretical_loss": 4.0244281935102055, + "tokens_seen": 414580736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415947843530592, + "loss": 1.6477, + "theoretical_loss": 4.024357954390645, + "tokens_seen": 414646272 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044158475426278837, + "loss": 1.628, + "theoretical_loss": 4.0242877294795525, + "tokens_seen": 414711808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044157472417251755, + "loss": 1.5962, + "theoretical_loss": 4.0242175187718106, + "tokens_seen": 414777344 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044156469408224673, + "loss": 1.586, + "theoretical_loss": 4.024147322262301, + "tokens_seen": 414842880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415546639919759, + "loss": 1.7548, + "theoretical_loss": 4.024077139945912, + "tokens_seen": 414908416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044154463390170515, + "loss": 1.7009, + "theoretical_loss": 4.024006971817532, + "tokens_seen": 414973952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415346038114343, + "loss": 1.5779, + "theoretical_loss": 4.023936817872053, + "tokens_seen": 415039488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004415245737211635, + "loss": 1.6546, + "theoretical_loss": 4.023866678104368, + "tokens_seen": 415105024 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044151454363089264, + "loss": 1.5734, + "theoretical_loss": 4.023796552509376, + "tokens_seen": 415170560 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044150451354062187, + "loss": 1.4088, + "theoretical_loss": 4.023726441081976, + "tokens_seen": 415236096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044149448345035105, + "loss": 1.4403, + "theoretical_loss": 4.02365634381707, + "tokens_seen": 415301632 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044148445336008023, + "loss": 1.6023, + "theoretical_loss": 4.023586260709564, + "tokens_seen": 415367168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414744232698094, + "loss": 1.5806, + "theoretical_loss": 4.023516191754363, + "tokens_seen": 415432704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044146439317953865, + "loss": 1.5968, + "theoretical_loss": 4.02344613694638, + "tokens_seen": 415498240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414543630892678, + "loss": 1.578, + "theoretical_loss": 4.023376096280527, + "tokens_seen": 415563776 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441444332998997, + "loss": 1.4895, + "theoretical_loss": 4.02330606975172, + "tokens_seen": 415629312 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414343029087262, + "loss": 1.7034, + "theoretical_loss": 4.023236057354875, + "tokens_seen": 415694848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414242728184554, + "loss": 1.6042, + "theoretical_loss": 4.023166059084916, + "tokens_seen": 415760384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004414142427281846, + "loss": 1.6903, + "theoretical_loss": 4.023096074936764, + "tokens_seen": 415825920 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044140421263791374, + "loss": 1.4815, + "theoretical_loss": 4.0230261049053455, + "tokens_seen": 415891456 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044139418254764297, + "loss": 1.6155, + "theoretical_loss": 4.022956148985589, + "tokens_seen": 415956992 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413841524573721, + "loss": 1.5934, + "theoretical_loss": 4.0228862071724265, + "tokens_seen": 416022528 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044137412236710133, + "loss": 1.6181, + "theoretical_loss": 4.022816279460791, + "tokens_seen": 416088064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 157543, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4787075519561768, + "objective/train/theoretical_loss": 4.02274636584562, + "objective/train/tokens_used": 436613600, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413640922768305, + "loss": 1.5951, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413540621865597, + "loss": 1.6007, + "theoretical_loss": 4.022676466321853, + "tokens_seen": 416219136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413440320962889, + "loss": 1.7291, + "theoretical_loss": 4.022606580884429, + "tokens_seen": 416284672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413340020060181, + "loss": 1.6926, + "theoretical_loss": 4.0225367095282945, + "tokens_seen": 416350208 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044132397191574724, + "loss": 1.7011, + "theoretical_loss": 4.022466852248396, + "tokens_seen": 416415744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413139418254765, + "loss": 1.5314, + "theoretical_loss": 4.0223970090396834, + "tokens_seen": 416481280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004413039117352056, + "loss": 1.6134, + "theoretical_loss": 4.0223271798971085, + "tokens_seen": 416546816 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044129388164493484, + "loss": 1.557, + "theoretical_loss": 4.0222573648156255, + "tokens_seen": 416612352 + }, + { + "epoch": 0.13, + "learning_rate": 0.000441283851554664, + "loss": 1.7059, + "theoretical_loss": 4.022187563790193, + "tokens_seen": 416677888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412738214643932, + "loss": 1.5696, + "theoretical_loss": 4.02211777681577, + "tokens_seen": 416743424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412637913741224, + "loss": 1.7186, + "theoretical_loss": 4.022048003887319, + "tokens_seen": 416808960 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044125376128385156, + "loss": 1.5311, + "theoretical_loss": 4.021978244999806, + "tokens_seen": 416874496 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044124373119358074, + "loss": 1.7531, + "theoretical_loss": 4.021908500148198, + "tokens_seen": 416940032 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044123370110331, + "loss": 1.6392, + "theoretical_loss": 4.0218387693274655, + "tokens_seen": 417005568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004412236710130391, + "loss": 1.4808, + "theoretical_loss": 4.021769052532582, + "tokens_seen": 417071104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044121364092276834, + "loss": 1.5403, + "theoretical_loss": 4.0216993497585225, + "tokens_seen": 417136640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044120361083249747, + "loss": 1.7052, + "theoretical_loss": 4.021629661000265, + "tokens_seen": 417202176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411935807422267, + "loss": 1.6202, + "theoretical_loss": 4.02155998625279, + "tokens_seen": 417267712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411835506519559, + "loss": 1.6717, + "theoretical_loss": 4.021490325511082, + "tokens_seen": 417333248 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044117352056168507, + "loss": 1.6208, + "theoretical_loss": 4.021420678770125, + "tokens_seen": 417398784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044116349047141425, + "loss": 1.6542, + "theoretical_loss": 4.02135104602491, + "tokens_seen": 417464320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411534603811435, + "loss": 1.5783, + "theoretical_loss": 4.021281427270425, + "tokens_seen": 417529856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411434302908726, + "loss": 1.6607, + "theoretical_loss": 4.021211822501666, + "tokens_seen": 417595392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044113340020060184, + "loss": 1.5687, + "theoretical_loss": 4.021142231713629, + "tokens_seen": 417660928 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044112337011033097, + "loss": 1.6064, + "theoretical_loss": 4.021072654901312, + "tokens_seen": 417726464 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411133400200602, + "loss": 1.4962, + "theoretical_loss": 4.021003092059716, + "tokens_seen": 417792000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004411033099297894, + "loss": 1.5505, + "theoretical_loss": 4.020933543183845, + "tokens_seen": 417857536 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044109327983951857, + "loss": 1.6054, + "theoretical_loss": 4.020864008268706, + "tokens_seen": 417923072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044108324974924775, + "loss": 1.4388, + "theoretical_loss": 4.020794487309308, + "tokens_seen": 417988608 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044107321965897693, + "loss": 1.6044, + "theoretical_loss": 4.020724980300663, + "tokens_seen": 418054144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410631895687061, + "loss": 1.6966, + "theoretical_loss": 4.020655487237785, + "tokens_seen": 418119680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044105315947843535, + "loss": 1.5881, + "theoretical_loss": 4.02058600811569, + "tokens_seen": 418185216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410431293881645, + "loss": 1.6501, + "theoretical_loss": 4.020516542929398, + "tokens_seen": 418250752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004410330992978937, + "loss": 1.4927, + "theoretical_loss": 4.0204470916739306, + "tokens_seen": 418316288 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044102306920762284, + "loss": 1.5905, + "theoretical_loss": 4.020377654344312, + "tokens_seen": 418381824 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044101303911735207, + "loss": 1.7022, + "theoretical_loss": 4.020308230935569, + "tokens_seen": 418447360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044100300902708125, + "loss": 1.7344, + "theoretical_loss": 4.020238821442733, + "tokens_seen": 418512896 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044099297893681043, + "loss": 1.6922, + "theoretical_loss": 4.020169425860834, + "tokens_seen": 418578432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409829488465396, + "loss": 1.6599, + "theoretical_loss": 4.0201000441849075, + "tokens_seen": 418643968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044097291875626885, + "loss": 1.7164, + "theoretical_loss": 4.0200306764099905, + "tokens_seen": 418709504 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440962888665998, + "loss": 1.6511, + "theoretical_loss": 4.019961322531122, + "tokens_seen": 418775040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409528585757272, + "loss": 1.619, + "theoretical_loss": 4.019891982543346, + "tokens_seen": 418840576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044094282848545634, + "loss": 1.6013, + "theoretical_loss": 4.019822656441706, + "tokens_seen": 418906112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409327983951856, + "loss": 1.5531, + "theoretical_loss": 4.01975334422125, + "tokens_seen": 418971648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044092276830491476, + "loss": 1.6157, + "theoretical_loss": 4.019684045877027, + "tokens_seen": 419037184 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044091273821464394, + "loss": 1.4693, + "theoretical_loss": 4.019614761404091, + "tokens_seen": 419102720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004409027081243731, + "loss": 1.6422, + "theoretical_loss": 4.019545490797495, + "tokens_seen": 419168256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408926780341023, + "loss": 1.5894, + "theoretical_loss": 4.0194762340522985, + "tokens_seen": 419233792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408826479438315, + "loss": 1.6012, + "theoretical_loss": 4.01940699116356, + "tokens_seen": 419299328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408726178535607, + "loss": 1.7057, + "theoretical_loss": 4.019337762126343, + "tokens_seen": 419364864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 158751, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7587889432907104, + "objective/train/theoretical_loss": 4.019268546935713, + "objective/train/tokens_used": 439890400, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044086258776328984, + "loss": 1.6253, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408525576730191, + "loss": 1.6119, + "theoretical_loss": 4.0191993455867365, + "tokens_seen": 419495936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408425275827482, + "loss": 1.7444, + "theoretical_loss": 4.019130158074484, + "tokens_seen": 419561472 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044083249749247744, + "loss": 1.6839, + "theoretical_loss": 4.019060984394028, + "tokens_seen": 419627008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408224674022066, + "loss": 1.6942, + "theoretical_loss": 4.018991824540445, + "tokens_seen": 419692544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004408124373119358, + "loss": 1.6578, + "theoretical_loss": 4.018922678508812, + "tokens_seen": 419758080 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440802407221665, + "loss": 1.6427, + "theoretical_loss": 4.018853546294208, + "tokens_seen": 419823616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407923771313942, + "loss": 1.6085, + "theoretical_loss": 4.018784427891718, + "tokens_seen": 419889152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044078234704112335, + "loss": 1.6521, + "theoretical_loss": 4.018715323296425, + "tokens_seen": 419954688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407723169508526, + "loss": 1.5085, + "theoretical_loss": 4.018646232503419, + "tokens_seen": 420020224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407622868605817, + "loss": 1.4569, + "theoretical_loss": 4.0185771555077885, + "tokens_seen": 420085760 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044075225677031094, + "loss": 1.5991, + "theoretical_loss": 4.018508092304628, + "tokens_seen": 420151296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407422266800401, + "loss": 1.6244, + "theoretical_loss": 4.018439042889032, + "tokens_seen": 420216832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407321965897693, + "loss": 1.6655, + "theoretical_loss": 4.018370007256098, + "tokens_seen": 420282368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004407221664994985, + "loss": 1.6802, + "theoretical_loss": 4.018300985400927, + "tokens_seen": 420347904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044071213640922767, + "loss": 1.6954, + "theoretical_loss": 4.018231977318622, + "tokens_seen": 420413440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044070210631895685, + "loss": 1.6707, + "theoretical_loss": 4.018162983004287, + "tokens_seen": 420478976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406920762286861, + "loss": 1.5231, + "theoretical_loss": 4.018094002453031, + "tokens_seen": 420544512 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044068204613841527, + "loss": 1.665, + "theoretical_loss": 4.018025035659965, + "tokens_seen": 420610048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044067201604814445, + "loss": 1.719, + "theoretical_loss": 4.017956082620201, + "tokens_seen": 420675584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406619859578737, + "loss": 1.6547, + "theoretical_loss": 4.0178871433288545, + "tokens_seen": 420741120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406519558676028, + "loss": 1.7135, + "theoretical_loss": 4.017818217781044, + "tokens_seen": 420806656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044064192577733204, + "loss": 1.6177, + "theoretical_loss": 4.017749305971889, + "tokens_seen": 420872192 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044063189568706117, + "loss": 1.5451, + "theoretical_loss": 4.0176804078965125, + "tokens_seen": 420937728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406218655967904, + "loss": 1.6438, + "theoretical_loss": 4.017611523550039, + "tokens_seen": 421003264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004406118355065196, + "loss": 1.7018, + "theoretical_loss": 4.017542652927599, + "tokens_seen": 421068800 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044060180541624877, + "loss": 1.6469, + "theoretical_loss": 4.01747379602432, + "tokens_seen": 421134336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044059177532597795, + "loss": 1.6439, + "theoretical_loss": 4.017404952835336, + "tokens_seen": 421199872 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044058174523570713, + "loss": 1.5624, + "theoretical_loss": 4.017336123355783, + "tokens_seen": 421265408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405717151454363, + "loss": 1.6525, + "theoretical_loss": 4.017267307580798, + "tokens_seen": 421330944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044056168505516555, + "loss": 1.676, + "theoretical_loss": 4.017198505505521, + "tokens_seen": 421396480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405516549648947, + "loss": 1.6619, + "theoretical_loss": 4.017129717125094, + "tokens_seen": 421462016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004405416248746239, + "loss": 1.6964, + "theoretical_loss": 4.017060942434664, + "tokens_seen": 421527552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044053159478435304, + "loss": 1.6006, + "theoretical_loss": 4.016992181429377, + "tokens_seen": 421593088 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044052156469408227, + "loss": 1.7304, + "theoretical_loss": 4.016923434104386, + "tokens_seen": 421658624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044051153460381145, + "loss": 1.6337, + "theoretical_loss": 4.01685470045484, + "tokens_seen": 421724160 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044050150451354063, + "loss": 1.6337, + "theoretical_loss": 4.0167859804758965, + "tokens_seen": 421789696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404914744232698, + "loss": 1.5827, + "theoretical_loss": 4.016717274162712, + "tokens_seen": 421855232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044048144433299905, + "loss": 1.5583, + "theoretical_loss": 4.016648581510448, + "tokens_seen": 421920768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404714142427282, + "loss": 1.481, + "theoretical_loss": 4.016579902514264, + "tokens_seen": 421986304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404613841524574, + "loss": 1.5797, + "theoretical_loss": 4.016511237169329, + "tokens_seen": 422051840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044045135406218654, + "loss": 1.6798, + "theoretical_loss": 4.016442585470807, + "tokens_seen": 422117376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404413239719158, + "loss": 1.5768, + "theoretical_loss": 4.01637394741387, + "tokens_seen": 422182912 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044043129388164496, + "loss": 1.6905, + "theoretical_loss": 4.016305322993689, + "tokens_seen": 422248448 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044042126379137414, + "loss": 1.5408, + "theoretical_loss": 4.016236712205439, + "tokens_seen": 422313984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404112337011033, + "loss": 1.6435, + "theoretical_loss": 4.016168115044298, + "tokens_seen": 422379520 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004404012036108325, + "loss": 1.6839, + "theoretical_loss": 4.016099531505445, + "tokens_seen": 422445056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403911735205617, + "loss": 1.5698, + "theoretical_loss": 4.016030961584063, + "tokens_seen": 422510592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403811434302909, + "loss": 1.5278, + "theoretical_loss": 4.015962405275336, + "tokens_seen": 422576128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044037111334002004, + "loss": 1.607, + "theoretical_loss": 4.015893862574451, + "tokens_seen": 422641664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 159824, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5696735382080078, + "objective/train/theoretical_loss": 4.015825333476598, + "objective/train/tokens_used": 443167200, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403610832497493, + "loss": 1.6174, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403510531594784, + "loss": 1.6986, + "theoretical_loss": 4.0157568179769685, + "tokens_seen": 422772736 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044034102306920764, + "loss": 1.6424, + "theoretical_loss": 4.015688316070757, + "tokens_seen": 422838272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403309929789368, + "loss": 1.6942, + "theoretical_loss": 4.01561982775316, + "tokens_seen": 422903808 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440320962888666, + "loss": 1.614, + "theoretical_loss": 4.015551353019378, + "tokens_seen": 422969344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403109327983952, + "loss": 1.6169, + "theoretical_loss": 4.015482891864612, + "tokens_seen": 423034880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004403009027081244, + "loss": 1.6523, + "theoretical_loss": 4.015414444284065, + "tokens_seen": 423100416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044029087261785355, + "loss": 1.5805, + "theoretical_loss": 4.015346010272946, + "tokens_seen": 423165952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402808425275828, + "loss": 1.6085, + "theoretical_loss": 4.0152775898264625, + "tokens_seen": 423231488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402708124373119, + "loss": 1.7381, + "theoretical_loss": 4.0152091829398255, + "tokens_seen": 423297024 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044026078234704114, + "loss": 1.5938, + "theoretical_loss": 4.0151407896082505, + "tokens_seen": 423362560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402507522567703, + "loss": 1.7033, + "theoretical_loss": 4.015072409826953, + "tokens_seen": 423428096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402407221664995, + "loss": 1.5266, + "theoretical_loss": 4.015004043591151, + "tokens_seen": 423493632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402306920762287, + "loss": 1.6337, + "theoretical_loss": 4.014935690896067, + "tokens_seen": 423559168 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044022066198595787, + "loss": 1.6738, + "theoretical_loss": 4.014867351736924, + "tokens_seen": 423624704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044021063189568705, + "loss": 1.6667, + "theoretical_loss": 4.014799026108948, + "tokens_seen": 423690240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004402006018054163, + "loss": 1.5954, + "theoretical_loss": 4.014730714007367, + "tokens_seen": 423755776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401905717151454, + "loss": 1.5705, + "theoretical_loss": 4.014662415427413, + "tokens_seen": 423821312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044018054162487465, + "loss": 1.5473, + "theoretical_loss": 4.014594130364319, + "tokens_seen": 423886848 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044017051153460383, + "loss": 1.5152, + "theoretical_loss": 4.01452585881332, + "tokens_seen": 423952384 + }, + { + "epoch": 0.13, + "learning_rate": 0.000440160481444333, + "loss": 1.6721, + "theoretical_loss": 4.014457600769655, + "tokens_seen": 424017920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401504513540622, + "loss": 1.7208, + "theoretical_loss": 4.0143893562285635, + "tokens_seen": 424083456 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044014042126379137, + "loss": 1.5955, + "theoretical_loss": 4.014321125185289, + "tokens_seen": 424148992 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044013039117352055, + "loss": 1.6279, + "theoretical_loss": 4.014252907635077, + "tokens_seen": 424214528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401203610832498, + "loss": 1.6753, + "theoretical_loss": 4.014184703573176, + "tokens_seen": 424280064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004401103309929789, + "loss": 1.7229, + "theoretical_loss": 4.014116512994835, + "tokens_seen": 424345600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044010030090270815, + "loss": 1.7664, + "theoretical_loss": 4.014048335895307, + "tokens_seen": 424411136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400902708124373, + "loss": 1.697, + "theoretical_loss": 4.0139801722698465, + "tokens_seen": 424476672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400802407221665, + "loss": 1.7094, + "theoretical_loss": 4.013912022113711, + "tokens_seen": 424542208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400702106318957, + "loss": 1.6549, + "theoretical_loss": 4.013843885422162, + "tokens_seen": 424607744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400601805416249, + "loss": 1.5949, + "theoretical_loss": 4.01377576219046, + "tokens_seen": 424673280 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044005015045135406, + "loss": 1.6337, + "theoretical_loss": 4.01370765241387, + "tokens_seen": 424738816 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044004012036108324, + "loss": 1.5525, + "theoretical_loss": 4.013639556087659, + "tokens_seen": 424804352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400300902708124, + "loss": 1.6801, + "theoretical_loss": 4.0135714732070955, + "tokens_seen": 424869888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044002006018054165, + "loss": 1.686, + "theoretical_loss": 4.013503403767453, + "tokens_seen": 424935424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004400100300902708, + "loss": 1.6065, + "theoretical_loss": 4.013435347764005, + "tokens_seen": 425000960 + }, + { + "epoch": 0.13, + "learning_rate": 0.00044, + "loss": 1.5081, + "theoretical_loss": 4.013367305192027, + "tokens_seen": 425066496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399899699097292, + "loss": 1.5848, + "theoretical_loss": 4.0132992760468, + "tokens_seen": 425132032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399799398194584, + "loss": 1.5535, + "theoretical_loss": 4.013231260323603, + "tokens_seen": 425197568 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043996990972918756, + "loss": 1.458, + "theoretical_loss": 4.0131632580177214, + "tokens_seen": 425263104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043995987963891674, + "loss": 1.7, + "theoretical_loss": 4.01309526912444, + "tokens_seen": 425328640 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399498495486459, + "loss": 1.7383, + "theoretical_loss": 4.013027293639049, + "tokens_seen": 425394176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043993981945837516, + "loss": 1.5979, + "theoretical_loss": 4.0129593315568375, + "tokens_seen": 425459712 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043992978936810434, + "loss": 1.669, + "theoretical_loss": 4.0128913828731, + "tokens_seen": 425525248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399197592778335, + "loss": 1.6448, + "theoretical_loss": 4.012823447583132, + "tokens_seen": 425590784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004399097291875627, + "loss": 1.7558, + "theoretical_loss": 4.012755525682231, + "tokens_seen": 425656320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398996990972919, + "loss": 1.6267, + "theoretical_loss": 4.012687617165698, + "tokens_seen": 425721856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398896690070211, + "loss": 1.6315, + "theoretical_loss": 4.012619722028835, + "tokens_seen": 425787392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043987963891675024, + "loss": 1.6059, + "theoretical_loss": 4.012551840266948, + "tokens_seen": 425852928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398696088264795, + "loss": 1.4826, + "theoretical_loss": 4.0124839718753424, + "tokens_seen": 425918464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 161088, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4721550941467285, + "objective/train/theoretical_loss": 4.01241611684933, + "objective/train/tokens_used": 446444000, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398595787362086, + "loss": 1.5725, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043984954864593784, + "loss": 1.6575, + "theoretical_loss": 4.012348275184223, + "tokens_seen": 426049536 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439839518555667, + "loss": 1.5322, + "theoretical_loss": 4.012280446875336, + "tokens_seen": 426115072 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398294884653962, + "loss": 1.6756, + "theoretical_loss": 4.012212631917986, + "tokens_seen": 426180608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398194583751254, + "loss": 1.6426, + "theoretical_loss": 4.012144830307491, + "tokens_seen": 426246144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004398094282848546, + "loss": 1.5476, + "theoretical_loss": 4.012077042039174, + "tokens_seen": 426311680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043979939819458375, + "loss": 1.643, + "theoretical_loss": 4.012009267108358, + "tokens_seen": 426377216 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439789368104313, + "loss": 1.6606, + "theoretical_loss": 4.01194150551037, + "tokens_seen": 426442752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397793380140421, + "loss": 1.6241, + "theoretical_loss": 4.011873757240539, + "tokens_seen": 426508288 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043976930792377134, + "loss": 1.7286, + "theoretical_loss": 4.011806022294196, + "tokens_seen": 426573824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397592778335005, + "loss": 1.6882, + "theoretical_loss": 4.011738300666674, + "tokens_seen": 426639360 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397492477432297, + "loss": 1.6655, + "theoretical_loss": 4.011670592353309, + "tokens_seen": 426704896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397392176529589, + "loss": 1.5453, + "theoretical_loss": 4.011602897349439, + "tokens_seen": 426770432 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043972918756268807, + "loss": 1.6268, + "theoretical_loss": 4.011535215650405, + "tokens_seen": 426835968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043971915747241725, + "loss": 1.54, + "theoretical_loss": 4.0114675472515495, + "tokens_seen": 426901504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004397091273821465, + "loss": 1.6337, + "theoretical_loss": 4.011399892148218, + "tokens_seen": 426967040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396990972918756, + "loss": 1.5858, + "theoretical_loss": 4.011332250335758, + "tokens_seen": 427032576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043968906720160485, + "loss": 1.6959, + "theoretical_loss": 4.011264621809518, + "tokens_seen": 427098112 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043967903711133403, + "loss": 1.6281, + "theoretical_loss": 4.011197006564853, + "tokens_seen": 427163648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396690070210632, + "loss": 1.6285, + "theoretical_loss": 4.011129404597114, + "tokens_seen": 427229184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396589769307924, + "loss": 1.74, + "theoretical_loss": 4.011061815901661, + "tokens_seen": 427294720 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043964894684052157, + "loss": 1.6239, + "theoretical_loss": 4.010994240473853, + "tokens_seen": 427360256 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043963891675025075, + "loss": 1.6543, + "theoretical_loss": 4.0109266783090485, + "tokens_seen": 427425792 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043962888665998, + "loss": 1.6412, + "theoretical_loss": 4.010859129402615, + "tokens_seen": 427491328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004396188565697091, + "loss": 1.5907, + "theoretical_loss": 4.010791593749916, + "tokens_seen": 427556864 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043960882647943835, + "loss": 1.6851, + "theoretical_loss": 4.010724071346322, + "tokens_seen": 427622400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395987963891675, + "loss": 1.6962, + "theoretical_loss": 4.0106565621872035, + "tokens_seen": 427687936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395887662988967, + "loss": 1.5947, + "theoretical_loss": 4.010589066267933, + "tokens_seen": 427753472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395787362086259, + "loss": 1.6973, + "theoretical_loss": 4.010521583583886, + "tokens_seen": 427819008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395687061183551, + "loss": 1.6504, + "theoretical_loss": 4.010454114130441, + "tokens_seen": 427884544 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043955867602808426, + "loss": 1.6148, + "theoretical_loss": 4.010386657902979, + "tokens_seen": 427950080 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043954864593781344, + "loss": 1.5815, + "theoretical_loss": 4.01031921489688, + "tokens_seen": 428015616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395386158475426, + "loss": 1.6102, + "theoretical_loss": 4.01025178510753, + "tokens_seen": 428081152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043952858575727185, + "loss": 1.5027, + "theoretical_loss": 4.010184368530318, + "tokens_seen": 428146688 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439518555667001, + "loss": 1.5646, + "theoretical_loss": 4.0101169651606305, + "tokens_seen": 428212224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004395085255767302, + "loss": 1.6898, + "theoretical_loss": 4.010049574993861, + "tokens_seen": 428277760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394984954864594, + "loss": 1.6339, + "theoretical_loss": 4.009982198025404, + "tokens_seen": 428343296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394884653961886, + "loss": 1.6373, + "theoretical_loss": 4.0099148342506545, + "tokens_seen": 428408832 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043947843530591776, + "loss": 1.6849, + "theoretical_loss": 4.009847483665011, + "tokens_seen": 428474368 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043946840521564694, + "loss": 1.7745, + "theoretical_loss": 4.009780146263877, + "tokens_seen": 428539904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394583751253761, + "loss": 1.5708, + "theoretical_loss": 4.009712822042652, + "tokens_seen": 428605440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043944834503510536, + "loss": 1.6202, + "theoretical_loss": 4.009645510996744, + "tokens_seen": 428670976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394383149448345, + "loss": 1.5119, + "theoretical_loss": 4.0095782131215625, + "tokens_seen": 428736512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394282848545637, + "loss": 1.5719, + "theoretical_loss": 4.0095109284125146, + "tokens_seen": 428802048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043941825476429285, + "loss": 1.69, + "theoretical_loss": 4.009443656865013, + "tokens_seen": 428867584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004394082246740221, + "loss": 1.5679, + "theoretical_loss": 4.009376398474475, + "tokens_seen": 428933120 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043939819458375126, + "loss": 1.5708, + "theoretical_loss": 4.009309153236316, + "tokens_seen": 428998656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043938816449348044, + "loss": 1.5969, + "theoretical_loss": 4.009241921145956, + "tokens_seen": 429064192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393781344032096, + "loss": 1.6527, + "theoretical_loss": 4.009174702198816, + "tokens_seen": 429129728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393681043129388, + "loss": 1.5353, + "theoretical_loss": 4.009107496390321, + "tokens_seen": 429195264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 162190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.668204426765442, + "objective/train/theoretical_loss": 4.009040303715897, + "objective/train/tokens_used": 449720800, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439358074222668, + "loss": 1.6871, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393480441323972, + "loss": 1.5753, + "theoretical_loss": 4.008973124170971, + "tokens_seen": 429326336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043933801404212635, + "loss": 1.7419, + "theoretical_loss": 4.008905957750976, + "tokens_seen": 429391872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004393279839518556, + "loss": 1.6761, + "theoretical_loss": 4.0088388044513446, + "tokens_seen": 429457408 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043931795386158477, + "loss": 1.5306, + "theoretical_loss": 4.0087716642675115, + "tokens_seen": 429522944 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043930792377131395, + "loss": 1.5777, + "theoretical_loss": 4.008704537194916, + "tokens_seen": 429588480 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043929789368104313, + "loss": 1.6277, + "theoretical_loss": 4.008637423228996, + "tokens_seen": 429654016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392878635907723, + "loss": 1.694, + "theoretical_loss": 4.008570322365195, + "tokens_seen": 429719552 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392778335005015, + "loss": 1.627, + "theoretical_loss": 4.008503234598958, + "tokens_seen": 429785088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392678034102307, + "loss": 1.6735, + "theoretical_loss": 4.008436159925731, + "tokens_seen": 429850624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043925777331995985, + "loss": 1.6635, + "theoretical_loss": 4.008369098340963, + "tokens_seen": 429916160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392477432296891, + "loss": 1.5266, + "theoretical_loss": 4.008302049840106, + "tokens_seen": 429981696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392377131394182, + "loss": 1.6393, + "theoretical_loss": 4.008235014418614, + "tokens_seen": 430047232 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043922768304914745, + "loss": 1.6913, + "theoretical_loss": 4.008167992071941, + "tokens_seen": 430112768 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043921765295887663, + "loss": 1.6329, + "theoretical_loss": 4.008100982795549, + "tokens_seen": 430178304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004392076228686058, + "loss": 1.5199, + "theoretical_loss": 4.008033986584895, + "tokens_seen": 430243840 + }, + { + "epoch": 0.13, + "learning_rate": 0.000439197592778335, + "loss": 1.6937, + "theoretical_loss": 4.007967003435443, + "tokens_seen": 430309376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043918756268806423, + "loss": 1.4624, + "theoretical_loss": 4.007900033342658, + "tokens_seen": 430374912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391775325977934, + "loss": 1.6202, + "theoretical_loss": 4.007833076302006, + "tokens_seen": 430440448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391675025075226, + "loss": 1.5801, + "theoretical_loss": 4.00776613230896, + "tokens_seen": 430505984 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043915747241725177, + "loss": 1.5377, + "theoretical_loss": 4.007699201358989, + "tokens_seen": 430571520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043914744232698095, + "loss": 1.7104, + "theoretical_loss": 4.007632283447567, + "tokens_seen": 430637056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391374122367102, + "loss": 1.5188, + "theoretical_loss": 4.007565378570171, + "tokens_seen": 430702592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391273821464393, + "loss": 1.6476, + "theoretical_loss": 4.00749848672228, + "tokens_seen": 430768128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043911735205616855, + "loss": 1.5484, + "theoretical_loss": 4.007431607899375, + "tokens_seen": 430833664 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004391073219658977, + "loss": 1.5016, + "theoretical_loss": 4.007364742096938, + "tokens_seen": 430899200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390972918756269, + "loss": 1.6125, + "theoretical_loss": 4.007297889310455, + "tokens_seen": 430964736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390872617853561, + "loss": 1.5894, + "theoretical_loss": 4.007231049535413, + "tokens_seen": 431030272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390772316950853, + "loss": 1.6049, + "theoretical_loss": 4.007164222767304, + "tokens_seen": 431095808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043906720160481446, + "loss": 1.6867, + "theoretical_loss": 4.0070974090016165, + "tokens_seen": 431161344 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043905717151454364, + "loss": 1.635, + "theoretical_loss": 4.007030608233848, + "tokens_seen": 431226880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390471414242728, + "loss": 1.7298, + "theoretical_loss": 4.006963820459494, + "tokens_seen": 431292416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043903711133400205, + "loss": 1.5794, + "theoretical_loss": 4.006897045674053, + "tokens_seen": 431357952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390270812437312, + "loss": 1.6086, + "theoretical_loss": 4.006830283873027, + "tokens_seen": 431423488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390170511534604, + "loss": 1.5528, + "theoretical_loss": 4.006763535051918, + "tokens_seen": 431489024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004390070210631896, + "loss": 1.5811, + "theoretical_loss": 4.006696799206233, + "tokens_seen": 431554560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389969909729188, + "loss": 1.6103, + "theoretical_loss": 4.006630076331478, + "tokens_seen": 431620096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043898696088264796, + "loss": 1.6231, + "theoretical_loss": 4.0065633664231655, + "tokens_seen": 431685632 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043897693079237714, + "loss": 1.6127, + "theoretical_loss": 4.006496669476805, + "tokens_seen": 431751168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389669007021063, + "loss": 1.6134, + "theoretical_loss": 4.006429985487913, + "tokens_seen": 431816704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043895687061183556, + "loss": 1.5712, + "theoretical_loss": 4.006363314452007, + "tokens_seen": 431882240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389468405215647, + "loss": 1.6351, + "theoretical_loss": 4.006296656364603, + "tokens_seen": 431947776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389368104312939, + "loss": 1.6703, + "theoretical_loss": 4.006230011221225, + "tokens_seen": 432013312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043892678034102305, + "loss": 1.7279, + "theoretical_loss": 4.006163379017394, + "tokens_seen": 432078848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004389167502507523, + "loss": 1.6184, + "theoretical_loss": 4.0060967597486385, + "tokens_seen": 432144384 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043890672016048146, + "loss": 1.6077, + "theoretical_loss": 4.006030153410483, + "tokens_seen": 432209920 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043889669007021064, + "loss": 1.6711, + "theoretical_loss": 4.005963559998461, + "tokens_seen": 432275456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388866599799398, + "loss": 1.5883, + "theoretical_loss": 4.005896979508103, + "tokens_seen": 432340992 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438876629889669, + "loss": 1.6648, + "theoretical_loss": 4.005830411934944, + "tokens_seen": 432406528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388665997993982, + "loss": 1.623, + "theoretical_loss": 4.005763857274521, + "tokens_seen": 432472064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 163281, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.586388349533081, + "objective/train/theoretical_loss": 4.005697315522371, + "objective/train/tokens_used": 452997600, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388565697091274, + "loss": 1.6405, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043884653961885655, + "loss": 1.6385, + "theoretical_loss": 4.005630786674039, + "tokens_seen": 432603136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004388365095285858, + "loss": 1.6585, + "theoretical_loss": 4.005564270725065, + "tokens_seen": 432668672 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043882647943831497, + "loss": 1.6611, + "theoretical_loss": 4.005497767670996, + "tokens_seen": 432734208 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043881644934804415, + "loss": 1.4957, + "theoretical_loss": 4.0054312775073795, + "tokens_seen": 432799744 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043880641925777333, + "loss": 1.5514, + "theoretical_loss": 4.005364800229766, + "tokens_seen": 432865280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387963891675025, + "loss": 1.6706, + "theoretical_loss": 4.005298335833708, + "tokens_seen": 432930816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387863590772317, + "loss": 1.5987, + "theoretical_loss": 4.005231884314759, + "tokens_seen": 432996352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387763289869609, + "loss": 1.5726, + "theoretical_loss": 4.005165445668477, + "tokens_seen": 433061888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043876629889669005, + "loss": 1.5511, + "theoretical_loss": 4.0050990198904195, + "tokens_seen": 433127424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387562688064193, + "loss": 1.6576, + "theoretical_loss": 4.005032606976149, + "tokens_seen": 433192960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387462387161484, + "loss": 1.6147, + "theoretical_loss": 4.004966206921227, + "tokens_seen": 433258496 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043873620862587765, + "loss": 1.5754, + "theoretical_loss": 4.004899819721221, + "tokens_seen": 433324032 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043872617853560683, + "loss": 1.6116, + "theoretical_loss": 4.004833445371697, + "tokens_seen": 433389568 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438716148445336, + "loss": 1.7138, + "theoretical_loss": 4.004767083868226, + "tokens_seen": 433455104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004387061183550652, + "loss": 1.7271, + "theoretical_loss": 4.004700735206381, + "tokens_seen": 433520640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043869608826479443, + "loss": 1.6696, + "theoretical_loss": 4.004634399381733, + "tokens_seen": 433586176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043868605817452356, + "loss": 1.7346, + "theoretical_loss": 4.0045680763898615, + "tokens_seen": 433651712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386760280842528, + "loss": 1.5532, + "theoretical_loss": 4.004501766226344, + "tokens_seen": 433717248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386659979939819, + "loss": 1.5621, + "theoretical_loss": 4.004435468886761, + "tokens_seen": 433782784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043865596790371115, + "loss": 1.7068, + "theoretical_loss": 4.0043691843666975, + "tokens_seen": 433848320 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043864593781344033, + "loss": 1.5181, + "theoretical_loss": 4.004302912661736, + "tokens_seen": 433913856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386359077231695, + "loss": 1.6125, + "theoretical_loss": 4.0042366537674665, + "tokens_seen": 433979392 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386258776328987, + "loss": 1.6021, + "theoretical_loss": 4.004170407679476, + "tokens_seen": 434044928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004386158475426279, + "loss": 1.6216, + "theoretical_loss": 4.004104174393359, + "tokens_seen": 434110464 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043860581745235706, + "loss": 1.6063, + "theoretical_loss": 4.004037953904708, + "tokens_seen": 434176000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385957873620863, + "loss": 1.5818, + "theoretical_loss": 4.003971746209119, + "tokens_seen": 434241536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385857572718154, + "loss": 1.7074, + "theoretical_loss": 4.003905551302191, + "tokens_seen": 434307072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043857572718154466, + "loss": 1.516, + "theoretical_loss": 4.0038393691795235, + "tokens_seen": 434372608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385656970912738, + "loss": 1.6694, + "theoretical_loss": 4.003773199836721, + "tokens_seen": 434438144 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438555667001003, + "loss": 1.5111, + "theoretical_loss": 4.003707043269387, + "tokens_seen": 434503680 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385456369107322, + "loss": 1.7503, + "theoretical_loss": 4.003640899473128, + "tokens_seen": 434569216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385356068204614, + "loss": 1.6426, + "theoretical_loss": 4.0035747684435545, + "tokens_seen": 434634752 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043852557673019056, + "loss": 1.6749, + "theoretical_loss": 4.003508650176278, + "tokens_seen": 434700288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385155466399198, + "loss": 1.684, + "theoretical_loss": 4.003442544666911, + "tokens_seen": 434765824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004385055165496489, + "loss": 1.7091, + "theoretical_loss": 4.003376451911071, + "tokens_seen": 434831360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043849548645937816, + "loss": 1.6333, + "theoretical_loss": 4.003310371904373, + "tokens_seen": 434896896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384854563691073, + "loss": 1.6396, + "theoretical_loss": 4.00324430464244, + "tokens_seen": 434962432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384754262788365, + "loss": 1.6451, + "theoretical_loss": 4.003178250120892, + "tokens_seen": 435027968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384653961885657, + "loss": 1.7371, + "theoretical_loss": 4.003112208335355, + "tokens_seen": 435093504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384553660982949, + "loss": 1.5608, + "theoretical_loss": 4.003046179281455, + "tokens_seen": 435159040 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043844533600802407, + "loss": 1.7006, + "theoretical_loss": 4.00298016295482, + "tokens_seen": 435224576 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043843530591775325, + "loss": 1.4524, + "theoretical_loss": 4.002914159351082, + "tokens_seen": 435290112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004384252758274825, + "loss": 1.6164, + "theoretical_loss": 4.0028481684658725, + "tokens_seen": 435355648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043841524573721166, + "loss": 1.6334, + "theoretical_loss": 4.002782190294829, + "tokens_seen": 435421184 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043840521564694084, + "loss": 1.7152, + "theoretical_loss": 4.002716224833588, + "tokens_seen": 435486720 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043839518555667, + "loss": 1.4992, + "theoretical_loss": 4.002650272077787, + "tokens_seen": 435552256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383851554663992, + "loss": 1.5719, + "theoretical_loss": 4.00258433202307, + "tokens_seen": 435617792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383751253761284, + "loss": 1.6351, + "theoretical_loss": 4.0025184046650795, + "tokens_seen": 435683328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383650952858576, + "loss": 1.7127, + "theoretical_loss": 4.002452489999463, + "tokens_seen": 435748864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 164592, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5357398986816406, + "objective/train/theoretical_loss": 4.002386588021867, + "objective/train/tokens_used": 456274400, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043835506519558675, + "loss": 1.6554, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438345035105316, + "loss": 1.5702, + "theoretical_loss": 4.002320698727942, + "tokens_seen": 435879936 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043833500501504517, + "loss": 1.6068, + "theoretical_loss": 4.002254822113341, + "tokens_seen": 435945472 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043832497492477435, + "loss": 1.6604, + "theoretical_loss": 4.0021889581737184, + "tokens_seen": 436011008 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043831494483450353, + "loss": 1.6968, + "theoretical_loss": 4.00212310690473, + "tokens_seen": 436076544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004383049147442327, + "loss": 1.5492, + "theoretical_loss": 4.0020572683020355, + "tokens_seen": 436142080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382948846539619, + "loss": 1.6507, + "theoretical_loss": 4.0019914423612954, + "tokens_seen": 436207616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382848545636911, + "loss": 1.5767, + "theoretical_loss": 4.0019256290781735, + "tokens_seen": 436273152 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043827482447342025, + "loss": 1.6884, + "theoretical_loss": 4.001859828448334, + "tokens_seen": 436338688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382647943831495, + "loss": 1.6091, + "theoretical_loss": 4.001794040467446, + "tokens_seen": 436404224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382547642928786, + "loss": 1.6119, + "theoretical_loss": 4.0017282651311765, + "tokens_seen": 436469760 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043824473420260785, + "loss": 1.7222, + "theoretical_loss": 4.001662502435199, + "tokens_seen": 436535296 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043823470411233703, + "loss": 1.4742, + "theoretical_loss": 4.001596752375186, + "tokens_seen": 436600832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382246740220662, + "loss": 1.6069, + "theoretical_loss": 4.001531014946815, + "tokens_seen": 436666368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004382146439317954, + "loss": 1.6728, + "theoretical_loss": 4.001465290145763, + "tokens_seen": 436731904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043820461384152463, + "loss": 1.6868, + "theoretical_loss": 4.00139957796771, + "tokens_seen": 436797440 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043819458375125376, + "loss": 1.7288, + "theoretical_loss": 4.001333878408339, + "tokens_seen": 436862976 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438184553660983, + "loss": 1.5364, + "theoretical_loss": 4.001268191463334, + "tokens_seen": 436928512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381745235707121, + "loss": 1.5395, + "theoretical_loss": 4.00120251712838, + "tokens_seen": 436994048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043816449348044135, + "loss": 1.6759, + "theoretical_loss": 4.001136855399168, + "tokens_seen": 437059584 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043815446339017053, + "loss": 1.679, + "theoretical_loss": 4.001071206271388, + "tokens_seen": 437125120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381444332998997, + "loss": 1.6249, + "theoretical_loss": 4.001005569740732, + "tokens_seen": 437190656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381344032096289, + "loss": 1.4855, + "theoretical_loss": 4.000939945802896, + "tokens_seen": 437256192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381243731193581, + "loss": 1.5538, + "theoretical_loss": 4.000874334453577, + "tokens_seen": 437321728 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043811434302908726, + "loss": 1.5816, + "theoretical_loss": 4.000808735688474, + "tokens_seen": 437387264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004381043129388165, + "loss": 1.6543, + "theoretical_loss": 4.000743149503288, + "tokens_seen": 437452800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380942828485456, + "loss": 1.6258, + "theoretical_loss": 4.000677575893723, + "tokens_seen": 437518336 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043808425275827486, + "loss": 1.4583, + "theoretical_loss": 4.000612014855484, + "tokens_seen": 437583872 + }, + { + "epoch": 0.13, + "learning_rate": 0.000438074222668004, + "loss": 1.7337, + "theoretical_loss": 4.00054646638428, + "tokens_seen": 437649408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380641925777332, + "loss": 1.5655, + "theoretical_loss": 4.000480930475819, + "tokens_seen": 437714944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380541624874624, + "loss": 1.6682, + "theoretical_loss": 4.000415407125814, + "tokens_seen": 437780480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380441323971916, + "loss": 1.6829, + "theoretical_loss": 4.000349896329979, + "tokens_seen": 437846016 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043803410230692076, + "loss": 1.6725, + "theoretical_loss": 4.0002843980840295, + "tokens_seen": 437911552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043802407221665, + "loss": 1.6039, + "theoretical_loss": 4.0002189123836835, + "tokens_seen": 437977088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004380140421263791, + "loss": 1.612, + "theoretical_loss": 4.000153439224662, + "tokens_seen": 438042624 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043800401203610836, + "loss": 1.4694, + "theoretical_loss": 4.0000879786026875, + "tokens_seen": 438108160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379939819458375, + "loss": 1.6377, + "theoretical_loss": 4.000022530513484, + "tokens_seen": 438173696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379839518555667, + "loss": 1.6564, + "theoretical_loss": 3.999957094952779, + "tokens_seen": 438239232 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379739217652959, + "loss": 1.6589, + "theoretical_loss": 3.9998916719163002, + "tokens_seen": 438304768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379638916750251, + "loss": 1.6301, + "theoretical_loss": 3.999826261399778, + "tokens_seen": 438370304 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043795386158475427, + "loss": 1.5243, + "theoretical_loss": 3.9997608633989463, + "tokens_seen": 438435840 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043794383149448345, + "loss": 1.6485, + "theoretical_loss": 3.9996954779095395, + "tokens_seen": 438501376 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043793380140421263, + "loss": 1.7028, + "theoretical_loss": 3.999630104927295, + "tokens_seen": 438566912 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043792377131394186, + "loss": 1.6941, + "theoretical_loss": 3.999564744447952, + "tokens_seen": 438632448 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437913741223671, + "loss": 1.7332, + "theoretical_loss": 3.9994993964672516, + "tokens_seen": 438697984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004379037111334002, + "loss": 1.6229, + "theoretical_loss": 3.9994340609809367, + "tokens_seen": 438763520 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043789368104312935, + "loss": 1.6692, + "theoretical_loss": 3.999368737984754, + "tokens_seen": 438829056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378836509528586, + "loss": 1.5752, + "theoretical_loss": 3.999303427474449, + "tokens_seen": 438894592 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043787362086258777, + "loss": 1.6478, + "theoretical_loss": 3.9992381294457733, + "tokens_seen": 438960128 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043786359077231695, + "loss": 1.6109, + "theoretical_loss": 3.9991728438944767, + "tokens_seen": 439025664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 165818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6244661808013916, + "objective/train/theoretical_loss": 3.9991075708163146, + "objective/train/tokens_used": 459551200, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043785356068204613, + "loss": 1.64, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043784353059177537, + "loss": 1.7065, + "theoretical_loss": 3.9990423102070416, + "tokens_seen": 439156736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378335005015045, + "loss": 1.664, + "theoretical_loss": 3.9989770620624165, + "tokens_seen": 439222272 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043782347041123373, + "loss": 1.7583, + "theoretical_loss": 3.9989118263781984, + "tokens_seen": 439287808 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043781344032096286, + "loss": 1.6511, + "theoretical_loss": 3.99884660315015, + "tokens_seen": 439353344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004378034102306921, + "loss": 1.7285, + "theoretical_loss": 3.9987813923740356, + "tokens_seen": 439418880 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043779338014042127, + "loss": 1.6042, + "theoretical_loss": 3.9987161940456204, + "tokens_seen": 439484416 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043778335005015045, + "loss": 1.5709, + "theoretical_loss": 3.9986510081606736, + "tokens_seen": 439549952 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043777331995987963, + "loss": 1.6805, + "theoretical_loss": 3.998585834714965, + "tokens_seen": 439615488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377632898696088, + "loss": 1.7419, + "theoretical_loss": 3.9985206737042676, + "tokens_seen": 439681024 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437753259779338, + "loss": 1.6647, + "theoretical_loss": 3.998455525124355, + "tokens_seen": 439746560 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043774322968906723, + "loss": 1.6187, + "theoretical_loss": 3.9983903889710044, + "tokens_seen": 439812096 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043773319959879636, + "loss": 1.6439, + "theoretical_loss": 3.998325265239994, + "tokens_seen": 439877632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377231695085256, + "loss": 1.6284, + "theoretical_loss": 3.998260153927105, + "tokens_seen": 439943168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004377131394182547, + "loss": 1.5595, + "theoretical_loss": 3.998195055028119, + "tokens_seen": 440008704 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043770310932798396, + "loss": 1.6122, + "theoretical_loss": 3.998129968538822, + "tokens_seen": 440074240 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043769307923771314, + "loss": 1.6059, + "theoretical_loss": 3.998064894455, + "tokens_seen": 440139776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376830491474423, + "loss": 1.6769, + "theoretical_loss": 3.997999832772443, + "tokens_seen": 440205312 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043767301905717155, + "loss": 1.6358, + "theoretical_loss": 3.997934783486941, + "tokens_seen": 440270848 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043766298896690073, + "loss": 1.6146, + "theoretical_loss": 3.997869746594287, + "tokens_seen": 440336384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376529588766299, + "loss": 1.7004, + "theoretical_loss": 3.9978047220902764, + "tokens_seen": 440401920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376429287863591, + "loss": 1.6321, + "theoretical_loss": 3.997739709970706, + "tokens_seen": 440467456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376328986960883, + "loss": 1.5431, + "theoretical_loss": 3.997674710231375, + "tokens_seen": 440532992 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043762286860581746, + "loss": 1.77, + "theoretical_loss": 3.997609722868085, + "tokens_seen": 440598528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376128385155467, + "loss": 1.5333, + "theoretical_loss": 3.9975447478766393, + "tokens_seen": 440664064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004376028084252758, + "loss": 1.6217, + "theoretical_loss": 3.997479785252842, + "tokens_seen": 440729600 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043759277833500506, + "loss": 1.6673, + "theoretical_loss": 3.997414834992502, + "tokens_seen": 440795136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375827482447342, + "loss": 1.589, + "theoretical_loss": 3.9973498970914285, + "tokens_seen": 440860672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375727181544634, + "loss": 1.6514, + "theoretical_loss": 3.9972849715454317, + "tokens_seen": 440926208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375626880641926, + "loss": 1.5819, + "theoretical_loss": 3.9972200583503263, + "tokens_seen": 440991744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375526579739218, + "loss": 1.6491, + "theoretical_loss": 3.9971551575019273, + "tokens_seen": 441057280 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043754262788365096, + "loss": 1.7092, + "theoretical_loss": 3.997090268996052, + "tokens_seen": 441122816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375325977933802, + "loss": 1.577, + "theoretical_loss": 3.9970253928285215, + "tokens_seen": 441188352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375225677031093, + "loss": 1.6014, + "theoretical_loss": 3.996960528995155, + "tokens_seen": 441253888 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043751253761283856, + "loss": 1.7318, + "theoretical_loss": 3.996895677491778, + "tokens_seen": 441319424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004375025075225677, + "loss": 1.6256, + "theoretical_loss": 3.996830838314216, + "tokens_seen": 441384960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374924774322969, + "loss": 1.5565, + "theoretical_loss": 3.996766011458296, + "tokens_seen": 441450496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374824473420261, + "loss": 1.6164, + "theoretical_loss": 3.9967011969198483, + "tokens_seen": 441516032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374724172517553, + "loss": 1.5436, + "theoretical_loss": 3.996636394694705, + "tokens_seen": 441581568 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043746238716148447, + "loss": 1.603, + "theoretical_loss": 3.9965716047786985, + "tokens_seen": 441647104 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043745235707121365, + "loss": 1.4097, + "theoretical_loss": 3.996506827167667, + "tokens_seen": 441712640 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043744232698094283, + "loss": 1.6578, + "theoretical_loss": 3.9964420618574454, + "tokens_seen": 441778176 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043743229689067206, + "loss": 1.675, + "theoretical_loss": 3.9963773088438765, + "tokens_seen": 441843712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374222668004012, + "loss": 1.6151, + "theoretical_loss": 3.996312568122801, + "tokens_seen": 441909248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004374122367101304, + "loss": 1.5843, + "theoretical_loss": 3.9962478396900627, + "tokens_seen": 441974784 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043740220661985955, + "loss": 1.5784, + "theoretical_loss": 3.996183123541507, + "tokens_seen": 442040320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373921765295888, + "loss": 1.7877, + "theoretical_loss": 3.9961184196729835, + "tokens_seen": 442105856 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043738214643931797, + "loss": 1.5426, + "theoretical_loss": 3.9960537280803408, + "tokens_seen": 442171392 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043737211634904715, + "loss": 1.5863, + "theoretical_loss": 3.995989048759432, + "tokens_seen": 442236928 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043736208625877633, + "loss": 1.6469, + "theoretical_loss": 3.99592438170611, + "tokens_seen": 442302464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 166951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5862295627593994, + "objective/train/theoretical_loss": 3.9958597269162315, + "objective/train/tokens_used": 462828000, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043735205616850557, + "loss": 1.6493, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373420260782347, + "loss": 1.5193, + "theoretical_loss": 3.9957950843856542, + "tokens_seen": 442433536 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043733199598796393, + "loss": 1.6236, + "theoretical_loss": 3.9957304541102396, + "tokens_seen": 442499072 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043732196589769306, + "loss": 1.602, + "theoretical_loss": 3.995665836085848, + "tokens_seen": 442564608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004373119358074223, + "loss": 1.4955, + "theoretical_loss": 3.9956012303083446, + "tokens_seen": 442630144 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043730190571715147, + "loss": 1.651, + "theoretical_loss": 3.9955366367735947, + "tokens_seen": 442695680 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043729187562688065, + "loss": 1.5377, + "theoretical_loss": 3.995472055477467, + "tokens_seen": 442761216 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043728184553660983, + "loss": 1.7269, + "theoretical_loss": 3.9954074864158313, + "tokens_seen": 442826752 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437271815446339, + "loss": 1.5599, + "theoretical_loss": 3.9953429295845604, + "tokens_seen": 442892288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372617853560682, + "loss": 1.5677, + "theoretical_loss": 3.9952783849795277, + "tokens_seen": 442957824 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043725175526579743, + "loss": 1.4442, + "theoretical_loss": 3.99521385259661, + "tokens_seen": 443023360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043724172517552656, + "loss": 1.7659, + "theoretical_loss": 3.9951493324316845, + "tokens_seen": 443088896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372316950852558, + "loss": 1.4801, + "theoretical_loss": 3.9950848244806316, + "tokens_seen": 443154432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004372216649949849, + "loss": 1.6358, + "theoretical_loss": 3.9950203287393347, + "tokens_seen": 443219968 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043721163490471416, + "loss": 1.5303, + "theoretical_loss": 3.994955845203676, + "tokens_seen": 443285504 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043720160481444334, + "loss": 1.6854, + "theoretical_loss": 3.994891373869543, + "tokens_seen": 443351040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371915747241725, + "loss": 1.5922, + "theoretical_loss": 3.9948269147328235, + "tokens_seen": 443416576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371815446339017, + "loss": 1.6397, + "theoretical_loss": 3.994762467789407, + "tokens_seen": 443482112 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043717151454363093, + "loss": 1.7108, + "theoretical_loss": 3.9946980330351862, + "tokens_seen": 443547648 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043716148445336006, + "loss": 1.6191, + "theoretical_loss": 3.994633610466056, + "tokens_seen": 443613184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371514543630893, + "loss": 1.5751, + "theoretical_loss": 3.994569200077911, + "tokens_seen": 443678720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371414242728184, + "loss": 1.611, + "theoretical_loss": 3.9945048018666496, + "tokens_seen": 443744256 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043713139418254766, + "loss": 1.6349, + "theoretical_loss": 3.9944404158281723, + "tokens_seen": 443809792 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043712136409227684, + "loss": 1.5689, + "theoretical_loss": 3.994376041958381, + "tokens_seen": 443875328 + }, + { + "epoch": 0.13, + "learning_rate": 0.000437111334002006, + "loss": 1.6008, + "theoretical_loss": 3.9943116802531797, + "tokens_seen": 443940864 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004371013039117352, + "loss": 1.5528, + "theoretical_loss": 3.9942473307084745, + "tokens_seen": 444006400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370912738214644, + "loss": 1.5968, + "theoretical_loss": 3.9941829933201736, + "tokens_seen": 444071936 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043708124373119356, + "loss": 1.6786, + "theoretical_loss": 3.9941186680841865, + "tokens_seen": 444137472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370712136409228, + "loss": 1.7048, + "theoretical_loss": 3.9940543549964254, + "tokens_seen": 444203008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370611835506519, + "loss": 1.4881, + "theoretical_loss": 3.993990054052804, + "tokens_seen": 444268544 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043705115346038116, + "loss": 1.5086, + "theoretical_loss": 3.993925765249239, + "tokens_seen": 444334080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370411233701103, + "loss": 1.6262, + "theoretical_loss": 3.993861488581647, + "tokens_seen": 444399616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370310932798395, + "loss": 1.7317, + "theoretical_loss": 3.9937972240459487, + "tokens_seen": 444465152 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370210631895687, + "loss": 1.6035, + "theoretical_loss": 3.9937329716380665, + "tokens_seen": 444530688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004370110330992979, + "loss": 1.6453, + "theoretical_loss": 3.9936687313539228, + "tokens_seen": 444596224 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043700100300902707, + "loss": 1.5767, + "theoretical_loss": 3.9936045031894443, + "tokens_seen": 444661760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369909729187563, + "loss": 1.5921, + "theoretical_loss": 3.9935402871405588, + "tokens_seen": 444727296 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043698094282848543, + "loss": 1.5939, + "theoretical_loss": 3.9934760832031957, + "tokens_seen": 444792832 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043697091273821467, + "loss": 1.6923, + "theoretical_loss": 3.9934118913732863, + "tokens_seen": 444858368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369608826479438, + "loss": 1.6749, + "theoretical_loss": 3.9933477116467646, + "tokens_seen": 444923904 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043695085255767303, + "loss": 1.7364, + "theoretical_loss": 3.993283544019567, + "tokens_seen": 444989440 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369408224674022, + "loss": 1.63, + "theoretical_loss": 3.9932193884876304, + "tokens_seen": 445054976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369307923771314, + "loss": 1.605, + "theoretical_loss": 3.9931552450468937, + "tokens_seen": 445120512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0004369207622868606, + "loss": 1.5971, + "theoretical_loss": 3.9930911136933, + "tokens_seen": 445186048 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043691073219658975, + "loss": 1.5759, + "theoretical_loss": 3.9930269944227907, + "tokens_seen": 445251584 + }, + { + "epoch": 0.13, + "learning_rate": 0.000436900702106319, + "loss": 1.65, + "theoretical_loss": 3.9929628872313128, + "tokens_seen": 445317120 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043689067201604817, + "loss": 1.6218, + "theoretical_loss": 3.9928987921148127, + "tokens_seen": 445382656 + }, + { + "epoch": 0.13, + "learning_rate": 0.00043688064192577735, + "loss": 1.5813, + "theoretical_loss": 3.992834709069241, + "tokens_seen": 445448192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043687061183550653, + "loss": 1.5931, + "theoretical_loss": 3.9927706380905477, + "tokens_seen": 445513728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043686058174523577, + "loss": 1.5904, + "theoretical_loss": 3.992706579174686, + "tokens_seen": 445579264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 168097, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7863926887512207, + "objective/train/theoretical_loss": 3.9926425323176122, + "objective/train/tokens_used": 466104800, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004368505516549649, + "loss": 1.703, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043684052156469413, + "loss": 1.5587, + "theoretical_loss": 3.992578497515283, + "tokens_seen": 445710336 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043683049147442326, + "loss": 1.6643, + "theoretical_loss": 3.9925144747636576, + "tokens_seen": 445775872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004368204613841525, + "loss": 1.5565, + "theoretical_loss": 3.992450464058696, + "tokens_seen": 445841408 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043681043129388167, + "loss": 1.5149, + "theoretical_loss": 3.9923864653963625, + "tokens_seen": 445906944 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043680040120361085, + "loss": 1.5866, + "theoretical_loss": 3.9923224787726213, + "tokens_seen": 445972480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043679037111334003, + "loss": 1.5626, + "theoretical_loss": 3.9922585041834395, + "tokens_seen": 446038016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367803410230692, + "loss": 1.5629, + "theoretical_loss": 3.9921945416247855, + "tokens_seen": 446103552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367703109327984, + "loss": 1.5531, + "theoretical_loss": 3.9921305910926312, + "tokens_seen": 446169088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043676028084252763, + "loss": 1.5521, + "theoretical_loss": 3.992066652582948, + "tokens_seen": 446234624 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043675025075225676, + "loss": 1.6186, + "theoretical_loss": 3.9920027260917115, + "tokens_seen": 446300160 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436740220661986, + "loss": 1.6824, + "theoretical_loss": 3.9919388116148977, + "tokens_seen": 446365696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367301905717151, + "loss": 1.5681, + "theoretical_loss": 3.991874909148485, + "tokens_seen": 446431232 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043672016048144436, + "loss": 1.5902, + "theoretical_loss": 3.9918110186884546, + "tokens_seen": 446496768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043671013039117354, + "loss": 1.6573, + "theoretical_loss": 3.991747140230788, + "tokens_seen": 446562304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004367001003009027, + "loss": 1.7275, + "theoretical_loss": 3.9916832737714705, + "tokens_seen": 446627840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366900702106319, + "loss": 1.6688, + "theoretical_loss": 3.9916194193064873, + "tokens_seen": 446693376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043668004012036113, + "loss": 1.6341, + "theoretical_loss": 3.9915555768318276, + "tokens_seen": 446758912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043667001003009026, + "loss": 1.5837, + "theoretical_loss": 3.9914917463434807, + "tokens_seen": 446824448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366599799398195, + "loss": 1.623, + "theoretical_loss": 3.9914279278374387, + "tokens_seen": 446889984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366499498495486, + "loss": 1.607, + "theoretical_loss": 3.9913641213096964, + "tokens_seen": 446955520 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043663991975927786, + "loss": 1.5329, + "theoretical_loss": 3.9913003267562486, + "tokens_seen": 447021056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043662988966900704, + "loss": 1.5442, + "theoretical_loss": 3.9912365441730944, + "tokens_seen": 447086592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366198595787362, + "loss": 1.7292, + "theoretical_loss": 3.9911727735562326, + "tokens_seen": 447152128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004366098294884654, + "loss": 1.4361, + "theoretical_loss": 3.9911090149016646, + "tokens_seen": 447217664 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365997993981946, + "loss": 1.6782, + "theoretical_loss": 3.9910452682053954, + "tokens_seen": 447283200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043658976930792376, + "loss": 1.5515, + "theoretical_loss": 3.9909815334634287, + "tokens_seen": 447348736 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436579739217653, + "loss": 1.5239, + "theoretical_loss": 3.990917810671773, + "tokens_seen": 447414272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043656970912738213, + "loss": 1.5636, + "theoretical_loss": 3.9908540998264375, + "tokens_seen": 447479808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043655967903711136, + "loss": 1.7094, + "theoretical_loss": 3.9907904009234336, + "tokens_seen": 447545344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365496489468405, + "loss": 1.5824, + "theoretical_loss": 3.9907267139587743, + "tokens_seen": 447610880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365396188565697, + "loss": 1.4794, + "theoretical_loss": 3.9906630389284747, + "tokens_seen": 447676416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365295887662989, + "loss": 1.5295, + "theoretical_loss": 3.990599375828552, + "tokens_seen": 447741952 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004365195586760281, + "loss": 1.5661, + "theoretical_loss": 3.9905357246550253, + "tokens_seen": 447807488 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043650952858575727, + "loss": 1.5121, + "theoretical_loss": 3.9904720854039137, + "tokens_seen": 447873024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364994984954865, + "loss": 1.6643, + "theoretical_loss": 3.9904084580712427, + "tokens_seen": 447938560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043648946840521563, + "loss": 1.5393, + "theoretical_loss": 3.9903448426530352, + "tokens_seen": 448004096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043647943831494487, + "loss": 1.7026, + "theoretical_loss": 3.990281239145318, + "tokens_seen": 448069632 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436469408224674, + "loss": 1.5971, + "theoretical_loss": 3.99021764754412, + "tokens_seen": 448135168 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043645937813440323, + "loss": 1.6086, + "theoretical_loss": 3.990154067845472, + "tokens_seen": 448200704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364493480441324, + "loss": 1.6212, + "theoretical_loss": 3.990090500045404, + "tokens_seen": 448266240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004364393179538616, + "loss": 1.6481, + "theoretical_loss": 3.990026944139953, + "tokens_seen": 448331776 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043642928786359077, + "loss": 1.496, + "theoretical_loss": 3.989963400125153, + "tokens_seen": 448397312 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043641925777331995, + "loss": 1.5569, + "theoretical_loss": 3.9898998679970434, + "tokens_seen": 448462848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043640922768304913, + "loss": 1.5572, + "theoretical_loss": 3.9898363477516634, + "tokens_seen": 448528384 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043639919759277837, + "loss": 1.6398, + "theoretical_loss": 3.9897728393850547, + "tokens_seen": 448593920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363891675025075, + "loss": 1.5123, + "theoretical_loss": 3.9897093428932613, + "tokens_seen": 448659456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043637913741223673, + "loss": 1.5145, + "theoretical_loss": 3.989645858272328, + "tokens_seen": 448724992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363691073219659, + "loss": 1.5507, + "theoretical_loss": 3.989582385518304, + "tokens_seen": 448790528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363590772316951, + "loss": 1.5823, + "theoretical_loss": 3.989518924627237, + "tokens_seen": 448856064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 169309, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6853488683700562, + "objective/train/theoretical_loss": 3.9894554755951788, + "objective/train/tokens_used": 469381600, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004363490471414243, + "loss": 1.6155, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043633901705115346, + "loss": 1.6034, + "theoretical_loss": 3.989392038418182, + "tokens_seen": 448987136 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043632898696088264, + "loss": 1.4644, + "theoretical_loss": 3.9893286130923027, + "tokens_seen": 449052672 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043631895687061187, + "loss": 1.6863, + "theoretical_loss": 3.989265199613597, + "tokens_seen": 449118208 + }, + { + "epoch": 0.14, + "learning_rate": 0.000436308926780341, + "loss": 1.5462, + "theoretical_loss": 3.989201797978124, + "tokens_seen": 449183744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043629889669007023, + "loss": 1.5136, + "theoretical_loss": 3.989138408181944, + "tokens_seen": 449249280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043628886659979936, + "loss": 1.6027, + "theoretical_loss": 3.9890750302211204, + "tokens_seen": 449314816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362788365095286, + "loss": 1.5953, + "theoretical_loss": 3.9890116640917173, + "tokens_seen": 449380352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362688064192578, + "loss": 1.6795, + "theoretical_loss": 3.9889483097898006, + "tokens_seen": 449445888 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043625877632898696, + "loss": 1.6751, + "theoretical_loss": 3.9888849673114386, + "tokens_seen": 449511424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043624874623871614, + "loss": 1.6542, + "theoretical_loss": 3.988821636652702, + "tokens_seen": 449576960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362387161484453, + "loss": 1.528, + "theoretical_loss": 3.988758317809662, + "tokens_seen": 449642496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004362286860581745, + "loss": 1.5911, + "theoretical_loss": 3.988695010778393, + "tokens_seen": 449708032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043621865596790374, + "loss": 1.6886, + "theoretical_loss": 3.9886317155549706, + "tokens_seen": 449773568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043620862587763286, + "loss": 1.5549, + "theoretical_loss": 3.988568432135472, + "tokens_seen": 449839104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361985957873621, + "loss": 1.5998, + "theoretical_loss": 3.988505160515978, + "tokens_seen": 449904640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361885656970913, + "loss": 1.6548, + "theoretical_loss": 3.9884419006925684, + "tokens_seen": 449970176 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043617853560682046, + "loss": 1.5587, + "theoretical_loss": 3.9883786526613267, + "tokens_seen": 450035712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361685055165497, + "loss": 1.5777, + "theoretical_loss": 3.988315416418339, + "tokens_seen": 450101248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361584754262788, + "loss": 1.5975, + "theoretical_loss": 3.988252191959692, + "tokens_seen": 450166784 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043614844533600806, + "loss": 1.7001, + "theoretical_loss": 3.9881889792814733, + "tokens_seen": 450232320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043613841524573724, + "loss": 1.6586, + "theoretical_loss": 3.9881257783797746, + "tokens_seen": 450297856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361283851554664, + "loss": 1.5811, + "theoretical_loss": 3.9880625892506885, + "tokens_seen": 450363392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361183550651956, + "loss": 1.6205, + "theoretical_loss": 3.9879994118903093, + "tokens_seen": 450428928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004361083249749248, + "loss": 1.6398, + "theoretical_loss": 3.9879362462947334, + "tokens_seen": 450494464 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043609829488465397, + "loss": 1.6605, + "theoretical_loss": 3.9878730924600583, + "tokens_seen": 450560000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360882647943832, + "loss": 1.6383, + "theoretical_loss": 3.9878099503823847, + "tokens_seen": 450625536 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043607823470411233, + "loss": 1.6036, + "theoretical_loss": 3.9877468200578146, + "tokens_seen": 450691072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043606820461384156, + "loss": 1.6273, + "theoretical_loss": 3.9876837014824513, + "tokens_seen": 450756608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360581745235707, + "loss": 1.6195, + "theoretical_loss": 3.9876205946524, + "tokens_seen": 450822144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360481444332999, + "loss": 1.545, + "theoretical_loss": 3.9875574995637697, + "tokens_seen": 450887680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360381143430291, + "loss": 1.6041, + "theoretical_loss": 3.987494416212668, + "tokens_seen": 450953216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360280842527583, + "loss": 1.6015, + "theoretical_loss": 3.987431344595207, + "tokens_seen": 451018752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043601805416248747, + "loss": 1.4559, + "theoretical_loss": 3.9873682847074994, + "tokens_seen": 451084288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004360080240722167, + "loss": 1.5907, + "theoretical_loss": 3.9873052365456605, + "tokens_seen": 451149824 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043599799398194583, + "loss": 1.4632, + "theoretical_loss": 3.9872422001058063, + "tokens_seen": 451215360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043598796389167507, + "loss": 1.5624, + "theoretical_loss": 3.987179175384056, + "tokens_seen": 451280896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359779338014042, + "loss": 1.5521, + "theoretical_loss": 3.9871161623765294, + "tokens_seen": 451346432 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043596790371113343, + "loss": 1.6241, + "theoretical_loss": 3.9870531610793494, + "tokens_seen": 451411968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359578736208626, + "loss": 1.6009, + "theoretical_loss": 3.9869901714886398, + "tokens_seen": 451477504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004359478435305918, + "loss": 1.5383, + "theoretical_loss": 3.9869271936005264, + "tokens_seen": 451543040 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043593781344032097, + "loss": 1.627, + "theoretical_loss": 3.9868642274111377, + "tokens_seen": 451608576 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043592778335005015, + "loss": 1.5836, + "theoretical_loss": 3.986801272916603, + "tokens_seen": 451674112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043591775325977933, + "loss": 1.5998, + "theoretical_loss": 3.9867383301130532, + "tokens_seen": 451739648 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043590772316950857, + "loss": 1.5886, + "theoretical_loss": 3.9866753989966224, + "tokens_seen": 451805184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358976930792377, + "loss": 1.7255, + "theoretical_loss": 3.986612479563445, + "tokens_seen": 451870720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043588766298896693, + "loss": 1.6661, + "theoretical_loss": 3.9865495718096584, + "tokens_seen": 451936256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358776328986961, + "loss": 1.5825, + "theoretical_loss": 3.9864866757314017, + "tokens_seen": 452001792 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358676028084253, + "loss": 1.6128, + "theoretical_loss": 3.9864237913248157, + "tokens_seen": 452067328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358575727181545, + "loss": 1.6216, + "theoretical_loss": 3.9863609185860422, + "tokens_seen": 452132864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 170598, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7061796188354492, + "objective/train/theoretical_loss": 3.9862980575112257, + "objective/train/tokens_used": 472658400, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043584754262788366, + "loss": 1.65, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043583751253761284, + "loss": 1.6671, + "theoretical_loss": 3.986235208096513, + "tokens_seen": 452263936 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043582748244734207, + "loss": 1.6602, + "theoretical_loss": 3.9861723703380516, + "tokens_seen": 452329472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004358174523570712, + "loss": 1.6709, + "theoretical_loss": 3.986109544231992, + "tokens_seen": 452395008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043580742226680043, + "loss": 1.6396, + "theoretical_loss": 3.9860467297744844, + "tokens_seen": 452460544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043579739217652956, + "loss": 1.5663, + "theoretical_loss": 3.985983926961684, + "tokens_seen": 452526080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357873620862588, + "loss": 1.6097, + "theoretical_loss": 3.985921135789745, + "tokens_seen": 452591616 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435777331995988, + "loss": 1.5236, + "theoretical_loss": 3.9858583562548247, + "tokens_seen": 452657152 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043576730190571716, + "loss": 1.6841, + "theoretical_loss": 3.9857955883530827, + "tokens_seen": 452722688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043575727181544634, + "loss": 1.663, + "theoretical_loss": 3.985732832080679, + "tokens_seen": 452788224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357472417251755, + "loss": 1.5879, + "theoretical_loss": 3.985670087433777, + "tokens_seen": 452853760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357372116349047, + "loss": 1.6473, + "theoretical_loss": 3.9856073544085406, + "tokens_seen": 452919296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043572718154463394, + "loss": 1.7349, + "theoretical_loss": 3.9855446330011364, + "tokens_seen": 452984832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043571715145436306, + "loss": 1.6418, + "theoretical_loss": 3.9854819232077316, + "tokens_seen": 453050368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004357071213640923, + "loss": 1.6025, + "theoretical_loss": 3.985419225024498, + "tokens_seen": 453115904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356970912738215, + "loss": 1.6735, + "theoretical_loss": 3.985356538447605, + "tokens_seen": 453181440 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043568706118355066, + "loss": 1.7369, + "theoretical_loss": 3.9852938634732276, + "tokens_seen": 453246976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043567703109327984, + "loss": 1.6243, + "theoretical_loss": 3.985231200097541, + "tokens_seen": 453312512 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435667001003009, + "loss": 1.5798, + "theoretical_loss": 3.9851685483167225, + "tokens_seen": 453378048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356569709127382, + "loss": 1.5738, + "theoretical_loss": 3.9851059081269504, + "tokens_seen": 453443584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043564694082246744, + "loss": 1.5437, + "theoretical_loss": 3.9850432795244055, + "tokens_seen": 453509120 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043563691073219657, + "loss": 1.518, + "theoretical_loss": 3.9849806625052713, + "tokens_seen": 453574656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004356268806419258, + "loss": 1.5312, + "theoretical_loss": 3.984918057065731, + "tokens_seen": 453640192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043561685055165493, + "loss": 1.5607, + "theoretical_loss": 3.984855463201972, + "tokens_seen": 453705728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043560682046138417, + "loss": 1.5257, + "theoretical_loss": 3.984792880910182, + "tokens_seen": 453771264 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043559679037111335, + "loss": 1.643, + "theoretical_loss": 3.98473031018655, + "tokens_seen": 453836800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043558676028084253, + "loss": 1.6142, + "theoretical_loss": 3.984667751027268, + "tokens_seen": 453902336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355767301905717, + "loss": 1.5535, + "theoretical_loss": 3.98460520342853, + "tokens_seen": 453967872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355667001003009, + "loss": 1.6178, + "theoretical_loss": 3.984542667386531, + "tokens_seen": 454033408 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043555667001003007, + "loss": 1.5738, + "theoretical_loss": 3.9844801428974685, + "tokens_seen": 454098944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004355466399197593, + "loss": 1.6088, + "theoretical_loss": 3.98441762995754, + "tokens_seen": 454164480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043553660982948843, + "loss": 1.5872, + "theoretical_loss": 3.984355128562947, + "tokens_seen": 454230016 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043552657973921767, + "loss": 1.6948, + "theoretical_loss": 3.984292638709892, + "tokens_seen": 454295552 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043551654964894685, + "loss": 1.5352, + "theoretical_loss": 3.9842301603945787, + "tokens_seen": 454361088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043550651955867603, + "loss": 1.6034, + "theoretical_loss": 3.9841676936132133, + "tokens_seen": 454426624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354964894684052, + "loss": 1.5569, + "theoretical_loss": 3.984105238362005, + "tokens_seen": 454492160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354864593781344, + "loss": 1.5615, + "theoretical_loss": 3.984042794637161, + "tokens_seen": 454557696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354764292878636, + "loss": 1.7295, + "theoretical_loss": 3.9839803624348944, + "tokens_seen": 454623232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354663991975928, + "loss": 1.5467, + "theoretical_loss": 3.983917941751418, + "tokens_seen": 454688768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043545636910732194, + "loss": 1.6195, + "theoretical_loss": 3.983855532582947, + "tokens_seen": 454754304 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043544633901705117, + "loss": 1.6128, + "theoretical_loss": 3.983793134925697, + "tokens_seen": 454819840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354363089267803, + "loss": 1.5187, + "theoretical_loss": 3.9837307487758884, + "tokens_seen": 454885376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043542627883650953, + "loss": 1.6018, + "theoretical_loss": 3.9836683741297403, + "tokens_seen": 454950912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043541624874623877, + "loss": 1.554, + "theoretical_loss": 3.9836060109834746, + "tokens_seen": 455016448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004354062186559679, + "loss": 1.5596, + "theoretical_loss": 3.983543659333316, + "tokens_seen": 455081984 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043539618856569713, + "loss": 1.6339, + "theoretical_loss": 3.9834813191754908, + "tokens_seen": 455147520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353861584754263, + "loss": 1.6659, + "theoretical_loss": 3.983418990506225, + "tokens_seen": 455213056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353761283851555, + "loss": 1.583, + "theoretical_loss": 3.9833566733217483, + "tokens_seen": 455278592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353660982948847, + "loss": 1.6376, + "theoretical_loss": 3.983294367618292, + "tokens_seen": 455344128 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043535606820461386, + "loss": 1.6196, + "theoretical_loss": 3.983232073392089, + "tokens_seen": 455409664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 171735, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6469066143035889, + "objective/train/theoretical_loss": 3.983169790639374, + "objective/train/tokens_used": 475935200, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043534603811434304, + "loss": 1.6923, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043533600802407227, + "loss": 1.5806, + "theoretical_loss": 3.983107519356383, + "tokens_seen": 455540736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004353259779338014, + "loss": 1.676, + "theoretical_loss": 3.9830452595393546, + "tokens_seen": 455606272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043531594784353063, + "loss": 1.6035, + "theoretical_loss": 3.9829830111845284, + "tokens_seen": 455671808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043530591775325976, + "loss": 1.6277, + "theoretical_loss": 3.9829207742881456, + "tokens_seen": 455737344 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435295887662989, + "loss": 1.5877, + "theoretical_loss": 3.982858548846451, + "tokens_seen": 455802880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352858575727182, + "loss": 1.6418, + "theoretical_loss": 3.982796334855689, + "tokens_seen": 455868416 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043527582748244736, + "loss": 1.5741, + "theoretical_loss": 3.9827341323121064, + "tokens_seen": 455933952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043526579739217654, + "loss": 1.6419, + "theoretical_loss": 3.9826719412119527, + "tokens_seen": 455999488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352557673019057, + "loss": 1.5426, + "theoretical_loss": 3.982609761551478, + "tokens_seen": 456065024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352457372116349, + "loss": 1.5866, + "theoretical_loss": 3.982547593326934, + "tokens_seen": 456130560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043523570712136414, + "loss": 1.5871, + "theoretical_loss": 3.982485436534576, + "tokens_seen": 456196096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043522567703109326, + "loss": 1.5681, + "theoretical_loss": 3.98242329117066, + "tokens_seen": 456261632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352156469408225, + "loss": 1.6478, + "theoretical_loss": 3.982361157231442, + "tokens_seen": 456327168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004352056168505517, + "loss": 1.5643, + "theoretical_loss": 3.982299034713183, + "tokens_seen": 456392704 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043519558676028086, + "loss": 1.5724, + "theoretical_loss": 3.9822369236121427, + "tokens_seen": 456458240 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043518555667001004, + "loss": 1.4882, + "theoretical_loss": 3.982174823924586, + "tokens_seen": 456523776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004351755265797392, + "loss": 1.6509, + "theoretical_loss": 3.9821127356467754, + "tokens_seen": 456589312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004351654964894684, + "loss": 1.5862, + "theoretical_loss": 3.9820506587749787, + "tokens_seen": 456654848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043515546639919764, + "loss": 1.4661, + "theoretical_loss": 3.9819885933054637, + "tokens_seen": 456720384 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043514543630892677, + "loss": 1.5519, + "theoretical_loss": 3.9819265392345, + "tokens_seen": 456785920 + }, + { + "epoch": 0.14, + "learning_rate": 0.000435135406218656, + "loss": 1.5291, + "theoretical_loss": 3.9818644965583596, + "tokens_seen": 456851456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043512537612838513, + "loss": 1.5276, + "theoretical_loss": 3.981802465273316, + "tokens_seen": 456916992 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043511534603811437, + "loss": 1.5641, + "theoretical_loss": 3.9817404453756446, + "tokens_seen": 456982528 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043510531594784355, + "loss": 1.6494, + "theoretical_loss": 3.981678436861622, + "tokens_seen": 457048064 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043509528585757273, + "loss": 1.5546, + "theoretical_loss": 3.981616439727527, + "tokens_seen": 457113600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350852557673019, + "loss": 1.648, + "theoretical_loss": 3.9815544539696397, + "tokens_seen": 457179136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350752256770311, + "loss": 1.6766, + "theoretical_loss": 3.981492479584243, + "tokens_seen": 457244672 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043506519558676027, + "loss": 1.5777, + "theoretical_loss": 3.9814305165676203, + "tokens_seen": 457310208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350551654964895, + "loss": 1.5736, + "theoretical_loss": 3.9813685649160577, + "tokens_seen": 457375744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043504513540621863, + "loss": 1.4253, + "theoretical_loss": 3.981306624625842, + "tokens_seen": 457441280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043503510531594787, + "loss": 1.6551, + "theoretical_loss": 3.9812446956932632, + "tokens_seen": 457506816 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043502507522567705, + "loss": 1.7039, + "theoretical_loss": 3.9811827781146123, + "tokens_seen": 457572352 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043501504513540623, + "loss": 1.5795, + "theoretical_loss": 3.981120871886181, + "tokens_seen": 457637888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004350050150451354, + "loss": 1.6133, + "theoretical_loss": 3.9810589770042637, + "tokens_seen": 457703424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349949849548646, + "loss": 1.6, + "theoretical_loss": 3.9809970934651577, + "tokens_seen": 457768960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349849548645938, + "loss": 1.6478, + "theoretical_loss": 3.9809352212651596, + "tokens_seen": 457834496 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434974924774323, + "loss": 1.6515, + "theoretical_loss": 3.9808733604005706, + "tokens_seen": 457900032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043496489468405214, + "loss": 1.5893, + "theoretical_loss": 3.9808115108676905, + "tokens_seen": 457965568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043495486459378137, + "loss": 1.6346, + "theoretical_loss": 3.9807496726628235, + "tokens_seen": 458031104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349448345035105, + "loss": 1.5142, + "theoretical_loss": 3.9806878457822736, + "tokens_seen": 458096640 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043493480441323973, + "loss": 1.6132, + "theoretical_loss": 3.9806260302223477, + "tokens_seen": 458162176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349247743229689, + "loss": 1.7152, + "theoretical_loss": 3.9805642259793546, + "tokens_seen": 458227712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349147442326981, + "loss": 1.7395, + "theoretical_loss": 3.9805024330496037, + "tokens_seen": 458293248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004349047141424273, + "loss": 1.6219, + "theoretical_loss": 3.980440651429407, + "tokens_seen": 458358784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348946840521565, + "loss": 1.6501, + "theoretical_loss": 3.9803788811150778, + "tokens_seen": 458424320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043488465396188564, + "loss": 1.5931, + "theoretical_loss": 3.980317122102931, + "tokens_seen": 458489856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348746238716149, + "loss": 1.5766, + "theoretical_loss": 3.9802553743892846, + "tokens_seen": 458555392 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434864593781344, + "loss": 1.5606, + "theoretical_loss": 3.9801936379704568, + "tokens_seen": 458620928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043485456369107324, + "loss": 1.6492, + "theoretical_loss": 3.9801319128427677, + "tokens_seen": 458686464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 173009, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5641891956329346, + "objective/train/theoretical_loss": 3.9800701990025393, + "objective/train/tokens_used": 479212000, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348445336008024, + "loss": 1.5385, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348345035105316, + "loss": 1.6516, + "theoretical_loss": 3.9800084964460964, + "tokens_seen": 458817536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004348244734202608, + "loss": 1.4985, + "theoretical_loss": 3.9799468051697633, + "tokens_seen": 458883072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043481444332998996, + "loss": 1.537, + "theoretical_loss": 3.9798851251698686, + "tokens_seen": 458948608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043480441323971914, + "loss": 1.481, + "theoretical_loss": 3.97982345644274, + "tokens_seen": 459014144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347943831494484, + "loss": 1.6231, + "theoretical_loss": 3.9797617989847094, + "tokens_seen": 459079680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347843530591775, + "loss": 1.6557, + "theoretical_loss": 3.979700152792108, + "tokens_seen": 459145216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043477432296890674, + "loss": 1.6152, + "theoretical_loss": 3.9796385178612717, + "tokens_seen": 459210752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043476429287863587, + "loss": 1.5633, + "theoretical_loss": 3.9795768941885346, + "tokens_seen": 459276288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347542627883651, + "loss": 1.6225, + "theoretical_loss": 3.9795152817702357, + "tokens_seen": 459341824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347442326980943, + "loss": 1.5792, + "theoretical_loss": 3.979453680602713, + "tokens_seen": 459407360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043473420260782346, + "loss": 1.6713, + "theoretical_loss": 3.9793920906823086, + "tokens_seen": 459472896 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043472417251755265, + "loss": 1.505, + "theoretical_loss": 3.979330512005365, + "tokens_seen": 459538432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004347141424272819, + "loss": 1.5946, + "theoretical_loss": 3.9792689445682257, + "tokens_seen": 459603968 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434704112337011, + "loss": 1.5467, + "theoretical_loss": 3.9792073883672385, + "tokens_seen": 459669504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043469408224674024, + "loss": 1.6051, + "theoretical_loss": 3.9791458433987503, + "tokens_seen": 459735040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346840521564694, + "loss": 1.6442, + "theoretical_loss": 3.97908430965911, + "tokens_seen": 459800576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346740220661986, + "loss": 1.6545, + "theoretical_loss": 3.979022787144671, + "tokens_seen": 459866112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043466399197592784, + "loss": 1.6252, + "theoretical_loss": 3.9789612758517836, + "tokens_seen": 459931648 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043465396188565697, + "loss": 1.6396, + "theoretical_loss": 3.9788997757768043, + "tokens_seen": 459997184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004346439317953862, + "loss": 1.6554, + "theoretical_loss": 3.9788382869160897, + "tokens_seen": 460062720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043463390170511533, + "loss": 1.619, + "theoretical_loss": 3.9787768092659963, + "tokens_seen": 460128256 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043462387161484457, + "loss": 1.6093, + "theoretical_loss": 3.978715342822885, + "tokens_seen": 460193792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043461384152457375, + "loss": 1.6679, + "theoretical_loss": 3.9786538875831168, + "tokens_seen": 460259328 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043460381143430293, + "loss": 1.6886, + "theoretical_loss": 3.9785924435430555, + "tokens_seen": 460324864 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345937813440321, + "loss": 1.5585, + "theoretical_loss": 3.978531010699066, + "tokens_seen": 460390400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345837512537613, + "loss": 1.5772, + "theoretical_loss": 3.9784695890475144, + "tokens_seen": 460455936 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043457372116349047, + "loss": 1.6245, + "theoretical_loss": 3.9784081785847687, + "tokens_seen": 460521472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345636910732197, + "loss": 1.5601, + "theoretical_loss": 3.9783467793071994, + "tokens_seen": 460587008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043455366098294883, + "loss": 1.5354, + "theoretical_loss": 3.978285391211178, + "tokens_seen": 460652544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043454363089267807, + "loss": 1.5741, + "theoretical_loss": 3.9782240142930787, + "tokens_seen": 460718080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043453360080240725, + "loss": 1.5992, + "theoretical_loss": 3.978162648549275, + "tokens_seen": 460783616 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043452357071213643, + "loss": 1.5587, + "theoretical_loss": 3.9781012939761453, + "tokens_seen": 460849152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345135406218656, + "loss": 1.5626, + "theoretical_loss": 3.978039950570067, + "tokens_seen": 460914688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004345035105315948, + "loss": 1.6752, + "theoretical_loss": 3.9779786183274197, + "tokens_seen": 460980224 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434493480441324, + "loss": 1.6138, + "theoretical_loss": 3.977917297244587, + "tokens_seen": 461045760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344834503510532, + "loss": 1.6834, + "theoretical_loss": 3.9778559873179504, + "tokens_seen": 461111296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043447342026078234, + "loss": 1.7465, + "theoretical_loss": 3.977794688543897, + "tokens_seen": 461176832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043446339017051157, + "loss": 1.53, + "theoretical_loss": 3.9777334009188117, + "tokens_seen": 461242368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344533600802407, + "loss": 1.5874, + "theoretical_loss": 3.9776721244390845, + "tokens_seen": 461307904 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043444332998996993, + "loss": 1.6339, + "theoretical_loss": 3.9776108591011052, + "tokens_seen": 461373440 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344332998996991, + "loss": 1.5365, + "theoretical_loss": 3.9775496049012657, + "tokens_seen": 461438976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344232698094283, + "loss": 1.5289, + "theoretical_loss": 3.9774883618359596, + "tokens_seen": 461504512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344132397191575, + "loss": 1.5503, + "theoretical_loss": 3.9774271299015824, + "tokens_seen": 461570048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004344032096288867, + "loss": 1.4554, + "theoretical_loss": 3.977365909094531, + "tokens_seen": 461635584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043439317953861584, + "loss": 1.7713, + "theoretical_loss": 3.977304699411204, + "tokens_seen": 461701120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343831494483451, + "loss": 1.7093, + "theoretical_loss": 3.977243500848001, + "tokens_seen": 461766656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343731193580742, + "loss": 1.626, + "theoretical_loss": 3.977182313401325, + "tokens_seen": 461832192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043436308926780344, + "loss": 1.6955, + "theoretical_loss": 3.977121137067579, + "tokens_seen": 461897728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343530591775326, + "loss": 1.5605, + "theoretical_loss": 3.9770599718431687, + "tokens_seen": 461963264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 174193, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7412657737731934, + "objective/train/theoretical_loss": 3.9769988177245015, + "objective/train/tokens_used": 482488800, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343430290872618, + "loss": 1.7028, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434332998996991, + "loss": 1.6875, + "theoretical_loss": 3.976937674707985, + "tokens_seen": 462094336 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043432296890672016, + "loss": 1.7276, + "theoretical_loss": 3.9768765427900306, + "tokens_seen": 462159872 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043431293881644934, + "loss": 1.4686, + "theoretical_loss": 3.97681542196705, + "tokens_seen": 462225408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004343029087261786, + "loss": 1.5656, + "theoretical_loss": 3.976754312235456, + "tokens_seen": 462290944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342928786359077, + "loss": 1.5854, + "theoretical_loss": 3.9766932135916653, + "tokens_seen": 462356480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043428284854563694, + "loss": 1.6845, + "theoretical_loss": 3.9766321260320945, + "tokens_seen": 462422016 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043427281845536607, + "loss": 1.5872, + "theoretical_loss": 3.976571049553162, + "tokens_seen": 462487552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342627883650953, + "loss": 1.6063, + "theoretical_loss": 3.9765099841512885, + "tokens_seen": 462553088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342527582748245, + "loss": 1.5904, + "theoretical_loss": 3.976448929822895, + "tokens_seen": 462618624 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043424272818455367, + "loss": 1.4805, + "theoretical_loss": 3.9763878865644067, + "tokens_seen": 462684160 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043423269809428285, + "loss": 1.5745, + "theoretical_loss": 3.9763268543722483, + "tokens_seen": 462749696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342226680040121, + "loss": 1.5525, + "theoretical_loss": 3.976265833242846, + "tokens_seen": 462815232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004342126379137412, + "loss": 1.6279, + "theoretical_loss": 3.9762048231726297, + "tokens_seen": 462880768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043420260782347044, + "loss": 1.7385, + "theoretical_loss": 3.9761438241580294, + "tokens_seen": 462946304 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043419257773319957, + "loss": 1.6089, + "theoretical_loss": 3.976082836195477, + "tokens_seen": 463011840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341825476429288, + "loss": 1.594, + "theoretical_loss": 3.976021859281406, + "tokens_seen": 463077376 + }, + { + "epoch": 0.14, + "learning_rate": 0.000434172517552658, + "loss": 1.5927, + "theoretical_loss": 3.975960893412251, + "tokens_seen": 463142912 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043416248746238717, + "loss": 1.5093, + "theoretical_loss": 3.9758999385844502, + "tokens_seen": 463208448 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043415245737211635, + "loss": 1.5046, + "theoretical_loss": 3.975838994794441, + "tokens_seen": 463273984 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043414242728184553, + "loss": 1.5561, + "theoretical_loss": 3.975778062038665, + "tokens_seen": 463339520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341323971915747, + "loss": 1.7012, + "theoretical_loss": 3.975717140313564, + "tokens_seen": 463405056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043412236710130395, + "loss": 1.6852, + "theoretical_loss": 3.9756562296155797, + "tokens_seen": 463470592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341123370110331, + "loss": 1.4593, + "theoretical_loss": 3.975595329941159, + "tokens_seen": 463536128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004341023069207623, + "loss": 1.5732, + "theoretical_loss": 3.975534441286748, + "tokens_seen": 463601664 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043409227683049144, + "loss": 1.5388, + "theoretical_loss": 3.975473563648796, + "tokens_seen": 463667200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043408224674022067, + "loss": 1.7195, + "theoretical_loss": 3.975412697023753, + "tokens_seen": 463732736 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043407221664994985, + "loss": 1.5805, + "theoretical_loss": 3.9753518414080693, + "tokens_seen": 463798272 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043406218655967903, + "loss": 1.5523, + "theoretical_loss": 3.9752909967982, + "tokens_seen": 463863808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340521564694082, + "loss": 1.6846, + "theoretical_loss": 3.9752301631906, + "tokens_seen": 463929344 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043404212637913745, + "loss": 1.5817, + "theoretical_loss": 3.9751693405817248, + "tokens_seen": 463994880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340320962888666, + "loss": 1.7416, + "theoretical_loss": 3.9751085289680343, + "tokens_seen": 464060416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340220661985958, + "loss": 1.8028, + "theoretical_loss": 3.975047728345987, + "tokens_seen": 464125952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043401203610832494, + "loss": 1.695, + "theoretical_loss": 3.9749869387120462, + "tokens_seen": 464191488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004340020060180542, + "loss": 1.5114, + "theoretical_loss": 3.974926160062674, + "tokens_seen": 464257024 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043399197592778336, + "loss": 1.578, + "theoretical_loss": 3.9748653923943356, + "tokens_seen": 464322560 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043398194583751254, + "loss": 1.7331, + "theoretical_loss": 3.9748046357034976, + "tokens_seen": 464388096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339719157472417, + "loss": 1.6143, + "theoretical_loss": 3.9747438899866276, + "tokens_seen": 464453632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339618856569709, + "loss": 1.5538, + "theoretical_loss": 3.9746831552401964, + "tokens_seen": 464519168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339518555667001, + "loss": 1.5601, + "theoretical_loss": 3.9746224314606753, + "tokens_seen": 464584704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339418254764293, + "loss": 1.5978, + "theoretical_loss": 3.9745617186445372, + "tokens_seen": 464650240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339317953861585, + "loss": 1.7158, + "theoretical_loss": 3.974501016788256, + "tokens_seen": 464715776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339217652958877, + "loss": 1.7198, + "theoretical_loss": 3.9744403258883096, + "tokens_seen": 464781312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004339117352056169, + "loss": 1.5176, + "theoretical_loss": 3.974379645941174, + "tokens_seen": 464846848 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043390170511534604, + "loss": 1.5724, + "theoretical_loss": 3.974318976943331, + "tokens_seen": 464912384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338916750250753, + "loss": 1.6569, + "theoretical_loss": 3.974258318891261, + "tokens_seen": 464977920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338816449348044, + "loss": 1.521, + "theoretical_loss": 3.974197671781446, + "tokens_seen": 465043456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043387161484453364, + "loss": 1.5281, + "theoretical_loss": 3.9741370356103713, + "tokens_seen": 465108992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338615847542628, + "loss": 1.5878, + "theoretical_loss": 3.9740764103745225, + "tokens_seen": 465174528 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433851554663992, + "loss": 1.6216, + "theoretical_loss": 3.974015796070388, + "tokens_seen": 465240064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 175471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3623638153076172, + "objective/train/theoretical_loss": 3.973955192694457, + "objective/train/tokens_used": 485765600, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338415245737212, + "loss": 1.4748, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043383149448345036, + "loss": 1.5396, + "theoretical_loss": 3.97389460024322, + "tokens_seen": 465371136 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043382146439317954, + "loss": 1.5843, + "theoretical_loss": 3.9738340187131698, + "tokens_seen": 465436672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338114343029088, + "loss": 1.6343, + "theoretical_loss": 3.973773448100801, + "tokens_seen": 465502208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004338014042126379, + "loss": 1.5577, + "theoretical_loss": 3.973712888402609, + "tokens_seen": 465567744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043379137412236714, + "loss": 1.6211, + "theoretical_loss": 3.9736523396150916, + "tokens_seen": 465633280 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043378134403209627, + "loss": 1.5075, + "theoretical_loss": 3.973591801734748, + "tokens_seen": 465698816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337713139418255, + "loss": 1.6293, + "theoretical_loss": 3.9735312747580775, + "tokens_seen": 465764352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337612838515547, + "loss": 1.6833, + "theoretical_loss": 3.973470758681584, + "tokens_seen": 465829888 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043375125376128387, + "loss": 1.6023, + "theoretical_loss": 3.9734102535017715, + "tokens_seen": 465895424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043374122367101305, + "loss": 1.5984, + "theoretical_loss": 3.9733497592151443, + "tokens_seen": 465960960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337311935807423, + "loss": 1.4795, + "theoretical_loss": 3.9732892758182103, + "tokens_seen": 466026496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004337211634904714, + "loss": 1.6089, + "theoretical_loss": 3.973228803307478, + "tokens_seen": 466092032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043371113340020064, + "loss": 1.434, + "theoretical_loss": 3.9731683416794583, + "tokens_seen": 466157568 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043370110330992977, + "loss": 1.5861, + "theoretical_loss": 3.9731078909306623, + "tokens_seen": 466223104 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433691073219659, + "loss": 1.7452, + "theoretical_loss": 3.9730474510576044, + "tokens_seen": 466288640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336810431293882, + "loss": 1.6368, + "theoretical_loss": 3.9729870220567993, + "tokens_seen": 466354176 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043367101303911737, + "loss": 1.6641, + "theoretical_loss": 3.9729266039247637, + "tokens_seen": 466419712 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043366098294884655, + "loss": 1.5839, + "theoretical_loss": 3.9728661966580163, + "tokens_seen": 466485248 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043365095285857573, + "loss": 1.5453, + "theoretical_loss": 3.9728058002530773, + "tokens_seen": 466550784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336409227683049, + "loss": 1.641, + "theoretical_loss": 3.972745414706468, + "tokens_seen": 466616320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043363089267803415, + "loss": 1.5115, + "theoretical_loss": 3.972685040014711, + "tokens_seen": 466681856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336208625877633, + "loss": 1.4682, + "theoretical_loss": 3.9726246761743322, + "tokens_seen": 466747392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004336108324974925, + "loss": 1.5479, + "theoretical_loss": 3.9725643231818575, + "tokens_seen": 466812928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043360080240722164, + "loss": 1.6214, + "theoretical_loss": 3.9725039810338147, + "tokens_seen": 466878464 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043359077231695087, + "loss": 1.64, + "theoretical_loss": 3.972443649726734, + "tokens_seen": 466944000 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043358074222668005, + "loss": 1.5694, + "theoretical_loss": 3.9723833292571458, + "tokens_seen": 467009536 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043357071213640923, + "loss": 1.5008, + "theoretical_loss": 3.9723230196215837, + "tokens_seen": 467075072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335606820461384, + "loss": 1.692, + "theoretical_loss": 3.9722627208165817, + "tokens_seen": 467140608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043355065195586765, + "loss": 1.5771, + "theoretical_loss": 3.9722024328386754, + "tokens_seen": 467206144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335406218655968, + "loss": 1.5436, + "theoretical_loss": 3.9721421556844034, + "tokens_seen": 467271680 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433530591775326, + "loss": 1.6146, + "theoretical_loss": 3.972081889350304, + "tokens_seen": 467337216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043352056168505514, + "loss": 1.6, + "theoretical_loss": 3.9720216338329184, + "tokens_seen": 467402752 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004335105315947844, + "loss": 1.6308, + "theoretical_loss": 3.971961389128789, + "tokens_seen": 467468288 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043350050150451356, + "loss": 1.5816, + "theoretical_loss": 3.97190115523446, + "tokens_seen": 467533824 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043349047141424274, + "loss": 1.5367, + "theoretical_loss": 3.9718409321464754, + "tokens_seen": 467599360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334804413239719, + "loss": 1.6589, + "theoretical_loss": 3.9717807198613837, + "tokens_seen": 467664896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334704112337011, + "loss": 1.6935, + "theoretical_loss": 3.971720518375734, + "tokens_seen": 467730432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334603811434303, + "loss": 1.622, + "theoretical_loss": 3.971660327686075, + "tokens_seen": 467795968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334503510531595, + "loss": 1.5104, + "theoretical_loss": 3.97160014778896, + "tokens_seen": 467861504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043344032096288864, + "loss": 1.6781, + "theoretical_loss": 3.9715399786809424, + "tokens_seen": 467927040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334302908726179, + "loss": 1.5574, + "theoretical_loss": 3.971479820358576, + "tokens_seen": 467992576 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433420260782347, + "loss": 1.5056, + "theoretical_loss": 3.9714196728184197, + "tokens_seen": 468058112 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043341023069207624, + "loss": 1.5213, + "theoretical_loss": 3.971359536057029, + "tokens_seen": 468123648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004334002006018054, + "loss": 1.5767, + "theoretical_loss": 3.9712994100709653, + "tokens_seen": 468189184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333901705115346, + "loss": 1.5915, + "theoretical_loss": 3.97123929485679, + "tokens_seen": 468254720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333801404212638, + "loss": 1.5784, + "theoretical_loss": 3.971179190411066, + "tokens_seen": 468320256 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433370110330993, + "loss": 1.5692, + "theoretical_loss": 3.9711190967303565, + "tokens_seen": 468385792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043336008024072215, + "loss": 1.5848, + "theoretical_loss": 3.971059013811229, + "tokens_seen": 468451328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333500501504514, + "loss": 1.5109, + "theoretical_loss": 3.9709989416502514, + "tokens_seen": 468516864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 176758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.715922236442566, + "objective/train/theoretical_loss": 3.9709388802439918, + "objective/train/tokens_used": 489042400, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333400200601805, + "loss": 1.6791, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043332998996990974, + "loss": 1.614, + "theoretical_loss": 3.9708788295890223, + "tokens_seen": 468647936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333199598796389, + "loss": 1.6626, + "theoretical_loss": 3.9708187896819136, + "tokens_seen": 468713472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004333099297893681, + "loss": 1.6496, + "theoretical_loss": 3.9707587605192414, + "tokens_seen": 468779008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332998996990973, + "loss": 1.5708, + "theoretical_loss": 3.9706987420975803, + "tokens_seen": 468844544 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043328986960882647, + "loss": 1.5956, + "theoretical_loss": 3.970638734413508, + "tokens_seen": 468910080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043327983951855565, + "loss": 1.6064, + "theoretical_loss": 3.970578737463603, + "tokens_seen": 468975616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332698094282849, + "loss": 1.6454, + "theoretical_loss": 3.9705187512444446, + "tokens_seen": 469041152 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433259779338014, + "loss": 1.6242, + "theoretical_loss": 3.9704587757526157, + "tokens_seen": 469106688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043324974924774325, + "loss": 1.6879, + "theoretical_loss": 3.9703988109846997, + "tokens_seen": 469172224 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043323971915747243, + "loss": 1.5462, + "theoretical_loss": 3.970338856937281, + "tokens_seen": 469237760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332296890672016, + "loss": 1.6001, + "theoretical_loss": 3.970278913606947, + "tokens_seen": 469303296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004332196589769308, + "loss": 1.626, + "theoretical_loss": 3.9702189809902846, + "tokens_seen": 469368832 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043320962888665997, + "loss": 1.6193, + "theoretical_loss": 3.9701590590838842, + "tokens_seen": 469434368 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043319959879638915, + "loss": 1.5291, + "theoretical_loss": 3.970099147884336, + "tokens_seen": 469499904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331895687061184, + "loss": 1.5658, + "theoretical_loss": 3.9700392473882347, + "tokens_seen": 469565440 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043317953861584757, + "loss": 1.6431, + "theoretical_loss": 3.969979357592173, + "tokens_seen": 469630976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043316950852557675, + "loss": 1.632, + "theoretical_loss": 3.969919478492747, + "tokens_seen": 469696512 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043315947843530593, + "loss": 1.6015, + "theoretical_loss": 3.9698596100865546, + "tokens_seen": 469762048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331494483450351, + "loss": 1.6542, + "theoretical_loss": 3.9697997523701947, + "tokens_seen": 469827584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043313941825476435, + "loss": 1.5888, + "theoretical_loss": 3.969739905340268, + "tokens_seen": 469893120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331293881644935, + "loss": 1.6825, + "theoretical_loss": 3.9696800689933758, + "tokens_seen": 469958656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004331193580742227, + "loss": 1.654, + "theoretical_loss": 3.9696202433261227, + "tokens_seen": 470024192 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043310932798395184, + "loss": 1.6159, + "theoretical_loss": 3.9695604283351136, + "tokens_seen": 470089728 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043309929789368107, + "loss": 1.5335, + "theoretical_loss": 3.9695006240169546, + "tokens_seen": 470155264 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043308926780341025, + "loss": 1.5791, + "theoretical_loss": 3.9694408303682556, + "tokens_seen": 470220800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043307923771313943, + "loss": 1.5515, + "theoretical_loss": 3.9693810473856246, + "tokens_seen": 470286336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330692076228686, + "loss": 1.5528, + "theoretical_loss": 3.969321275065674, + "tokens_seen": 470351872 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043305917753259785, + "loss": 1.65, + "theoretical_loss": 3.969261513405017, + "tokens_seen": 470417408 + }, + { + "epoch": 0.14, + "learning_rate": 0.000433049147442327, + "loss": 1.6879, + "theoretical_loss": 3.9692017624002682, + "tokens_seen": 470482944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330391173520562, + "loss": 1.5898, + "theoretical_loss": 3.9691420220480422, + "tokens_seen": 470548480 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043302908726178534, + "loss": 1.497, + "theoretical_loss": 3.9690822923449582, + "tokens_seen": 470614016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004330190571715146, + "loss": 1.6901, + "theoretical_loss": 3.969022573287635, + "tokens_seen": 470679552 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043300902708124376, + "loss": 1.5434, + "theoretical_loss": 3.968962864872693, + "tokens_seen": 470745088 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043299899699097294, + "loss": 1.6143, + "theoretical_loss": 3.9689031670967543, + "tokens_seen": 470810624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329889669007021, + "loss": 1.6104, + "theoretical_loss": 3.9688434799564427, + "tokens_seen": 470876160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329789368104313, + "loss": 1.5799, + "theoretical_loss": 3.9687838034483844, + "tokens_seen": 470941696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329689067201605, + "loss": 1.5479, + "theoretical_loss": 3.9687241375692053, + "tokens_seen": 471007232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329588766298897, + "loss": 1.6066, + "theoretical_loss": 3.9686644823155337, + "tokens_seen": 471072768 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043294884653961884, + "loss": 1.6939, + "theoretical_loss": 3.9686048376840004, + "tokens_seen": 471138304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329388164493481, + "loss": 1.6934, + "theoretical_loss": 3.968545203671236, + "tokens_seen": 471203840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329287863590772, + "loss": 1.602, + "theoretical_loss": 3.9684855802738745, + "tokens_seen": 471269376 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043291875626880644, + "loss": 1.6302, + "theoretical_loss": 3.9684259674885496, + "tokens_seen": 471334912 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004329087261785356, + "loss": 1.4893, + "theoretical_loss": 3.9683663653118977, + "tokens_seen": 471400448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328986960882648, + "loss": 1.5726, + "theoretical_loss": 3.968306773740556, + "tokens_seen": 471465984 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432888665997994, + "loss": 1.7182, + "theoretical_loss": 3.9682471927711642, + "tokens_seen": 471531520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328786359077232, + "loss": 1.5456, + "theoretical_loss": 3.9681876224003627, + "tokens_seen": 471597056 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043286860581745235, + "loss": 1.7013, + "theoretical_loss": 3.9681280626247943, + "tokens_seen": 471662592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328585757271816, + "loss": 1.6266, + "theoretical_loss": 3.9680685134411022, + "tokens_seen": 471728128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328485456369107, + "loss": 1.581, + "theoretical_loss": 3.9680089748459313, + "tokens_seen": 471793664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 177948, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6643733978271484, + "objective/train/theoretical_loss": 3.967949446835929, + "objective/train/tokens_used": 492319200, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043283851554663994, + "loss": 1.6318, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328284854563691, + "loss": 1.4491, + "theoretical_loss": 3.967889929407743, + "tokens_seen": 471924736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328184553660983, + "loss": 1.5843, + "theoretical_loss": 3.9678304225580248, + "tokens_seen": 471990272 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004328084252758275, + "loss": 1.4757, + "theoretical_loss": 3.967770926283424, + "tokens_seen": 472055808 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043279839518555667, + "loss": 1.546, + "theoretical_loss": 3.9677114405805938, + "tokens_seen": 472121344 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043278836509528585, + "loss": 1.5218, + "theoretical_loss": 3.967651965446189, + "tokens_seen": 472186880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327783350050151, + "loss": 1.6447, + "theoretical_loss": 3.9675925008768655, + "tokens_seen": 472252416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327683049147442, + "loss": 1.6105, + "theoretical_loss": 3.9675330468692804, + "tokens_seen": 472317952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043275827482447345, + "loss": 1.5399, + "theoretical_loss": 3.9674736034200935, + "tokens_seen": 472383488 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043274824473420263, + "loss": 1.6563, + "theoretical_loss": 3.967414170525965, + "tokens_seen": 472449024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004327382146439318, + "loss": 1.6214, + "theoretical_loss": 3.9673547481835567, + "tokens_seen": 472514560 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432728184553661, + "loss": 1.6284, + "theoretical_loss": 3.9672953363895322, + "tokens_seen": 472580096 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043271815446339017, + "loss": 1.5824, + "theoretical_loss": 3.9672359351405566, + "tokens_seen": 472645632 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043270812437311935, + "loss": 1.6748, + "theoretical_loss": 3.967176544433296, + "tokens_seen": 472711168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326980942828486, + "loss": 1.6254, + "theoretical_loss": 3.96711716426442, + "tokens_seen": 472776704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326880641925777, + "loss": 1.472, + "theoretical_loss": 3.9670577946305965, + "tokens_seen": 472842240 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043267803410230695, + "loss": 1.7315, + "theoretical_loss": 3.9669984355284975, + "tokens_seen": 472907776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326680040120361, + "loss": 1.5911, + "theoretical_loss": 3.9669390869547954, + "tokens_seen": 472973312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326579739217653, + "loss": 1.5755, + "theoretical_loss": 3.966879748906165, + "tokens_seen": 473038848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326479438314945, + "loss": 1.4913, + "theoretical_loss": 3.966820421379281, + "tokens_seen": 473104384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326379137412237, + "loss": 1.6733, + "theoretical_loss": 3.966761104370821, + "tokens_seen": 473169920 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043262788365095286, + "loss": 1.5374, + "theoretical_loss": 3.9667017978774632, + "tokens_seen": 473235456 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043261785356068204, + "loss": 1.655, + "theoretical_loss": 3.966642501895889, + "tokens_seen": 473300992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004326078234704112, + "loss": 1.7306, + "theoretical_loss": 3.9665832164227792, + "tokens_seen": 473366528 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043259779338014045, + "loss": 1.5821, + "theoretical_loss": 3.966523941454817, + "tokens_seen": 473432064 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325877632898696, + "loss": 1.6172, + "theoretical_loss": 3.9664646769886875, + "tokens_seen": 473497600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325777331995988, + "loss": 1.632, + "theoretical_loss": 3.966405423021077, + "tokens_seen": 473563136 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432567703109328, + "loss": 1.5138, + "theoretical_loss": 3.9663461795486725, + "tokens_seen": 473628672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325576730190572, + "loss": 1.5873, + "theoretical_loss": 3.9662869465681636, + "tokens_seen": 473694208 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043254764292878636, + "loss": 1.6017, + "theoretical_loss": 3.9662277240762416, + "tokens_seen": 473759744 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043253761283851554, + "loss": 1.6438, + "theoretical_loss": 3.9661685120695975, + "tokens_seen": 473825280 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325275827482447, + "loss": 1.5406, + "theoretical_loss": 3.9661093105449265, + "tokens_seen": 473890816 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043251755265797396, + "loss": 1.6186, + "theoretical_loss": 3.966050119498923, + "tokens_seen": 473956352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004325075225677031, + "loss": 1.5102, + "theoretical_loss": 3.9659909389282832, + "tokens_seen": 474021888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324974924774323, + "loss": 1.6362, + "theoretical_loss": 3.9659317688297064, + "tokens_seen": 474087424 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043248746238716145, + "loss": 1.5893, + "theoretical_loss": 3.965872609199892, + "tokens_seen": 474152960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324774322968907, + "loss": 1.5123, + "theoretical_loss": 3.965813460035541, + "tokens_seen": 474218496 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043246740220661986, + "loss": 1.6472, + "theoretical_loss": 3.965754321333356, + "tokens_seen": 474284032 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043245737211634904, + "loss": 1.6755, + "theoretical_loss": 3.965695193090042, + "tokens_seen": 474349568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324473420260782, + "loss": 1.6111, + "theoretical_loss": 3.9656360753023034, + "tokens_seen": 474415104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324373119358074, + "loss": 1.6335, + "theoretical_loss": 3.9655769679668484, + "tokens_seen": 474480640 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043242728184553664, + "loss": 1.6962, + "theoretical_loss": 3.965517871080386, + "tokens_seen": 474546176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004324172517552658, + "loss": 1.5839, + "theoretical_loss": 3.965458784639625, + "tokens_seen": 474611712 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432407221664995, + "loss": 1.7632, + "theoretical_loss": 3.965399708641278, + "tokens_seen": 474677248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323971915747242, + "loss": 1.6459, + "theoretical_loss": 3.9653406430820586, + "tokens_seen": 474742784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323871614844534, + "loss": 1.5705, + "theoretical_loss": 3.9652815879586805, + "tokens_seen": 474808320 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043237713139418255, + "loss": 1.5003, + "theoretical_loss": 3.96522254326786, + "tokens_seen": 474873856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323671013039118, + "loss": 1.6802, + "theoretical_loss": 3.9651635090063153, + "tokens_seen": 474939392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323570712136409, + "loss": 1.5787, + "theoretical_loss": 3.965104485170765, + "tokens_seen": 475004928 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043234704112337014, + "loss": 1.643, + "theoretical_loss": 3.9650454717579295, + "tokens_seen": 475070464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 179087, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.485135793685913, + "objective/train/theoretical_loss": 3.9649864687645318, + "objective/train/tokens_used": 495596000, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323370110330993, + "loss": 1.4941, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323269809428285, + "loss": 1.559, + "theoretical_loss": 3.9649274761872944, + "tokens_seen": 475201536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004323169508525577, + "loss": 1.5752, + "theoretical_loss": 3.9648684940229426, + "tokens_seen": 475267072 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043230692076228687, + "loss": 1.5772, + "theoretical_loss": 3.9648095222682036, + "tokens_seen": 475332608 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043229689067201605, + "loss": 1.597, + "theoretical_loss": 3.9647505609198044, + "tokens_seen": 475398144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322868605817453, + "loss": 1.7557, + "theoretical_loss": 3.9646916099744756, + "tokens_seen": 475463680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322768304914744, + "loss": 1.6273, + "theoretical_loss": 3.9646326694289473, + "tokens_seen": 475529216 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043226680040120365, + "loss": 1.6725, + "theoretical_loss": 3.9645737392799516, + "tokens_seen": 475594752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043225677031093283, + "loss": 1.5569, + "theoretical_loss": 3.964514819524224, + "tokens_seen": 475660288 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432246740220662, + "loss": 1.5537, + "theoretical_loss": 3.964455910158498, + "tokens_seen": 475725824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322367101303912, + "loss": 1.5135, + "theoretical_loss": 3.9643970111795115, + "tokens_seen": 475791360 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043222668004012037, + "loss": 1.6534, + "theoretical_loss": 3.964338122584003, + "tokens_seen": 475856896 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043221664994984955, + "loss": 1.5124, + "theoretical_loss": 3.964279244368712, + "tokens_seen": 475922432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004322066198595788, + "loss": 1.5244, + "theoretical_loss": 3.9642203765303794, + "tokens_seen": 475987968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321965897693079, + "loss": 1.4405, + "theoretical_loss": 3.9641615190657484, + "tokens_seen": 476053504 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043218655967903715, + "loss": 1.6238, + "theoretical_loss": 3.9641026719715633, + "tokens_seen": 476119040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321765295887663, + "loss": 1.5812, + "theoretical_loss": 3.9640438352445697, + "tokens_seen": 476184576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321664994984955, + "loss": 1.6401, + "theoretical_loss": 3.9639850088815143, + "tokens_seen": 476250112 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321564694082247, + "loss": 1.548, + "theoretical_loss": 3.9639261928791463, + "tokens_seen": 476315648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321464393179539, + "loss": 1.5307, + "theoretical_loss": 3.963867387234216, + "tokens_seen": 476381184 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043213640922768306, + "loss": 1.5158, + "theoretical_loss": 3.963808591943474, + "tokens_seen": 476446720 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043212637913741224, + "loss": 1.5446, + "theoretical_loss": 3.9637498070036745, + "tokens_seen": 476512256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004321163490471414, + "loss": 1.6055, + "theoretical_loss": 3.963691032411571, + "tokens_seen": 476577792 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043210631895687065, + "loss": 1.735, + "theoretical_loss": 3.9636322681639196, + "tokens_seen": 476643328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320962888665998, + "loss": 1.4905, + "theoretical_loss": 3.9635735142574786, + "tokens_seen": 476708864 + }, + { + "epoch": 0.14, + "learning_rate": 0.000432086258776329, + "loss": 1.6796, + "theoretical_loss": 3.963514770689006, + "tokens_seen": 476774400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320762286860582, + "loss": 1.5147, + "theoretical_loss": 3.9634560374552623, + "tokens_seen": 476839936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320661985957874, + "loss": 1.5999, + "theoretical_loss": 3.96339731455301, + "tokens_seen": 476905472 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043205616850551656, + "loss": 1.4423, + "theoretical_loss": 3.963338601979011, + "tokens_seen": 476971008 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043204613841524574, + "loss": 1.5623, + "theoretical_loss": 3.9632798997300314, + "tokens_seen": 477036544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320361083249749, + "loss": 1.6519, + "theoretical_loss": 3.963221207802836, + "tokens_seen": 477102080 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043202607823470416, + "loss": 1.4346, + "theoretical_loss": 3.9631625261941945, + "tokens_seen": 477167616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320160481444333, + "loss": 1.5458, + "theoretical_loss": 3.963103854900874, + "tokens_seen": 477233152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004320060180541625, + "loss": 1.5706, + "theoretical_loss": 3.963045193919646, + "tokens_seen": 477298688 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043199598796389165, + "loss": 1.5021, + "theoretical_loss": 3.9629865432472817, + "tokens_seen": 477364224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319859578736209, + "loss": 1.6098, + "theoretical_loss": 3.9629279028805553, + "tokens_seen": 477429760 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043197592778335006, + "loss": 1.687, + "theoretical_loss": 3.962869272816242, + "tokens_seen": 477495296 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043196589769307924, + "loss": 1.6296, + "theoretical_loss": 3.9628106530511173, + "tokens_seen": 477560832 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319558676028084, + "loss": 1.6486, + "theoretical_loss": 3.9627520435819594, + "tokens_seen": 477626368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319458375125376, + "loss": 1.6086, + "theoretical_loss": 3.9626934444055477, + "tokens_seen": 477691904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319358074222668, + "loss": 1.5719, + "theoretical_loss": 3.962634855518663, + "tokens_seen": 477757440 + }, + { + "epoch": 0.14, + "learning_rate": 0.000431925777331996, + "loss": 1.515, + "theoretical_loss": 3.962576276918087, + "tokens_seen": 477822976 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043191574724172515, + "loss": 1.611, + "theoretical_loss": 3.962517708600603, + "tokens_seen": 477888512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004319057171514544, + "loss": 1.6405, + "theoretical_loss": 3.9624591505629976, + "tokens_seen": 477954048 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043189568706118357, + "loss": 1.6774, + "theoretical_loss": 3.962400602802056, + "tokens_seen": 478019584 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043188565697091275, + "loss": 1.6135, + "theoretical_loss": 3.962342065314566, + "tokens_seen": 478085120 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043187562688064193, + "loss": 1.5613, + "theoretical_loss": 3.962283538097318, + "tokens_seen": 478150656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318655967903711, + "loss": 1.5943, + "theoretical_loss": 3.962225021147102, + "tokens_seen": 478216192 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318555667001003, + "loss": 1.4936, + "theoretical_loss": 3.9621665144607103, + "tokens_seen": 478281728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318455366098295, + "loss": 1.6194, + "theoretical_loss": 3.9621080180349373, + "tokens_seen": 478347264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 180329, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6273953914642334, + "objective/train/theoretical_loss": 3.9620495318665774, + "objective/train/tokens_used": 498872800, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.00043183550651955865, + "loss": 1.627, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0004318254764292879, + "loss": 1.7117, + "theoretical_loss": 3.9619910559524274, + "tokens_seen": 478478336 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431815446339017, + "loss": 1.5235, + "theoretical_loss": 3.9619325902892855, + "tokens_seen": 478543872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043180541624874625, + "loss": 1.5848, + "theoretical_loss": 3.961874134873951, + "tokens_seen": 478609408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043179538615847543, + "loss": 1.5878, + "theoretical_loss": 3.9618156897032257, + "tokens_seen": 478674944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317853560682046, + "loss": 1.6548, + "theoretical_loss": 3.961757254773911, + "tokens_seen": 478740480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317753259779338, + "loss": 1.5596, + "theoretical_loss": 3.9616988300828098, + "tokens_seen": 478806016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043176529588766303, + "loss": 1.6083, + "theoretical_loss": 3.9616404156267286, + "tokens_seen": 478871552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043175526579739215, + "loss": 1.4984, + "theoretical_loss": 3.961582011402475, + "tokens_seen": 478937088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317452357071214, + "loss": 1.5039, + "theoretical_loss": 3.9615236174068547, + "tokens_seen": 479002624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317352056168505, + "loss": 1.6335, + "theoretical_loss": 3.961465233636679, + "tokens_seen": 479068160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043172517552657975, + "loss": 1.6182, + "theoretical_loss": 3.9614068600887578, + "tokens_seen": 479133696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043171514543630893, + "loss": 1.5639, + "theoretical_loss": 3.9613484967599044, + "tokens_seen": 479199232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004317051153460381, + "loss": 1.4311, + "theoretical_loss": 3.961290143646932, + "tokens_seen": 479264768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316950852557673, + "loss": 1.6668, + "theoretical_loss": 3.961231800746656, + "tokens_seen": 479330304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316850551654965, + "loss": 1.5323, + "theoretical_loss": 3.961173468055893, + "tokens_seen": 479395840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316750250752257, + "loss": 1.4503, + "theoretical_loss": 3.961115145571461, + "tokens_seen": 479461376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316649949849549, + "loss": 1.5982, + "theoretical_loss": 3.96105683329018, + "tokens_seen": 479526912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316549648946841, + "loss": 1.579, + "theoretical_loss": 3.9609985312088707, + "tokens_seen": 479592448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043164493480441326, + "loss": 1.6274, + "theoretical_loss": 3.960940239324355, + "tokens_seen": 479657984 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043163490471414244, + "loss": 1.6166, + "theoretical_loss": 3.960881957633457, + "tokens_seen": 479723520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004316248746238716, + "loss": 1.6517, + "theoretical_loss": 3.9608236861330024, + "tokens_seen": 479789056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043161484453360085, + "loss": 1.5726, + "theoretical_loss": 3.960765424819817, + "tokens_seen": 479854592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043160481444333, + "loss": 1.5286, + "theoretical_loss": 3.9607071736907296, + "tokens_seen": 479920128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315947843530592, + "loss": 1.7319, + "theoretical_loss": 3.9606489327425694, + "tokens_seen": 479985664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315847542627884, + "loss": 1.6394, + "theoretical_loss": 3.960590701972167, + "tokens_seen": 480051200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315747241725176, + "loss": 1.5152, + "theoretical_loss": 3.960532481376355, + "tokens_seen": 480116736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043156469408224676, + "loss": 1.5483, + "theoretical_loss": 3.960474270951967, + "tokens_seen": 480182272 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043155466399197594, + "loss": 1.5129, + "theoretical_loss": 3.9604160706958385, + "tokens_seen": 480247808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315446339017051, + "loss": 1.6118, + "theoretical_loss": 3.960357880604806, + "tokens_seen": 480313344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043153460381143436, + "loss": 1.5545, + "theoretical_loss": 3.9602997006757072, + "tokens_seen": 480378880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315245737211635, + "loss": 1.639, + "theoretical_loss": 3.960241530905382, + "tokens_seen": 480444416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004315145436308927, + "loss": 1.5782, + "theoretical_loss": 3.96018337129067, + "tokens_seen": 480509952 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043150451354062185, + "loss": 1.6807, + "theoretical_loss": 3.960125221828415, + "tokens_seen": 480575488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314944834503511, + "loss": 1.6607, + "theoretical_loss": 3.9600670825154594, + "tokens_seen": 480641024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043148445336008026, + "loss": 1.6602, + "theoretical_loss": 3.960008953348649, + "tokens_seen": 480706560 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043147442326980944, + "loss": 1.5968, + "theoretical_loss": 3.9599508343248297, + "tokens_seen": 480772096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314643931795386, + "loss": 1.6885, + "theoretical_loss": 3.9598927254408505, + "tokens_seen": 480837632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314543630892678, + "loss": 1.629, + "theoretical_loss": 3.9598346266935596, + "tokens_seen": 480903168 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431444332998997, + "loss": 1.5604, + "theoretical_loss": 3.959776538079808, + "tokens_seen": 480968704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314343029087262, + "loss": 1.5629, + "theoretical_loss": 3.9597184595964476, + "tokens_seen": 481034240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043142427281845535, + "loss": 1.4866, + "theoretical_loss": 3.959660391240332, + "tokens_seen": 481099776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004314142427281846, + "loss": 1.5509, + "theoretical_loss": 3.9596023330083163, + "tokens_seen": 481165312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043140421263791377, + "loss": 1.555, + "theoretical_loss": 3.9595442848972566, + "tokens_seen": 481230848 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043139418254764295, + "loss": 1.648, + "theoretical_loss": 3.959486246904011, + "tokens_seen": 481296384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043138415245737213, + "loss": 1.5635, + "theoretical_loss": 3.959428219025438, + "tokens_seen": 481361920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313741223671013, + "loss": 1.6077, + "theoretical_loss": 3.9593702012583987, + "tokens_seen": 481427456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313640922768305, + "loss": 1.5835, + "theoretical_loss": 3.9593121935997546, + "tokens_seen": 481492992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313540621865597, + "loss": 1.547, + "theoretical_loss": 3.95925419604637, + "tokens_seen": 481558528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043134403209628885, + "loss": 1.5462, + "theoretical_loss": 3.959196208595108, + "tokens_seen": 481624064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 181505, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.577242374420166, + "objective/train/theoretical_loss": 3.959138231242836, + "objective/train/tokens_used": 502149600, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313340020060181, + "loss": 1.5371, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004313239719157472, + "loss": 1.6737, + "theoretical_loss": 3.959080263986421, + "tokens_seen": 481755136 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043131394182547645, + "loss": 1.5535, + "theoretical_loss": 3.9590223068227317, + "tokens_seen": 481820672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043130391173520563, + "loss": 1.5995, + "theoretical_loss": 3.9589643597486393, + "tokens_seen": 481886208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312938816449348, + "loss": 1.6605, + "theoretical_loss": 3.9589064227610145, + "tokens_seen": 481951744 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431283851554664, + "loss": 1.6033, + "theoretical_loss": 3.9588484958567314, + "tokens_seen": 482017280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043127382146439323, + "loss": 1.6751, + "theoretical_loss": 3.958790579032663, + "tokens_seen": 482082816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043126379137412236, + "loss": 1.6219, + "theoretical_loss": 3.958732672285687, + "tokens_seen": 482148352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312537612838516, + "loss": 1.6871, + "theoretical_loss": 3.95867477561268, + "tokens_seen": 482213888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312437311935807, + "loss": 1.5788, + "theoretical_loss": 3.95861688901052, + "tokens_seen": 482279424 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043123370110330995, + "loss": 1.4963, + "theoretical_loss": 3.958559012476088, + "tokens_seen": 482344960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043122367101303913, + "loss": 1.539, + "theoretical_loss": 3.958501146006265, + "tokens_seen": 482410496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312136409227683, + "loss": 1.6054, + "theoretical_loss": 3.958443289597934, + "tokens_seen": 482476032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004312036108324975, + "loss": 1.6055, + "theoretical_loss": 3.958385443247979, + "tokens_seen": 482541568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311935807422267, + "loss": 1.5608, + "theoretical_loss": 3.958327606953286, + "tokens_seen": 482607104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043118355065195586, + "loss": 1.608, + "theoretical_loss": 3.958269780710742, + "tokens_seen": 482672640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311735205616851, + "loss": 1.5692, + "theoretical_loss": 3.9582119645172344, + "tokens_seen": 482738176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311634904714142, + "loss": 1.4965, + "theoretical_loss": 3.9581541583696547, + "tokens_seen": 482803712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043115346038114346, + "loss": 1.578, + "theoretical_loss": 3.958096362264893, + "tokens_seen": 482869248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311434302908726, + "loss": 1.6303, + "theoretical_loss": 3.958038576199842, + "tokens_seen": 482934784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311334002006018, + "loss": 1.6521, + "theoretical_loss": 3.957980800171396, + "tokens_seen": 483000320 + }, + { + "epoch": 0.15, + "learning_rate": 0.000431123370110331, + "loss": 1.5955, + "theoretical_loss": 3.9579230341764493, + "tokens_seen": 483065856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004311133400200602, + "loss": 1.6194, + "theoretical_loss": 3.9578652782118997, + "tokens_seen": 483131392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043110330992978936, + "loss": 1.6027, + "theoretical_loss": 3.957807532274645, + "tokens_seen": 483196928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310932798395186, + "loss": 1.3501, + "theoretical_loss": 3.957749796361585, + "tokens_seen": 483262464 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310832497492477, + "loss": 1.5606, + "theoretical_loss": 3.9576920704696192, + "tokens_seen": 483328000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043107321965897696, + "loss": 1.5521, + "theoretical_loss": 3.957634354595651, + "tokens_seen": 483393536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310631895687061, + "loss": 1.5895, + "theoretical_loss": 3.957576648736584, + "tokens_seen": 483459072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310531594784353, + "loss": 1.5176, + "theoretical_loss": 3.9575189528893233, + "tokens_seen": 483524608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310431293881645, + "loss": 1.5317, + "theoretical_loss": 3.957461267050774, + "tokens_seen": 483590144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310330992978937, + "loss": 1.5939, + "theoretical_loss": 3.957403591217845, + "tokens_seen": 483655680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043102306920762286, + "loss": 1.6214, + "theoretical_loss": 3.9573459253874455, + "tokens_seen": 483721216 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043101303911735205, + "loss": 1.6115, + "theoretical_loss": 3.9572882695564853, + "tokens_seen": 483786752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004310030090270812, + "loss": 1.5349, + "theoretical_loss": 3.9572306237218764, + "tokens_seen": 483852288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043099297893681046, + "loss": 1.4879, + "theoretical_loss": 3.957172987880532, + "tokens_seen": 483917824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309829488465396, + "loss": 1.5778, + "theoretical_loss": 3.9571153620293673, + "tokens_seen": 483983360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309729187562688, + "loss": 1.6048, + "theoretical_loss": 3.9570577461652974, + "tokens_seen": 484048896 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043096288866599795, + "loss": 1.6672, + "theoretical_loss": 3.95700014028524, + "tokens_seen": 484114432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309528585757272, + "loss": 1.5129, + "theoretical_loss": 3.956942544386114, + "tokens_seen": 484179968 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043094282848545637, + "loss": 1.5813, + "theoretical_loss": 3.9568849584648396, + "tokens_seen": 484245504 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043093279839518555, + "loss": 1.5234, + "theoretical_loss": 3.9568273825183375, + "tokens_seen": 484311040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004309227683049148, + "loss": 1.6182, + "theoretical_loss": 3.9567698165435305, + "tokens_seen": 484376576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043091273821464397, + "loss": 1.7198, + "theoretical_loss": 3.9567122605373437, + "tokens_seen": 484442112 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043090270812437315, + "loss": 1.5264, + "theoretical_loss": 3.956654714496702, + "tokens_seen": 484507648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043089267803410233, + "loss": 1.5982, + "theoretical_loss": 3.9565971784185328, + "tokens_seen": 484573184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308826479438315, + "loss": 1.6076, + "theoretical_loss": 3.9565396522997633, + "tokens_seen": 484638720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308726178535607, + "loss": 1.577, + "theoretical_loss": 3.9564821361373244, + "tokens_seen": 484704256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308625877632899, + "loss": 1.6117, + "theoretical_loss": 3.956424629928146, + "tokens_seen": 484769792 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043085255767301905, + "loss": 1.5773, + "theoretical_loss": 3.9563671336691613, + "tokens_seen": 484835328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308425275827483, + "loss": 1.6727, + "theoretical_loss": 3.956309647357303, + "tokens_seen": 484900864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 182723, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4056205749511719, + "objective/train/theoretical_loss": 3.956252170989507, + "objective/train/tokens_used": 505426400, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004308324974924774, + "loss": 1.5073, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043082246740220665, + "loss": 1.577, + "theoretical_loss": 3.9561947045627095, + "tokens_seen": 485031936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043081243731193583, + "loss": 1.6515, + "theoretical_loss": 3.9561372480738486, + "tokens_seen": 485097472 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430802407221665, + "loss": 1.5822, + "theoretical_loss": 3.956079801519863, + "tokens_seen": 485163008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307923771313942, + "loss": 1.5889, + "theoretical_loss": 3.9560223648976924, + "tokens_seen": 485228544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043078234704112343, + "loss": 1.6548, + "theoretical_loss": 3.9559649382042803, + "tokens_seen": 485294080 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043077231695085256, + "loss": 1.5639, + "theoretical_loss": 3.955907521436569, + "tokens_seen": 485359616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307622868605818, + "loss": 1.6688, + "theoretical_loss": 3.9558501145915033, + "tokens_seen": 485425152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307522567703109, + "loss": 1.5774, + "theoretical_loss": 3.9557927176660286, + "tokens_seen": 485490688 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043074222668004015, + "loss": 1.612, + "theoretical_loss": 3.9557353306570926, + "tokens_seen": 485556224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043073219658976933, + "loss": 1.5561, + "theoretical_loss": 3.9556779535616444, + "tokens_seen": 485621760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307221664994985, + "loss": 1.6194, + "theoretical_loss": 3.955620586376633, + "tokens_seen": 485687296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307121364092277, + "loss": 1.5437, + "theoretical_loss": 3.9555632290990097, + "tokens_seen": 485752832 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004307021063189569, + "loss": 1.6538, + "theoretical_loss": 3.9555058817257285, + "tokens_seen": 485818368 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043069207622868606, + "loss": 1.5791, + "theoretical_loss": 3.9554485442537417, + "tokens_seen": 485883904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306820461384153, + "loss": 1.4565, + "theoretical_loss": 3.955391216680006, + "tokens_seen": 485949440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306720160481444, + "loss": 1.6182, + "theoretical_loss": 3.9553338990014772, + "tokens_seen": 486014976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043066198595787366, + "loss": 1.5855, + "theoretical_loss": 3.9552765912151138, + "tokens_seen": 486080512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306519558676028, + "loss": 1.6145, + "theoretical_loss": 3.9552192933178745, + "tokens_seen": 486146048 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430641925777332, + "loss": 1.586, + "theoretical_loss": 3.9551620053067214, + "tokens_seen": 486211584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306318956870612, + "loss": 1.4667, + "theoretical_loss": 3.9551047271786155, + "tokens_seen": 486277120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306218655967904, + "loss": 1.7272, + "theoretical_loss": 3.9550474589305202, + "tokens_seen": 486342656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043061183550651956, + "loss": 1.559, + "theoretical_loss": 3.954990200559401, + "tokens_seen": 486408192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004306018054162488, + "loss": 1.5581, + "theoretical_loss": 3.9549329520622236, + "tokens_seen": 486473728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305917753259779, + "loss": 1.4871, + "theoretical_loss": 3.954875713435955, + "tokens_seen": 486539264 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043058174523570716, + "loss": 1.5535, + "theoretical_loss": 3.9548184846775647, + "tokens_seen": 486604800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305717151454363, + "loss": 1.5732, + "theoretical_loss": 3.9547612657840228, + "tokens_seen": 486670336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305616850551655, + "loss": 1.5921, + "theoretical_loss": 3.9547040567523, + "tokens_seen": 486735872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305516549648947, + "loss": 1.6334, + "theoretical_loss": 3.9546468575793696, + "tokens_seen": 486801408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305416248746239, + "loss": 1.5996, + "theoretical_loss": 3.9545896682622064, + "tokens_seen": 486866944 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043053159478435306, + "loss": 1.5344, + "theoretical_loss": 3.954532488797785, + "tokens_seen": 486932480 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043052156469408225, + "loss": 1.5263, + "theoretical_loss": 3.954475319183082, + "tokens_seen": 486998016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004305115346038114, + "loss": 1.5674, + "theoretical_loss": 3.9544181594150762, + "tokens_seen": 487063552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043050150451354066, + "loss": 1.5956, + "theoretical_loss": 3.9543610094907464, + "tokens_seen": 487129088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304914744232698, + "loss": 1.471, + "theoretical_loss": 3.9543038694070747, + "tokens_seen": 487194624 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430481444332999, + "loss": 1.5378, + "theoretical_loss": 3.9542467391610416, + "tokens_seen": 487260160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043047141424272815, + "loss": 1.6113, + "theoretical_loss": 3.9541896187496315, + "tokens_seen": 487325696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304613841524574, + "loss": 1.5231, + "theoretical_loss": 3.9541325081698293, + "tokens_seen": 487391232 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043045135406218657, + "loss": 1.5779, + "theoretical_loss": 3.954075407418621, + "tokens_seen": 487456768 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043044132397191575, + "loss": 1.4687, + "theoretical_loss": 3.954018316492994, + "tokens_seen": 487522304 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043043129388164493, + "loss": 1.4778, + "theoretical_loss": 3.9539612353899365, + "tokens_seen": 487587840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043042126379137417, + "loss": 1.5214, + "theoretical_loss": 3.9539041641064396, + "tokens_seen": 487653376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004304112337011033, + "loss": 1.502, + "theoretical_loss": 3.953847102639494, + "tokens_seen": 487718912 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043040120361083253, + "loss": 1.6145, + "theoretical_loss": 3.9537900509860933, + "tokens_seen": 487784448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043039117352056165, + "loss": 1.6771, + "theoretical_loss": 3.953733009143231, + "tokens_seen": 487849984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303811434302909, + "loss": 1.5203, + "theoretical_loss": 3.9536759771079018, + "tokens_seen": 487915520 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043037111334002007, + "loss": 1.5582, + "theoretical_loss": 3.9536189548771037, + "tokens_seen": 487981056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043036108324974925, + "loss": 1.5911, + "theoretical_loss": 3.953561942447834, + "tokens_seen": 488046592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043035105315947843, + "loss": 1.5443, + "theoretical_loss": 3.953504939817093, + "tokens_seen": 488112128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303410230692076, + "loss": 1.5563, + "theoretical_loss": 3.953447946981881, + "tokens_seen": 488177664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 183865, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6494295597076416, + "objective/train/theoretical_loss": 3.9533909639391993, + "objective/train/tokens_used": 508703200, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303309929789368, + "loss": 1.6599, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043032096288866603, + "loss": 1.5277, + "theoretical_loss": 3.953333990686051, + "tokens_seen": 488308736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043031093279839516, + "loss": 1.5885, + "theoretical_loss": 3.9532770272194426, + "tokens_seen": 488374272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004303009027081244, + "loss": 1.5373, + "theoretical_loss": 3.9532200735363787, + "tokens_seen": 488439808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302908726178535, + "loss": 1.6473, + "theoretical_loss": 3.953163129633867, + "tokens_seen": 488505344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043028084252758276, + "loss": 1.4146, + "theoretical_loss": 3.9531061955089157, + "tokens_seen": 488570880 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043027081243731194, + "loss": 1.5276, + "theoretical_loss": 3.9530492711585357, + "tokens_seen": 488636416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302607823470411, + "loss": 1.5714, + "theoretical_loss": 3.9529923565797374, + "tokens_seen": 488701952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302507522567703, + "loss": 1.5115, + "theoretical_loss": 3.952935451769533, + "tokens_seen": 488767488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043024072216649953, + "loss": 1.5373, + "theoretical_loss": 3.9528785567249374, + "tokens_seen": 488833024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043023069207622866, + "loss": 1.604, + "theoretical_loss": 3.9528216714429654, + "tokens_seen": 488898560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004302206619859579, + "loss": 1.5902, + "theoretical_loss": 3.9527647959206336, + "tokens_seen": 488964096 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430210631895687, + "loss": 1.5865, + "theoretical_loss": 3.9527079301549595, + "tokens_seen": 489029632 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043020060180541626, + "loss": 1.6, + "theoretical_loss": 3.9526510741429624, + "tokens_seen": 489095168 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043019057171514544, + "loss": 1.5869, + "theoretical_loss": 3.9525942278816633, + "tokens_seen": 489160704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301805416248746, + "loss": 1.616, + "theoretical_loss": 3.952537391368083, + "tokens_seen": 489226240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043017051153460386, + "loss": 1.4484, + "theoretical_loss": 3.9524805645992447, + "tokens_seen": 489291776 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430160481444333, + "loss": 1.5815, + "theoretical_loss": 3.9524237475721726, + "tokens_seen": 489357312 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301504513540622, + "loss": 1.6213, + "theoretical_loss": 3.952366940283894, + "tokens_seen": 489422848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301404212637914, + "loss": 1.5728, + "theoretical_loss": 3.952310142731434, + "tokens_seen": 489488384 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301303911735206, + "loss": 1.6064, + "theoretical_loss": 3.9522533549118215, + "tokens_seen": 489553920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043012036108324976, + "loss": 1.5206, + "theoretical_loss": 3.952196576822086, + "tokens_seen": 489619456 + }, + { + "epoch": 0.15, + "learning_rate": 0.000430110330992979, + "loss": 1.6918, + "theoretical_loss": 3.9521398084592585, + "tokens_seen": 489684992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004301003009027081, + "loss": 1.6578, + "theoretical_loss": 3.952083049820372, + "tokens_seen": 489750528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043009027081243736, + "loss": 1.5156, + "theoretical_loss": 3.9520263009024585, + "tokens_seen": 489816064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300802407221665, + "loss": 1.6178, + "theoretical_loss": 3.951969561702554, + "tokens_seen": 489881600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300702106318957, + "loss": 1.6547, + "theoretical_loss": 3.9519128322176935, + "tokens_seen": 489947136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300601805416249, + "loss": 1.6999, + "theoretical_loss": 3.951856112444915, + "tokens_seen": 490012672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004300501504513541, + "loss": 1.5814, + "theoretical_loss": 3.951799402381258, + "tokens_seen": 490078208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043004012036108326, + "loss": 1.496, + "theoretical_loss": 3.951742702023761, + "tokens_seen": 490143744 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043003009027081245, + "loss": 1.4963, + "theoretical_loss": 3.951686011369466, + "tokens_seen": 490209280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043002006018054163, + "loss": 1.6299, + "theoretical_loss": 3.9516293304154155, + "tokens_seen": 490274816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043001003009027086, + "loss": 1.5705, + "theoretical_loss": 3.9515726591586544, + "tokens_seen": 490340352 + }, + { + "epoch": 0.15, + "learning_rate": 0.00043, + "loss": 1.6236, + "theoretical_loss": 3.9515159975962257, + "tokens_seen": 490405888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299899699097292, + "loss": 1.4522, + "theoretical_loss": 3.951459345725178, + "tokens_seen": 490471424 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042997993981945835, + "loss": 1.5354, + "theoretical_loss": 3.9514027035425583, + "tokens_seen": 490536960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299699097291876, + "loss": 1.6264, + "theoretical_loss": 3.951346071045415, + "tokens_seen": 490602496 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042995987963891677, + "loss": 1.585, + "theoretical_loss": 3.951289448230799, + "tokens_seen": 490668032 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042994984954864595, + "loss": 1.5451, + "theoretical_loss": 3.9512328350957624, + "tokens_seen": 490733568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042993981945837513, + "loss": 1.6072, + "theoretical_loss": 3.9511762316373575, + "tokens_seen": 490799104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042992978936810437, + "loss": 1.6614, + "theoretical_loss": 3.951119637852639, + "tokens_seen": 490864640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004299197592778335, + "loss": 1.5357, + "theoretical_loss": 3.951063053738662, + "tokens_seen": 490930176 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042990972918756273, + "loss": 1.5834, + "theoretical_loss": 3.951006479292484, + "tokens_seen": 490995712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042989969909729185, + "loss": 1.6221, + "theoretical_loss": 3.9509499145111624, + "tokens_seen": 491061248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298896690070211, + "loss": 1.5829, + "theoretical_loss": 3.950893359391756, + "tokens_seen": 491126784 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042987963891675027, + "loss": 1.4805, + "theoretical_loss": 3.9508368139313275, + "tokens_seen": 491192320 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042986960882647945, + "loss": 1.5871, + "theoretical_loss": 3.950780278126937, + "tokens_seen": 491257856 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042985957873620863, + "loss": 1.6076, + "theoretical_loss": 3.9507237519756484, + "tokens_seen": 491323392 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298495486459378, + "loss": 1.5873, + "theoretical_loss": 3.9506672354745263, + "tokens_seen": 491388928 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429839518555667, + "loss": 1.5696, + "theoretical_loss": 3.950610728620636, + "tokens_seen": 491454464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 185162, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5476281642913818, + "objective/train/theoretical_loss": 3.950554231411046, + "objective/train/tokens_used": 511980000, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042982948846539623, + "loss": 1.5836, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042981945837512536, + "loss": 1.5243, + "theoretical_loss": 3.9504977438428233, + "tokens_seen": 491585536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004298094282848546, + "loss": 1.5139, + "theoretical_loss": 3.950441265913038, + "tokens_seen": 491651072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297993981945837, + "loss": 1.4715, + "theoretical_loss": 3.950384797618761, + "tokens_seen": 491716608 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042978936810431296, + "loss": 1.6328, + "theoretical_loss": 3.950328338957064, + "tokens_seen": 491782144 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042977933801404214, + "loss": 1.5393, + "theoretical_loss": 3.950271889925022, + "tokens_seen": 491847680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297693079237713, + "loss": 1.6496, + "theoretical_loss": 3.9502154505197087, + "tokens_seen": 491913216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297592778335005, + "loss": 1.6647, + "theoretical_loss": 3.9501590207382, + "tokens_seen": 491978752 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042974924774322973, + "loss": 1.6055, + "theoretical_loss": 3.9501026005775732, + "tokens_seen": 492044288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042973921765295886, + "loss": 1.5856, + "theoretical_loss": 3.9500461900349078, + "tokens_seen": 492109824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297291875626881, + "loss": 1.6018, + "theoretical_loss": 3.949989789107283, + "tokens_seen": 492175360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004297191574724172, + "loss": 1.5565, + "theoretical_loss": 3.9499333977917805, + "tokens_seen": 492240896 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042970912738214646, + "loss": 1.6647, + "theoretical_loss": 3.949877016085482, + "tokens_seen": 492306432 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042969909729187564, + "loss": 1.5812, + "theoretical_loss": 3.949820643985472, + "tokens_seen": 492371968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296890672016048, + "loss": 1.654, + "theoretical_loss": 3.9497642814888345, + "tokens_seen": 492437504 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429679037111334, + "loss": 1.5564, + "theoretical_loss": 3.949707928592657, + "tokens_seen": 492503040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296690070210632, + "loss": 1.6549, + "theoretical_loss": 3.949651585294026, + "tokens_seen": 492568576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042965897693079236, + "loss": 1.6361, + "theoretical_loss": 3.9495952515900306, + "tokens_seen": 492634112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296489468405216, + "loss": 1.6319, + "theoretical_loss": 3.9495389274777613, + "tokens_seen": 492699648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296389167502507, + "loss": 1.457, + "theoretical_loss": 3.9494826129543084, + "tokens_seen": 492765184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042962888665997996, + "loss": 1.6359, + "theoretical_loss": 3.9494263080167658, + "tokens_seen": 492830720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296188565697091, + "loss": 1.6388, + "theoretical_loss": 3.949370012662227, + "tokens_seen": 492896256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004296088264794383, + "loss": 1.5711, + "theoretical_loss": 3.9493137268877865, + "tokens_seen": 492961792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295987963891675, + "loss": 1.5987, + "theoretical_loss": 3.9492574506905416, + "tokens_seen": 493027328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295887662988967, + "loss": 1.5196, + "theoretical_loss": 3.949201184067589, + "tokens_seen": 493092864 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042957873620862587, + "loss": 1.6229, + "theoretical_loss": 3.949144927016029, + "tokens_seen": 493158400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295687061183551, + "loss": 1.6135, + "theoretical_loss": 3.9490886795329603, + "tokens_seen": 493223936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042955867602808423, + "loss": 1.561, + "theoretical_loss": 3.9490324416154854, + "tokens_seen": 493289472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042954864593781347, + "loss": 1.6215, + "theoretical_loss": 3.9489762132607065, + "tokens_seen": 493355008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295386158475426, + "loss": 1.6051, + "theoretical_loss": 3.948919994465728, + "tokens_seen": 493420544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042952858575727183, + "loss": 1.7563, + "theoretical_loss": 3.9488637852276547, + "tokens_seen": 493486080 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429518555667001, + "loss": 1.6436, + "theoretical_loss": 3.9488075855435936, + "tokens_seen": 493551616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004295085255767302, + "loss": 1.7358, + "theoretical_loss": 3.9487513954106523, + "tokens_seen": 493617152 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042949849548645937, + "loss": 1.5956, + "theoretical_loss": 3.94869521482594, + "tokens_seen": 493682688 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042948846539618855, + "loss": 1.5184, + "theoretical_loss": 3.9486390437865664, + "tokens_seen": 493748224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042947843530591773, + "loss": 1.5729, + "theoretical_loss": 3.948582882289644, + "tokens_seen": 493813760 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042946840521564697, + "loss": 1.5823, + "theoretical_loss": 3.9485267303322846, + "tokens_seen": 493879296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294583751253761, + "loss": 1.4933, + "theoretical_loss": 3.9484705879116033, + "tokens_seen": 493944832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042944834503510533, + "loss": 1.4816, + "theoretical_loss": 3.9484144550247144, + "tokens_seen": 494010368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294383149448345, + "loss": 1.6871, + "theoretical_loss": 3.948358331668735, + "tokens_seen": 494075904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004294282848545637, + "loss": 1.5649, + "theoretical_loss": 3.9483022178407827, + "tokens_seen": 494141440 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042941825476429293, + "loss": 1.7247, + "theoretical_loss": 3.9482461135379774, + "tokens_seen": 494206976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042940822467402206, + "loss": 1.6067, + "theoretical_loss": 3.9481900187574386, + "tokens_seen": 494272512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293981945837513, + "loss": 1.5952, + "theoretical_loss": 3.948133933496288, + "tokens_seen": 494338048 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042938816449348047, + "loss": 1.5544, + "theoretical_loss": 3.948077857751649, + "tokens_seen": 494403584 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042937813440320965, + "loss": 1.6513, + "theoretical_loss": 3.948021791520645, + "tokens_seen": 494469120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042936810431293883, + "loss": 1.6457, + "theoretical_loss": 3.9479657348004014, + "tokens_seen": 494534656 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429358074222668, + "loss": 1.6661, + "theoretical_loss": 3.9479096875880453, + "tokens_seen": 494600192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293480441323972, + "loss": 1.7534, + "theoretical_loss": 3.947853649880704, + "tokens_seen": 494665728 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042933801404212643, + "loss": 1.5908, + "theoretical_loss": 3.947797621675507, + "tokens_seen": 494731264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 186433, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.629323959350586, + "objective/train/theoretical_loss": 3.947741602969585, + "objective/train/tokens_used": 515256800, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042932798395185556, + "loss": 1.5342, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293179538615848, + "loss": 1.4802, + "theoretical_loss": 3.9476855937600686, + "tokens_seen": 494862336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004293079237713139, + "loss": 1.4599, + "theoretical_loss": 3.9476295940440913, + "tokens_seen": 494927872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042929789368104316, + "loss": 1.6211, + "theoretical_loss": 3.9475736038187867, + "tokens_seen": 494993408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042928786359077234, + "loss": 1.5377, + "theoretical_loss": 3.947517623081291, + "tokens_seen": 495058944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292778335005015, + "loss": 1.5854, + "theoretical_loss": 3.9474616518287395, + "tokens_seen": 495124480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292678034102307, + "loss": 1.6095, + "theoretical_loss": 3.9474056900582717, + "tokens_seen": 495190016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042925777331995993, + "loss": 1.7139, + "theoretical_loss": 3.9473497377670244, + "tokens_seen": 495255552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042924774322968906, + "loss": 1.7294, + "theoretical_loss": 3.94729379495214, + "tokens_seen": 495321088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292377131394183, + "loss": 1.548, + "theoretical_loss": 3.947237861610759, + "tokens_seen": 495386624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004292276830491474, + "loss": 1.5953, + "theoretical_loss": 3.9471819377400243, + "tokens_seen": 495452160 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042921765295887666, + "loss": 1.5013, + "theoretical_loss": 3.94712602333708, + "tokens_seen": 495517696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042920762286860584, + "loss": 1.4782, + "theoretical_loss": 3.9470701183990715, + "tokens_seen": 495583232 + }, + { + "epoch": 0.15, + "learning_rate": 0.000429197592778335, + "loss": 1.6671, + "theoretical_loss": 3.947014222923145, + "tokens_seen": 495648768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291875626880642, + "loss": 1.5573, + "theoretical_loss": 3.946958336906449, + "tokens_seen": 495714304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291775325977934, + "loss": 1.4754, + "theoretical_loss": 3.9469024603461307, + "tokens_seen": 495779840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042916750250752256, + "loss": 1.7248, + "theoretical_loss": 3.946846593239342, + "tokens_seen": 495845376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291574724172518, + "loss": 1.6535, + "theoretical_loss": 3.946790735583234, + "tokens_seen": 495910912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291474423269809, + "loss": 1.5362, + "theoretical_loss": 3.9467348873749586, + "tokens_seen": 495976448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042913741223671016, + "loss": 1.6367, + "theoretical_loss": 3.9466790486116707, + "tokens_seen": 496041984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291273821464393, + "loss": 1.6201, + "theoretical_loss": 3.946623219290525, + "tokens_seen": 496107520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291173520561685, + "loss": 1.5697, + "theoretical_loss": 3.9465673994086776, + "tokens_seen": 496173056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004291073219658977, + "loss": 1.6239, + "theoretical_loss": 3.9465115889632862, + "tokens_seen": 496238592 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290972918756269, + "loss": 1.6298, + "theoretical_loss": 3.9464557879515105, + "tokens_seen": 496304128 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042908726178535607, + "loss": 1.5026, + "theoretical_loss": 3.946399996370509, + "tokens_seen": 496369664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290772316950853, + "loss": 1.5842, + "theoretical_loss": 3.9463442142174445, + "tokens_seen": 496435200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042906720160481443, + "loss": 1.6479, + "theoretical_loss": 3.9462884414894788, + "tokens_seen": 496500736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042905717151454367, + "loss": 1.5409, + "theoretical_loss": 3.946232678183775, + "tokens_seen": 496566272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290471414242728, + "loss": 1.5351, + "theoretical_loss": 3.9461769242974998, + "tokens_seen": 496631808 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042903711133400203, + "loss": 1.6405, + "theoretical_loss": 3.946121179827818, + "tokens_seen": 496697344 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290270812437312, + "loss": 1.6361, + "theoretical_loss": 3.9460654447718975, + "tokens_seen": 496762880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004290170511534604, + "loss": 1.6107, + "theoretical_loss": 3.9460097191269066, + "tokens_seen": 496828416 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042900702106318957, + "loss": 1.5966, + "theoretical_loss": 3.9459540028900157, + "tokens_seen": 496893952 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042899699097291875, + "loss": 1.5883, + "theoretical_loss": 3.945898296058396, + "tokens_seen": 496959488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042898696088264793, + "loss": 1.6245, + "theoretical_loss": 3.9458425986292194, + "tokens_seen": 497025024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042897693079237717, + "loss": 1.5626, + "theoretical_loss": 3.9457869105996597, + "tokens_seen": 497090560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289669007021063, + "loss": 1.5399, + "theoretical_loss": 3.9457312319668914, + "tokens_seen": 497156096 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042895687061183553, + "loss": 1.6431, + "theoretical_loss": 3.9456755627280904, + "tokens_seen": 497221632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289468405215647, + "loss": 1.4595, + "theoretical_loss": 3.945619902880434, + "tokens_seen": 497287168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289368104312939, + "loss": 1.6723, + "theoretical_loss": 3.945564252421102, + "tokens_seen": 497352704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004289267803410231, + "loss": 1.687, + "theoretical_loss": 3.9455086113472717, + "tokens_seen": 497418240 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042891675025075226, + "loss": 1.559, + "theoretical_loss": 3.9454529796561255, + "tokens_seen": 497483776 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042890672016048144, + "loss": 1.6145, + "theoretical_loss": 3.9453973573448455, + "tokens_seen": 497549312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042889669007021067, + "loss": 1.5505, + "theoretical_loss": 3.945341744410615, + "tokens_seen": 497614848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288866599799398, + "loss": 1.5695, + "theoretical_loss": 3.9452861408506172, + "tokens_seen": 497680384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042887662988966903, + "loss": 1.5995, + "theoretical_loss": 3.945230546662039, + "tokens_seen": 497745920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042886659979939816, + "loss": 1.544, + "theoretical_loss": 3.945174961842068, + "tokens_seen": 497811456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288565697091274, + "loss": 1.6131, + "theoretical_loss": 3.9451193863878906, + "tokens_seen": 497876992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288465396188566, + "loss": 1.5737, + "theoretical_loss": 3.945063820296698, + "tokens_seen": 497942528 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042883650952858576, + "loss": 1.6282, + "theoretical_loss": 3.9450082635656796, + "tokens_seen": 498008064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 187487, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.694678783416748, + "objective/train/theoretical_loss": 3.9449527161920273, + "objective/train/tokens_used": 518533600, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042882647943831494, + "loss": 1.6521, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288164493480441, + "loss": 1.5451, + "theoretical_loss": 3.944897178172935, + "tokens_seen": 498139136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004288064192577733, + "loss": 1.6822, + "theoretical_loss": 3.9448416495055962, + "tokens_seen": 498204672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042879638916750254, + "loss": 1.6093, + "theoretical_loss": 3.9447861301872065, + "tokens_seen": 498270208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042878635907723166, + "loss": 1.4672, + "theoretical_loss": 3.944730620214962, + "tokens_seen": 498335744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287763289869609, + "loss": 1.4993, + "theoretical_loss": 3.944675119586061, + "tokens_seen": 498401280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287662988966901, + "loss": 1.5953, + "theoretical_loss": 3.944619628297704, + "tokens_seen": 498466816 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042875626880641926, + "loss": 1.5777, + "theoretical_loss": 3.944564146347089, + "tokens_seen": 498532352 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042874623871614844, + "loss": 1.4656, + "theoretical_loss": 3.9445086737314186, + "tokens_seen": 498597888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287362086258776, + "loss": 1.6247, + "theoretical_loss": 3.9444532104478958, + "tokens_seen": 498663424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004287261785356068, + "loss": 1.611, + "theoretical_loss": 3.9443977564937236, + "tokens_seen": 498728960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042871614844533604, + "loss": 1.6169, + "theoretical_loss": 3.9443423118661074, + "tokens_seen": 498794496 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042870611835506517, + "loss": 1.5706, + "theoretical_loss": 3.944286876562254, + "tokens_seen": 498860032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286960882647944, + "loss": 1.4902, + "theoretical_loss": 3.9442314505793705, + "tokens_seen": 498925568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042868605817452353, + "loss": 1.6781, + "theoretical_loss": 3.944176033914666, + "tokens_seen": 498991104 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042867602808425276, + "loss": 1.632, + "theoretical_loss": 3.9441206265653497, + "tokens_seen": 499056640 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428665997993982, + "loss": 1.5624, + "theoretical_loss": 3.9440652285286335, + "tokens_seen": 499122176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286559679037111, + "loss": 1.6079, + "theoretical_loss": 3.9440098398017294, + "tokens_seen": 499187712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042864593781344036, + "loss": 1.5742, + "theoretical_loss": 3.9439544603818506, + "tokens_seen": 499253248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286359077231695, + "loss": 1.6694, + "theoretical_loss": 3.9438990902662128, + "tokens_seen": 499318784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286258776328987, + "loss": 1.595, + "theoretical_loss": 3.9438437294520305, + "tokens_seen": 499384320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286158475426279, + "loss": 1.7192, + "theoretical_loss": 3.943788377936522, + "tokens_seen": 499449856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004286058174523571, + "loss": 1.5943, + "theoretical_loss": 3.943733035716905, + "tokens_seen": 499515392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042859578736208627, + "loss": 1.5428, + "theoretical_loss": 3.9436777027903993, + "tokens_seen": 499580928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285857572718155, + "loss": 1.7177, + "theoretical_loss": 3.9436223791542258, + "tokens_seen": 499646464 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042857572718154463, + "loss": 1.52, + "theoretical_loss": 3.943567064805606, + "tokens_seen": 499712000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042856569709127387, + "loss": 1.477, + "theoretical_loss": 3.943511759741763, + "tokens_seen": 499777536 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428555667001003, + "loss": 1.4918, + "theoretical_loss": 3.9434564639599214, + "tokens_seen": 499843072 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042854563691073223, + "loss": 1.5972, + "theoretical_loss": 3.9434011774573063, + "tokens_seen": 499908608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285356068204614, + "loss": 1.5364, + "theoretical_loss": 3.9433459002311446, + "tokens_seen": 499974144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004285255767301906, + "loss": 1.53, + "theoretical_loss": 3.9432906322786643, + "tokens_seen": 500039680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042851554663991977, + "loss": 1.5922, + "theoretical_loss": 3.943235373597094, + "tokens_seen": 500105216 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042850551654964895, + "loss": 1.6553, + "theoretical_loss": 3.943180124183665, + "tokens_seen": 500170752 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042849548645937813, + "loss": 1.5962, + "theoretical_loss": 3.9431248840356075, + "tokens_seen": 500236288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042848545636910737, + "loss": 1.6162, + "theoretical_loss": 3.943069653150155, + "tokens_seen": 500301824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284754262788365, + "loss": 1.5875, + "theoretical_loss": 3.9430144315245412, + "tokens_seen": 500367360 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042846539618856573, + "loss": 1.6092, + "theoretical_loss": 3.942959219156, + "tokens_seen": 500432896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284553660982949, + "loss": 1.7119, + "theoretical_loss": 3.942904016041769, + "tokens_seen": 500498432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284453360080241, + "loss": 1.6227, + "theoretical_loss": 3.9428488221790854, + "tokens_seen": 500563968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004284353059177533, + "loss": 1.656, + "theoretical_loss": 3.9427936375651877, + "tokens_seen": 500629504 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042842527582748246, + "loss": 1.5814, + "theoretical_loss": 3.9427384621973154, + "tokens_seen": 500695040 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042841524573721164, + "loss": 1.5323, + "theoretical_loss": 3.9426832960727087, + "tokens_seen": 500760576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042840521564694087, + "loss": 1.482, + "theoretical_loss": 3.9426281391886113, + "tokens_seen": 500826112 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042839518555667, + "loss": 1.5692, + "theoretical_loss": 3.9425729915422654, + "tokens_seen": 500891648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042838515546639923, + "loss": 1.5238, + "theoretical_loss": 3.942517853130916, + "tokens_seen": 500957184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042837512537612836, + "loss": 1.5412, + "theoretical_loss": 3.9424627239518086, + "tokens_seen": 501022720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283650952858576, + "loss": 1.6765, + "theoretical_loss": 3.9424076040021903, + "tokens_seen": 501088256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283550651955868, + "loss": 1.5744, + "theoretical_loss": 3.9423524932793086, + "tokens_seen": 501153792 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042834503510531596, + "loss": 1.6131, + "theoretical_loss": 3.9422973917804134, + "tokens_seen": 501219328 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042833500501504514, + "loss": 1.5726, + "theoretical_loss": 3.9422422995027544, + "tokens_seen": 501284864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 188727, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.42486572265625, + "objective/train/theoretical_loss": 3.9421872164435836, + "objective/train/tokens_used": 521810400, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283249749247743, + "loss": 1.5566, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004283149448345035, + "loss": 1.5118, + "theoretical_loss": 3.942132142600154, + "tokens_seen": 501415936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042830491474423274, + "loss": 1.5508, + "theoretical_loss": 3.942077077969719, + "tokens_seen": 501481472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042829488465396186, + "loss": 1.672, + "theoretical_loss": 3.9420220225495335, + "tokens_seen": 501547008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282848545636911, + "loss": 1.5009, + "theoretical_loss": 3.9419669763368548, + "tokens_seen": 501612544 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282748244734203, + "loss": 1.5951, + "theoretical_loss": 3.9419119393289397, + "tokens_seen": 501678080 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042826479438314946, + "loss": 1.5054, + "theoretical_loss": 3.941856911523047, + "tokens_seen": 501743616 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042825476429287864, + "loss": 1.5326, + "theoretical_loss": 3.9418018929164362, + "tokens_seen": 501809152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282447342026078, + "loss": 1.5976, + "theoretical_loss": 3.9417468835063687, + "tokens_seen": 501874688 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428234704112337, + "loss": 1.6105, + "theoretical_loss": 3.9416918832901064, + "tokens_seen": 501940224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042822467402206624, + "loss": 1.583, + "theoretical_loss": 3.9416368922649134, + "tokens_seen": 502005760 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042821464393179537, + "loss": 1.5377, + "theoretical_loss": 3.9415819104280527, + "tokens_seen": 502071296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004282046138415246, + "loss": 1.5823, + "theoretical_loss": 3.941526937776791, + "tokens_seen": 502136832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042819458375125373, + "loss": 1.4496, + "theoretical_loss": 3.941471974308395, + "tokens_seen": 502202368 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042818455366098296, + "loss": 1.6807, + "theoretical_loss": 3.9414170200201326, + "tokens_seen": 502267904 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042817452357071215, + "loss": 1.5031, + "theoretical_loss": 3.9413620749092733, + "tokens_seen": 502333440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281644934804413, + "loss": 1.4701, + "theoretical_loss": 3.9413071389730874, + "tokens_seen": 502398976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281544633901705, + "loss": 1.4724, + "theoretical_loss": 3.941252212208846, + "tokens_seen": 502464512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281444332998997, + "loss": 1.5893, + "theoretical_loss": 3.941197294613822, + "tokens_seen": 502530048 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042813440320962887, + "loss": 1.6058, + "theoretical_loss": 3.9411423861852892, + "tokens_seen": 502595584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004281243731193581, + "loss": 1.5946, + "theoretical_loss": 3.941087486920523, + "tokens_seen": 502661120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042811434302908723, + "loss": 1.6005, + "theoretical_loss": 3.9410325968167994, + "tokens_seen": 502726656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042810431293881647, + "loss": 1.5257, + "theoretical_loss": 3.9409777158713952, + "tokens_seen": 502792192 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042809428284854565, + "loss": 1.5996, + "theoretical_loss": 3.9409228440815895, + "tokens_seen": 502857728 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042808425275827483, + "loss": 1.6405, + "theoretical_loss": 3.9408679814446623, + "tokens_seen": 502923264 + }, + { + "epoch": 0.15, + "learning_rate": 0.000428074222668004, + "loss": 1.561, + "theoretical_loss": 3.9408131279578935, + "tokens_seen": 502988800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280641925777332, + "loss": 1.5236, + "theoretical_loss": 3.9407582836185657, + "tokens_seen": 503054336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280541624874624, + "loss": 1.6492, + "theoretical_loss": 3.940703448423962, + "tokens_seen": 503119872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280441323971916, + "loss": 1.6644, + "theoretical_loss": 3.9406486223713664, + "tokens_seen": 503185408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042803410230692074, + "loss": 1.6296, + "theoretical_loss": 3.940593805458065, + "tokens_seen": 503250944 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042802407221664997, + "loss": 1.5414, + "theoretical_loss": 3.940538997681344, + "tokens_seen": 503316480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004280140421263791, + "loss": 1.5787, + "theoretical_loss": 3.9404841990384916, + "tokens_seen": 503382016 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042800401203610833, + "loss": 1.5814, + "theoretical_loss": 3.940429409526796, + "tokens_seen": 503447552 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279939819458375, + "loss": 1.6119, + "theoretical_loss": 3.940374629143548, + "tokens_seen": 503513088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279839518555667, + "loss": 1.631, + "theoretical_loss": 3.940319857886039, + "tokens_seen": 503578624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279739217652959, + "loss": 1.4683, + "theoretical_loss": 3.9402650957515606, + "tokens_seen": 503644160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279638916750251, + "loss": 1.588, + "theoretical_loss": 3.940210342737407, + "tokens_seen": 503709696 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042795386158475424, + "loss": 1.544, + "theoretical_loss": 3.940155598840873, + "tokens_seen": 503775232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279438314944835, + "loss": 1.4988, + "theoretical_loss": 3.9401008640592545, + "tokens_seen": 503840768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279338014042126, + "loss": 1.574, + "theoretical_loss": 3.940046138389848, + "tokens_seen": 503906304 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042792377131394184, + "loss": 1.6756, + "theoretical_loss": 3.9399914218299523, + "tokens_seen": 503971840 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042791374122367107, + "loss": 1.7332, + "theoretical_loss": 3.9399367143768664, + "tokens_seen": 504037376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004279037111334002, + "loss": 1.5862, + "theoretical_loss": 3.939882016027891, + "tokens_seen": 504102912 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042789368104312943, + "loss": 1.5942, + "theoretical_loss": 3.939827326780328, + "tokens_seen": 504168448 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042788365095285856, + "loss": 1.6421, + "theoretical_loss": 3.9397726466314795, + "tokens_seen": 504233984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278736208625878, + "loss": 1.5112, + "theoretical_loss": 3.93971797557865, + "tokens_seen": 504299520 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427863590772317, + "loss": 1.6456, + "theoretical_loss": 3.939663313619145, + "tokens_seen": 504365056 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042785356068204616, + "loss": 1.53, + "theoretical_loss": 3.9396086607502694, + "tokens_seen": 504430592 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042784353059177534, + "loss": 1.5656, + "theoretical_loss": 3.939554016969332, + "tokens_seen": 504496128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278335005015045, + "loss": 1.5701, + "theoretical_loss": 3.939499382273641, + "tokens_seen": 504561664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 189911, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4532643556594849, + "objective/train/theoretical_loss": 3.939444756660505, + "objective/train/tokens_used": 525087200, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004278234704112337, + "loss": 1.5581, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042781344032096294, + "loss": 1.5492, + "theoretical_loss": 3.9393901401272364, + "tokens_seen": 504692736 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042780341023069206, + "loss": 1.5654, + "theoretical_loss": 3.939335532671146, + "tokens_seen": 504758272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277933801404213, + "loss": 1.5552, + "theoretical_loss": 3.939280934289548, + "tokens_seen": 504823808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277833500501505, + "loss": 1.5685, + "theoretical_loss": 3.9392263449797564, + "tokens_seen": 504889344 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042777331995987966, + "loss": 1.6232, + "theoretical_loss": 3.939171764739086, + "tokens_seen": 504954880 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042776328986960884, + "loss": 1.5494, + "theoretical_loss": 3.9391171935648535, + "tokens_seen": 505020416 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427753259779338, + "loss": 1.6371, + "theoretical_loss": 3.9390626314543766, + "tokens_seen": 505085952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277432296890672, + "loss": 1.7475, + "theoretical_loss": 3.939008078404975, + "tokens_seen": 505151488 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042773319959879644, + "loss": 1.6285, + "theoretical_loss": 3.938953534413968, + "tokens_seen": 505217024 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042772316950852557, + "loss": 1.5094, + "theoretical_loss": 3.9388989994786767, + "tokens_seen": 505282560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004277131394182548, + "loss": 1.5886, + "theoretical_loss": 3.938844473596423, + "tokens_seen": 505348096 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042770310932798393, + "loss": 1.5763, + "theoretical_loss": 3.9387899567645315, + "tokens_seen": 505413632 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042769307923771316, + "loss": 1.5602, + "theoretical_loss": 3.9387354489803252, + "tokens_seen": 505479168 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042768304914744235, + "loss": 1.5957, + "theoretical_loss": 3.938680950241131, + "tokens_seen": 505544704 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042767301905717153, + "loss": 1.6137, + "theoretical_loss": 3.938626460544275, + "tokens_seen": 505610240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276629889669007, + "loss": 1.6373, + "theoretical_loss": 3.9385719798870857, + "tokens_seen": 505675776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276529588766299, + "loss": 1.6853, + "theoretical_loss": 3.9385175082668926, + "tokens_seen": 505741312 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042764292878635907, + "loss": 1.5857, + "theoretical_loss": 3.9384630456810243, + "tokens_seen": 505806848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004276328986960883, + "loss": 1.5867, + "theoretical_loss": 3.9384085921268133, + "tokens_seen": 505872384 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042762286860581743, + "loss": 1.5557, + "theoretical_loss": 3.9383541476015926, + "tokens_seen": 505937920 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042761283851554667, + "loss": 1.5969, + "theoretical_loss": 3.9382997121026944, + "tokens_seen": 506003456 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042760280842527585, + "loss": 1.6576, + "theoretical_loss": 3.938245285627455, + "tokens_seen": 506068992 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042759277833500503, + "loss": 1.5168, + "theoretical_loss": 3.9381908681732085, + "tokens_seen": 506134528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275827482447342, + "loss": 1.652, + "theoretical_loss": 3.9381364597372936, + "tokens_seen": 506200064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275727181544634, + "loss": 1.7296, + "theoretical_loss": 3.938082060317048, + "tokens_seen": 506265600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275626880641926, + "loss": 1.6652, + "theoretical_loss": 3.9380276699098102, + "tokens_seen": 506331136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275526579739218, + "loss": 1.485, + "theoretical_loss": 3.9379732885129215, + "tokens_seen": 506396672 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042754262788365094, + "loss": 1.5539, + "theoretical_loss": 3.9379189161237234, + "tokens_seen": 506462208 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042753259779338017, + "loss": 1.6838, + "theoretical_loss": 3.9378645527395575, + "tokens_seen": 506527744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275225677031093, + "loss": 1.5001, + "theoretical_loss": 3.937810198357769, + "tokens_seen": 506593280 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042751253761283853, + "loss": 1.5751, + "theoretical_loss": 3.9377558529757026, + "tokens_seen": 506658816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004275025075225677, + "loss": 1.6591, + "theoretical_loss": 3.937701516590703, + "tokens_seen": 506724352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274924774322969, + "loss": 1.5251, + "theoretical_loss": 3.9376471892001197, + "tokens_seen": 506789888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274824473420261, + "loss": 1.6937, + "theoretical_loss": 3.9375928708012986, + "tokens_seen": 506855424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274724172517553, + "loss": 1.6079, + "theoretical_loss": 3.9375385613915905, + "tokens_seen": 506920960 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042746238716148444, + "loss": 1.5753, + "theoretical_loss": 3.937484260968346, + "tokens_seen": 506986496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274523570712137, + "loss": 1.6723, + "theoretical_loss": 3.9374299695289157, + "tokens_seen": 507052032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274423269809428, + "loss": 1.6001, + "theoretical_loss": 3.9373756870706536, + "tokens_seen": 507117568 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042743229689067204, + "loss": 1.6276, + "theoretical_loss": 3.937321413590913, + "tokens_seen": 507183104 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274222668004012, + "loss": 1.7285, + "theoretical_loss": 3.9372671490870492, + "tokens_seen": 507248640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274122367101304, + "loss": 1.4768, + "theoretical_loss": 3.937212893556418, + "tokens_seen": 507314176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004274022066198596, + "loss": 1.5603, + "theoretical_loss": 3.9371586469963775, + "tokens_seen": 507379712 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042739217652958876, + "loss": 1.6031, + "theoretical_loss": 3.937104409404285, + "tokens_seen": 507445248 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042738214643931794, + "loss": 1.5136, + "theoretical_loss": 3.9370501807775007, + "tokens_seen": 507510784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273721163490472, + "loss": 1.5397, + "theoretical_loss": 3.936995961113385, + "tokens_seen": 507576320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273620862587763, + "loss": 1.5714, + "theoretical_loss": 3.9369417504093, + "tokens_seen": 507641856 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042735205616850554, + "loss": 1.5862, + "theoretical_loss": 3.936887548662608, + "tokens_seen": 507707392 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042734202607823467, + "loss": 1.5017, + "theoretical_loss": 3.9368333558706734, + "tokens_seen": 507772928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273319959879639, + "loss": 1.6642, + "theoretical_loss": 3.936779172030861, + "tokens_seen": 507838464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 191028, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.51791250705719, + "objective/train/theoretical_loss": 3.936724997140538, + "objective/train/tokens_used": 528364000, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004273219658976931, + "loss": 1.5084, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042731193580742226, + "loss": 1.6021, + "theoretical_loss": 3.9366708311970706, + "tokens_seen": 507969536 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042730190571715145, + "loss": 1.4644, + "theoretical_loss": 3.9366166741978272, + "tokens_seen": 508035072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272918756268807, + "loss": 1.5362, + "theoretical_loss": 3.936562526140178, + "tokens_seen": 508100608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272818455366098, + "loss": 1.4824, + "theoretical_loss": 3.936508387021494, + "tokens_seen": 508166144 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042727181544633904, + "loss": 1.5854, + "theoretical_loss": 3.9364542568391454, + "tokens_seen": 508231680 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042726178535606817, + "loss": 1.5187, + "theoretical_loss": 3.9364001355905067, + "tokens_seen": 508297216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272517552657974, + "loss": 1.5758, + "theoretical_loss": 3.936346023272952, + "tokens_seen": 508362752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272417251755266, + "loss": 1.5454, + "theoretical_loss": 3.936291919883855, + "tokens_seen": 508428288 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042723169508525577, + "loss": 1.5783, + "theoretical_loss": 3.9362378254205925, + "tokens_seen": 508493824 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042722166499498495, + "loss": 1.6914, + "theoretical_loss": 3.936183739880543, + "tokens_seen": 508559360 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042721163490471413, + "loss": 1.6657, + "theoretical_loss": 3.9361296632610827, + "tokens_seen": 508624896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004272016048144433, + "loss": 1.6042, + "theoretical_loss": 3.9360755955595934, + "tokens_seen": 508690432 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042719157472417255, + "loss": 1.7123, + "theoretical_loss": 3.936021536773455, + "tokens_seen": 508755968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271815446339017, + "loss": 1.5741, + "theoretical_loss": 3.9359674869000485, + "tokens_seen": 508821504 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271715145436309, + "loss": 1.5791, + "theoretical_loss": 3.9359134459367575, + "tokens_seen": 508887040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271614844533601, + "loss": 1.6005, + "theoretical_loss": 3.9358594138809657, + "tokens_seen": 508952576 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042715145436308927, + "loss": 1.6175, + "theoretical_loss": 3.9358053907300588, + "tokens_seen": 509018112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004271414242728185, + "loss": 1.5756, + "theoretical_loss": 3.935751376481422, + "tokens_seen": 509083648 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042713139418254763, + "loss": 1.5003, + "theoretical_loss": 3.9356973711324432, + "tokens_seen": 509149184 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042712136409227687, + "loss": 1.6029, + "theoretical_loss": 3.9356433746805113, + "tokens_seen": 509214720 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042711133400200605, + "loss": 1.6153, + "theoretical_loss": 3.9355893871230143, + "tokens_seen": 509280256 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042710130391173523, + "loss": 1.5979, + "theoretical_loss": 3.935535408457344, + "tokens_seen": 509345792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270912738214644, + "loss": 1.5065, + "theoretical_loss": 3.9354814386808923, + "tokens_seen": 509411328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270812437311936, + "loss": 1.5641, + "theoretical_loss": 3.935427477791051, + "tokens_seen": 509476864 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270712136409228, + "loss": 1.5792, + "theoretical_loss": 3.9353735257852147, + "tokens_seen": 509542400 + }, + { + "epoch": 0.15, + "learning_rate": 0.000427061183550652, + "loss": 1.5241, + "theoretical_loss": 3.935319582660778, + "tokens_seen": 509607936 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042705115346038114, + "loss": 1.6277, + "theoretical_loss": 3.935265648415138, + "tokens_seen": 509673472 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042704112337011037, + "loss": 1.4361, + "theoretical_loss": 3.9352117230456902, + "tokens_seen": 509739008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270310932798395, + "loss": 1.6054, + "theoretical_loss": 3.9351578065498343, + "tokens_seen": 509804544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042702106318956873, + "loss": 1.6786, + "theoretical_loss": 3.935103898924969, + "tokens_seen": 509870080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270110330992979, + "loss": 1.6773, + "theoretical_loss": 3.935050000168495, + "tokens_seen": 509935616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004270010030090271, + "loss": 1.4469, + "theoretical_loss": 3.9349961102778135, + "tokens_seen": 510001152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269909729187563, + "loss": 1.6223, + "theoretical_loss": 3.934942229250328, + "tokens_seen": 510066688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269809428284855, + "loss": 1.5088, + "theoretical_loss": 3.934888357083442, + "tokens_seen": 510132224 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042697091273821464, + "loss": 1.5273, + "theoretical_loss": 3.9348344937745594, + "tokens_seen": 510197760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269608826479439, + "loss": 1.48, + "theoretical_loss": 3.9347806393210876, + "tokens_seen": 510263296 + }, + { + "epoch": 0.15, + "learning_rate": 0.000426950852557673, + "loss": 1.6809, + "theoretical_loss": 3.934726793720433, + "tokens_seen": 510328832 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042694082246740224, + "loss": 1.5232, + "theoretical_loss": 3.9346729569700036, + "tokens_seen": 510394368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269307923771314, + "loss": 1.5121, + "theoretical_loss": 3.934619129067209, + "tokens_seen": 510459904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269207622868606, + "loss": 1.5403, + "theoretical_loss": 3.934565310009459, + "tokens_seen": 510525440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004269107321965898, + "loss": 1.6441, + "theoretical_loss": 3.9345114997941657, + "tokens_seen": 510590976 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042690070210631896, + "loss": 1.5797, + "theoretical_loss": 3.934457698418741, + "tokens_seen": 510656512 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042689067201604814, + "loss": 1.5842, + "theoretical_loss": 3.9344039058805986, + "tokens_seen": 510722048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268806419257774, + "loss": 1.5519, + "theoretical_loss": 3.934350122177154, + "tokens_seen": 510787584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268706118355065, + "loss": 1.4652, + "theoretical_loss": 3.9342963473058212, + "tokens_seen": 510853120 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042686058174523574, + "loss": 1.5736, + "theoretical_loss": 3.9342425812640194, + "tokens_seen": 510918656 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042685055165496487, + "loss": 1.4159, + "theoretical_loss": 3.9341888240491647, + "tokens_seen": 510984192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268405215646941, + "loss": 1.6681, + "theoretical_loss": 3.9341350756586766, + "tokens_seen": 511049728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268304914744233, + "loss": 1.5796, + "theoretical_loss": 3.934081336089976, + "tokens_seen": 511115264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 192198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6225566864013672, + "objective/train/theoretical_loss": 3.934027605340483, + "objective/train/tokens_used": 531640800, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042682046138415246, + "loss": 1.6606, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042681043129388165, + "loss": 1.4728, + "theoretical_loss": 3.93397388340762, + "tokens_seen": 511246336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0004268004012036109, + "loss": 1.5927, + "theoretical_loss": 3.933920170288811, + "tokens_seen": 511311872 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042679037111334, + "loss": 1.5244, + "theoretical_loss": 3.93386646598148, + "tokens_seen": 511377408 + }, + { + "epoch": 0.15, + "learning_rate": 0.00042678034102306924, + "loss": 1.588, + "theoretical_loss": 3.9338127704830534, + "tokens_seen": 511442944 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042677031093279837, + "loss": 1.5715, + "theoretical_loss": 3.933759083790957, + "tokens_seen": 511508480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267602808425276, + "loss": 1.5242, + "theoretical_loss": 3.933705405902619, + "tokens_seen": 511574016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267502507522568, + "loss": 1.5787, + "theoretical_loss": 3.9336517368154666, + "tokens_seen": 511639552 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042674022066198597, + "loss": 1.6593, + "theoretical_loss": 3.9335980765269323, + "tokens_seen": 511705088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042673019057171515, + "loss": 1.3871, + "theoretical_loss": 3.9335444250344445, + "tokens_seen": 511770624 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042672016048144433, + "loss": 1.5369, + "theoretical_loss": 3.933490782335437, + "tokens_seen": 511836160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004267101303911735, + "loss": 1.6443, + "theoretical_loss": 3.933437148427342, + "tokens_seen": 511901696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042670010030090275, + "loss": 1.5665, + "theoretical_loss": 3.933383523307594, + "tokens_seen": 511967232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266900702106319, + "loss": 1.6689, + "theoretical_loss": 3.9333299069736283, + "tokens_seen": 512032768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266800401203611, + "loss": 1.4352, + "theoretical_loss": 3.9332762994228814, + "tokens_seen": 512098304 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042667001003009024, + "loss": 1.5354, + "theoretical_loss": 3.93322270065279, + "tokens_seen": 512163840 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042665997993981947, + "loss": 1.5227, + "theoretical_loss": 3.933169110660793, + "tokens_seen": 512229376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042664994984954865, + "loss": 1.5629, + "theoretical_loss": 3.93311552944433, + "tokens_seen": 512294912 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042663991975927783, + "loss": 1.7515, + "theoretical_loss": 3.9330619570008416, + "tokens_seen": 512360448 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426629889669007, + "loss": 1.4894, + "theoretical_loss": 3.9330083933277695, + "tokens_seen": 512425984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042661985957873625, + "loss": 1.5246, + "theoretical_loss": 3.9329548384225568, + "tokens_seen": 512491520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004266098294884654, + "loss": 1.5567, + "theoretical_loss": 3.932901292282647, + "tokens_seen": 512557056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265997993981946, + "loss": 1.6734, + "theoretical_loss": 3.932847754905485, + "tokens_seen": 512622592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042658976930792374, + "loss": 1.6442, + "theoretical_loss": 3.9327942262885167, + "tokens_seen": 512688128 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426579739217653, + "loss": 1.6182, + "theoretical_loss": 3.9327407064291897, + "tokens_seen": 512753664 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042656970912738216, + "loss": 1.4822, + "theoretical_loss": 3.9326871953249514, + "tokens_seen": 512819200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042655967903711134, + "loss": 1.489, + "theoretical_loss": 3.9326336929732513, + "tokens_seen": 512884736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265496489468405, + "loss": 1.5549, + "theoretical_loss": 3.9325801993715404, + "tokens_seen": 512950272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265396188565697, + "loss": 1.5589, + "theoretical_loss": 3.9325267145172687, + "tokens_seen": 513015808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265295887662989, + "loss": 1.6497, + "theoretical_loss": 3.932473238407889, + "tokens_seen": 513081344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004265195586760281, + "loss": 1.6456, + "theoretical_loss": 3.932419771040856, + "tokens_seen": 513146880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042650952858575724, + "loss": 1.4981, + "theoretical_loss": 3.932366312413623, + "tokens_seen": 513212416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264994984954865, + "loss": 1.5922, + "theoretical_loss": 3.932312862523646, + "tokens_seen": 513277952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264894684052156, + "loss": 1.5702, + "theoretical_loss": 3.9322594213683812, + "tokens_seen": 513343488 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042647943831494484, + "loss": 1.4812, + "theoretical_loss": 3.9322059889452867, + "tokens_seen": 513409024 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426469408224674, + "loss": 1.5984, + "theoretical_loss": 3.9321525652518217, + "tokens_seen": 513474560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264593781344032, + "loss": 1.5915, + "theoretical_loss": 3.9320991502854454, + "tokens_seen": 513540096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264493480441324, + "loss": 1.6555, + "theoretical_loss": 3.932045744043619, + "tokens_seen": 513605632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004264393179538616, + "loss": 1.6434, + "theoretical_loss": 3.9319923465238054, + "tokens_seen": 513671168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042642928786359075, + "loss": 1.4776, + "theoretical_loss": 3.9319389577234665, + "tokens_seen": 513736704 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042641925777332, + "loss": 1.6226, + "theoretical_loss": 3.931885577640066, + "tokens_seen": 513802240 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042640922768304916, + "loss": 1.539, + "theoretical_loss": 3.9318322062710704, + "tokens_seen": 513867776 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042639919759277834, + "loss": 1.4487, + "theoretical_loss": 3.931778843613945, + "tokens_seen": 513933312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263891675025076, + "loss": 1.5655, + "theoretical_loss": 3.9317254896661575, + "tokens_seen": 513998848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263791374122367, + "loss": 1.6407, + "theoretical_loss": 3.9316721444251765, + "tokens_seen": 514064384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042636910732196594, + "loss": 1.5529, + "theoretical_loss": 3.9316188078884706, + "tokens_seen": 514129920 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042635907723169507, + "loss": 1.4951, + "theoretical_loss": 3.931565480053511, + "tokens_seen": 514195456 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263490471414243, + "loss": 1.5838, + "theoretical_loss": 3.9315121609177694, + "tokens_seen": 514260992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263390170511535, + "loss": 1.5612, + "theoretical_loss": 3.931458850478718, + "tokens_seen": 514326528 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042632898696088266, + "loss": 1.5891, + "theoretical_loss": 3.9314055487338297, + "tokens_seen": 514392064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 193343, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4870069026947021, + "objective/train/theoretical_loss": 3.93135225568058, + "objective/train/tokens_used": 534917600, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042631895687061185, + "loss": 1.5851, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004263089267803411, + "loss": 1.5957, + "theoretical_loss": 3.9312989713164455, + "tokens_seen": 514523136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262988966900702, + "loss": 1.5655, + "theoretical_loss": 3.9312456956389017, + "tokens_seen": 514588672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042628886659979944, + "loss": 1.6758, + "theoretical_loss": 3.931192428645427, + "tokens_seen": 514654208 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042627883650952857, + "loss": 1.5096, + "theoretical_loss": 3.9311391703334997, + "tokens_seen": 514719744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262688064192578, + "loss": 1.4992, + "theoretical_loss": 3.9310859207006006, + "tokens_seen": 514785280 + }, + { + "epoch": 0.16, + "learning_rate": 0.000426258776328987, + "loss": 1.6096, + "theoretical_loss": 3.9310326797442103, + "tokens_seen": 514850816 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042624874623871617, + "loss": 1.5603, + "theoretical_loss": 3.930979447461811, + "tokens_seen": 514916352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042623871614844535, + "loss": 1.6004, + "theoretical_loss": 3.9309262238508857, + "tokens_seen": 514981888 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042622868605817453, + "loss": 1.59, + "theoretical_loss": 3.930873008908919, + "tokens_seen": 515047424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004262186559679037, + "loss": 1.6681, + "theoretical_loss": 3.9308198026333954, + "tokens_seen": 515112960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042620862587763295, + "loss": 1.6108, + "theoretical_loss": 3.9307666050218018, + "tokens_seen": 515178496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261985957873621, + "loss": 1.5174, + "theoretical_loss": 3.930713416071625, + "tokens_seen": 515244032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261885656970913, + "loss": 1.5744, + "theoretical_loss": 3.930660235780354, + "tokens_seen": 515309568 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042617853560682044, + "loss": 1.677, + "theoretical_loss": 3.9306070641454776, + "tokens_seen": 515375104 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042616850551654967, + "loss": 1.5542, + "theoretical_loss": 3.9305539011644863, + "tokens_seen": 515440640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042615847542627885, + "loss": 1.5257, + "theoretical_loss": 3.9305007468348725, + "tokens_seen": 515506176 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042614844533600803, + "loss": 1.6597, + "theoretical_loss": 3.9304476011541274, + "tokens_seen": 515571712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261384152457372, + "loss": 1.5597, + "theoretical_loss": 3.930394464119746, + "tokens_seen": 515637248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042612838515546645, + "loss": 1.6519, + "theoretical_loss": 3.9303413357292216, + "tokens_seen": 515702784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261183550651956, + "loss": 1.5617, + "theoretical_loss": 3.930288215980051, + "tokens_seen": 515768320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004261083249749248, + "loss": 1.6126, + "theoretical_loss": 3.9302351048697304, + "tokens_seen": 515833856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042609829488465394, + "loss": 1.5765, + "theoretical_loss": 3.9301820023957577, + "tokens_seen": 515899392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260882647943832, + "loss": 1.5897, + "theoretical_loss": 3.930128908555632, + "tokens_seen": 515964928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042607823470411236, + "loss": 1.6758, + "theoretical_loss": 3.9300758233468525, + "tokens_seen": 516030464 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042606820461384154, + "loss": 1.5612, + "theoretical_loss": 3.9300227467669204, + "tokens_seen": 516096000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260581745235707, + "loss": 1.5881, + "theoretical_loss": 3.929969678813338, + "tokens_seen": 516161536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260481444332999, + "loss": 1.5479, + "theoretical_loss": 3.929916619483608, + "tokens_seen": 516227072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260381143430291, + "loss": 1.4935, + "theoretical_loss": 3.9298635687752337, + "tokens_seen": 516292608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260280842527583, + "loss": 1.4857, + "theoretical_loss": 3.9298105266857215, + "tokens_seen": 516358144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042601805416248744, + "loss": 1.5681, + "theoretical_loss": 3.929757493212577, + "tokens_seen": 516423680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004260080240722167, + "loss": 1.5745, + "theoretical_loss": 3.929704468353307, + "tokens_seen": 516489216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259979939819458, + "loss": 1.5553, + "theoretical_loss": 3.9296514521054196, + "tokens_seen": 516554752 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042598796389167504, + "loss": 1.5519, + "theoretical_loss": 3.929598444466425, + "tokens_seen": 516620288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259779338014042, + "loss": 1.5994, + "theoretical_loss": 3.9295454454338326, + "tokens_seen": 516685824 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259679037111334, + "loss": 1.5233, + "theoretical_loss": 3.9294924550051533, + "tokens_seen": 516751360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259578736208626, + "loss": 1.5221, + "theoretical_loss": 3.9294394731779008, + "tokens_seen": 516816896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259478435305918, + "loss": 1.6084, + "theoretical_loss": 3.929386499949587, + "tokens_seen": 516882432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042593781344032095, + "loss": 1.68, + "theoretical_loss": 3.929333535317727, + "tokens_seen": 516947968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259277833500502, + "loss": 1.443, + "theoretical_loss": 3.929280579279837, + "tokens_seen": 517013504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004259177532597793, + "loss": 1.4398, + "theoretical_loss": 3.929227631833432, + "tokens_seen": 517079040 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042590772316950854, + "loss": 1.5551, + "theoretical_loss": 3.92917469297603, + "tokens_seen": 517144576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258976930792377, + "loss": 1.6002, + "theoretical_loss": 3.9291217627051496, + "tokens_seen": 517210112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258876629889669, + "loss": 1.5054, + "theoretical_loss": 3.9290688410183106, + "tokens_seen": 517275648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258776328986961, + "loss": 1.5997, + "theoretical_loss": 3.9290159279130332, + "tokens_seen": 517341184 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042586760280842527, + "loss": 1.5361, + "theoretical_loss": 3.928963023386839, + "tokens_seen": 517406720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042585757271815445, + "loss": 1.5586, + "theoretical_loss": 3.9289101274372515, + "tokens_seen": 517472256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258475426278837, + "loss": 1.4258, + "theoretical_loss": 3.9288572400617934, + "tokens_seen": 517537792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258375125376128, + "loss": 1.5222, + "theoretical_loss": 3.9288043612579893, + "tokens_seen": 517603328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042582748244734205, + "loss": 1.5875, + "theoretical_loss": 3.928751491023366, + "tokens_seen": 517668864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 194581, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6015167236328125, + "objective/train/theoretical_loss": 3.9286986293554493, + "objective/train/tokens_used": 538194400, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258174523570712, + "loss": 1.557, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004258074222668004, + "loss": 1.5653, + "theoretical_loss": 3.9286457762517673, + "tokens_seen": 517799936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257973921765296, + "loss": 1.5751, + "theoretical_loss": 3.928592931709849, + "tokens_seen": 517865472 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042578736208625877, + "loss": 1.6106, + "theoretical_loss": 3.9285400957272234, + "tokens_seen": 517931008 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042577733199598795, + "loss": 1.5065, + "theoretical_loss": 3.9284872683014225, + "tokens_seen": 517996544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257673019057172, + "loss": 1.5025, + "theoretical_loss": 3.9284344494299774, + "tokens_seen": 518062080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257572718154463, + "loss": 1.616, + "theoretical_loss": 3.9283816391104214, + "tokens_seen": 518127616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042574724172517555, + "loss": 1.5236, + "theoretical_loss": 3.9283288373402883, + "tokens_seen": 518193152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257372116349047, + "loss": 1.5829, + "theoretical_loss": 3.9282760441171134, + "tokens_seen": 518258688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257271815446339, + "loss": 1.4848, + "theoretical_loss": 3.9282232594384316, + "tokens_seen": 518324224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257171514543631, + "loss": 1.4347, + "theoretical_loss": 3.9281704833017814, + "tokens_seen": 518389760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004257071213640923, + "loss": 1.5502, + "theoretical_loss": 3.9281177157047003, + "tokens_seen": 518455296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042569709127382145, + "loss": 1.5138, + "theoretical_loss": 3.928064956644726, + "tokens_seen": 518520832 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042568706118355064, + "loss": 1.5405, + "theoretical_loss": 3.9280122061194005, + "tokens_seen": 518586368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256770310932798, + "loss": 1.5352, + "theoretical_loss": 3.927959464126264, + "tokens_seen": 518651904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042566700100300905, + "loss": 1.5561, + "theoretical_loss": 3.9279067306628583, + "tokens_seen": 518717440 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042565697091273823, + "loss": 1.5047, + "theoretical_loss": 3.9278540057267266, + "tokens_seen": 518782976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256469408224674, + "loss": 1.6708, + "theoretical_loss": 3.927801289315414, + "tokens_seen": 518848512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042563691073219665, + "loss": 1.5124, + "theoretical_loss": 3.9277485814264645, + "tokens_seen": 518914048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004256268806419258, + "loss": 1.4886, + "theoretical_loss": 3.9276958820574253, + "tokens_seen": 518979584 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425616850551655, + "loss": 1.569, + "theoretical_loss": 3.9276431912058425, + "tokens_seen": 519045120 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042560682046138414, + "loss": 1.6174, + "theoretical_loss": 3.9275905088692653, + "tokens_seen": 519110656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255967903711134, + "loss": 1.498, + "theoretical_loss": 3.927537835045242, + "tokens_seen": 519176192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042558676028084256, + "loss": 1.5174, + "theoretical_loss": 3.927485169731323, + "tokens_seen": 519241728 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042557673019057174, + "loss": 1.575, + "theoretical_loss": 3.9274325129250602, + "tokens_seen": 519307264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255667001003009, + "loss": 1.4867, + "theoretical_loss": 3.9273798646240055, + "tokens_seen": 519372800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255566700100301, + "loss": 1.5595, + "theoretical_loss": 3.9273272248257123, + "tokens_seen": 519438336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255466399197593, + "loss": 1.5156, + "theoretical_loss": 3.9272745935277347, + "tokens_seen": 519503872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255366098294885, + "loss": 1.4599, + "theoretical_loss": 3.927221970727628, + "tokens_seen": 519569408 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042552657973921764, + "loss": 1.4816, + "theoretical_loss": 3.927169356422948, + "tokens_seen": 519634944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004255165496489469, + "loss": 1.5013, + "theoretical_loss": 3.9271167506112534, + "tokens_seen": 519700480 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425506519558676, + "loss": 1.4826, + "theoretical_loss": 3.927064153290101, + "tokens_seen": 519766016 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042549648946840524, + "loss": 1.5539, + "theoretical_loss": 3.9270115644570516, + "tokens_seen": 519831552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254864593781344, + "loss": 1.5934, + "theoretical_loss": 3.9269589841096644, + "tokens_seen": 519897088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254764292878636, + "loss": 1.634, + "theoretical_loss": 3.9269064122455006, + "tokens_seen": 519962624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254663991975928, + "loss": 1.4588, + "theoretical_loss": 3.926853848862124, + "tokens_seen": 520028160 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425456369107322, + "loss": 1.4728, + "theoretical_loss": 3.9268012939570967, + "tokens_seen": 520093696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042544633901705115, + "loss": 1.4825, + "theoretical_loss": 3.9267487475279834, + "tokens_seen": 520159232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254363089267804, + "loss": 1.6303, + "theoretical_loss": 3.9266962095723494, + "tokens_seen": 520224768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254262788365095, + "loss": 1.546, + "theoretical_loss": 3.9266436800877615, + "tokens_seen": 520290304 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042541624874623874, + "loss": 1.5555, + "theoretical_loss": 3.926591159071787, + "tokens_seen": 520355840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004254062186559679, + "loss": 1.5196, + "theoretical_loss": 3.9265386465219936, + "tokens_seen": 520421376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253961885656971, + "loss": 1.4961, + "theoretical_loss": 3.9264861424359516, + "tokens_seen": 520486912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253861584754263, + "loss": 1.5978, + "theoretical_loss": 3.9264336468112315, + "tokens_seen": 520552448 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042537612838515547, + "loss": 1.4584, + "theoretical_loss": 3.9263811596454037, + "tokens_seen": 520617984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042536609829488465, + "loss": 1.5606, + "theoretical_loss": 3.9263286809360407, + "tokens_seen": 520683520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253560682046139, + "loss": 1.5294, + "theoretical_loss": 3.9262762106807174, + "tokens_seen": 520749056 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425346038114343, + "loss": 1.4863, + "theoretical_loss": 3.926223748877007, + "tokens_seen": 520814592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042533600802407225, + "loss": 1.4706, + "theoretical_loss": 3.926171295522485, + "tokens_seen": 520880128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253259779338014, + "loss": 1.5603, + "theoretical_loss": 3.9261188506147278, + "tokens_seen": 520945664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 195804, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6601482629776, + "objective/train/theoretical_loss": 3.926066414151313, + "objective/train/tokens_used": 541471200, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253159478435306, + "loss": 1.6084, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004253059177532598, + "loss": 1.611, + "theoretical_loss": 3.926013986129819, + "tokens_seen": 521076736 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042529588766298897, + "loss": 1.5765, + "theoretical_loss": 3.9259615665478256, + "tokens_seen": 521142272 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042528585757271815, + "loss": 1.4864, + "theoretical_loss": 3.9259091554029126, + "tokens_seen": 521207808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252758274824474, + "loss": 1.5988, + "theoretical_loss": 3.9258567526926615, + "tokens_seen": 521273344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252657973921765, + "loss": 1.4925, + "theoretical_loss": 3.9258043584146556, + "tokens_seen": 521338880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042525576730190575, + "loss": 1.5621, + "theoretical_loss": 3.925751972566477, + "tokens_seen": 521404416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252457372116349, + "loss": 1.6711, + "theoretical_loss": 3.9256995951457103, + "tokens_seen": 521469952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252357071213641, + "loss": 1.5288, + "theoretical_loss": 3.925647226149942, + "tokens_seen": 521535488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252256770310933, + "loss": 1.7424, + "theoretical_loss": 3.925594865576758, + "tokens_seen": 521601024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004252156469408225, + "loss": 1.4721, + "theoretical_loss": 3.9255425134237445, + "tokens_seen": 521666560 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042520561685055165, + "loss": 1.6511, + "theoretical_loss": 3.9254901696884916, + "tokens_seen": 521732096 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042519558676028084, + "loss": 1.4804, + "theoretical_loss": 3.925437834368588, + "tokens_seen": 521797632 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042518555667001, + "loss": 1.5503, + "theoretical_loss": 3.925385507461624, + "tokens_seen": 521863168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042517552657973925, + "loss": 1.4834, + "theoretical_loss": 3.9253331889651912, + "tokens_seen": 521928704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251654964894684, + "loss": 1.4836, + "theoretical_loss": 3.9252808788768814, + "tokens_seen": 521994240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251554663991976, + "loss": 1.515, + "theoretical_loss": 3.9252285771942876, + "tokens_seen": 522059776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251454363089268, + "loss": 1.6255, + "theoretical_loss": 3.9251762839150057, + "tokens_seen": 522125312 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425135406218656, + "loss": 1.6474, + "theoretical_loss": 3.9251239990366305, + "tokens_seen": 522190848 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042512537612838516, + "loss": 1.6249, + "theoretical_loss": 3.925071722556757, + "tokens_seen": 522256384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042511534603811434, + "loss": 1.5325, + "theoretical_loss": 3.9250194544729844, + "tokens_seen": 522321920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004251053159478435, + "loss": 1.628, + "theoretical_loss": 3.9249671947829095, + "tokens_seen": 522387456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042509528585757276, + "loss": 1.5054, + "theoretical_loss": 3.924914943484133, + "tokens_seen": 522452992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250852557673019, + "loss": 1.4325, + "theoretical_loss": 3.9248627005742533, + "tokens_seen": 522518528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250752256770311, + "loss": 1.5065, + "theoretical_loss": 3.9248104660508734, + "tokens_seen": 522584064 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042506519558676024, + "loss": 1.6397, + "theoretical_loss": 3.9247582399115943, + "tokens_seen": 522649600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250551654964895, + "loss": 1.5915, + "theoretical_loss": 3.92470602215402, + "tokens_seen": 522715136 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042504513540621866, + "loss": 1.606, + "theoretical_loss": 3.9246538127757544, + "tokens_seen": 522780672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042503510531594784, + "loss": 1.6264, + "theoretical_loss": 3.9246016117744036, + "tokens_seen": 522846208 + }, + { + "epoch": 0.16, + "learning_rate": 0.000425025075225677, + "loss": 1.5289, + "theoretical_loss": 3.924549419147572, + "tokens_seen": 522911744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250150451354062, + "loss": 1.4297, + "theoretical_loss": 3.9244972348928684, + "tokens_seen": 522977280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004250050150451354, + "loss": 1.6237, + "theoretical_loss": 3.9244450590079003, + "tokens_seen": 523042816 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249949849548646, + "loss": 1.5632, + "theoretical_loss": 3.9243928914902764, + "tokens_seen": 523108352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042498495486459375, + "loss": 1.606, + "theoretical_loss": 3.924340732337608, + "tokens_seen": 523173888 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424974924774323, + "loss": 1.6007, + "theoretical_loss": 3.924288581547505, + "tokens_seen": 523239424 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042496489468405216, + "loss": 1.5246, + "theoretical_loss": 3.9242364391175797, + "tokens_seen": 523304960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042495486459378135, + "loss": 1.5215, + "theoretical_loss": 3.924184305045446, + "tokens_seen": 523370496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249448345035105, + "loss": 1.4807, + "theoretical_loss": 3.9241321793287174, + "tokens_seen": 523436032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249348044132397, + "loss": 1.5945, + "theoretical_loss": 3.9240800619650082, + "tokens_seen": 523501568 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042492477432296894, + "loss": 1.5604, + "theoretical_loss": 3.9240279529519357, + "tokens_seen": 523567104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249147442326981, + "loss": 1.6932, + "theoretical_loss": 3.9239758522871164, + "tokens_seen": 523632640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004249047141424273, + "loss": 1.494, + "theoretical_loss": 3.923923759968168, + "tokens_seen": 523698176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248946840521565, + "loss": 1.5151, + "theoretical_loss": 3.9238716759927095, + "tokens_seen": 523763712 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042488465396188567, + "loss": 1.5815, + "theoretical_loss": 3.923819600358361, + "tokens_seen": 523829248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042487462387161485, + "loss": 1.696, + "theoretical_loss": 3.923767533062743, + "tokens_seen": 523894784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248645937813441, + "loss": 1.472, + "theoretical_loss": 3.9237154741034783, + "tokens_seen": 523960320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248545636910732, + "loss": 1.5753, + "theoretical_loss": 3.923663423478189, + "tokens_seen": 524025856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042484453360080245, + "loss": 1.6162, + "theoretical_loss": 3.9236113811844984, + "tokens_seen": 524091392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248345035105316, + "loss": 1.5474, + "theoretical_loss": 3.9235593472200323, + "tokens_seen": 524156928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004248244734202608, + "loss": 1.5528, + "theoretical_loss": 3.9235073215824166, + "tokens_seen": 524222464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 196964, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4527997970581055, + "objective/train/theoretical_loss": 3.9234553042692766, + "objective/train/tokens_used": 544748000, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042481444332999, + "loss": 1.4901, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042480441323971917, + "loss": 1.543, + "theoretical_loss": 3.9234032952782414, + "tokens_seen": 524353536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042479438314944835, + "loss": 1.6199, + "theoretical_loss": 3.923351294606939, + "tokens_seen": 524419072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247843530591776, + "loss": 1.5538, + "theoretical_loss": 3.923299302253, + "tokens_seen": 524484608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247743229689067, + "loss": 1.5247, + "theoretical_loss": 3.923247318214054, + "tokens_seen": 524550144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042476429287863595, + "loss": 1.4545, + "theoretical_loss": 3.923195342487733, + "tokens_seen": 524615680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247542627883651, + "loss": 1.5703, + "theoretical_loss": 3.9231433750716693, + "tokens_seen": 524681216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247442326980943, + "loss": 1.5786, + "theoretical_loss": 3.9230914159634964, + "tokens_seen": 524746752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247342026078235, + "loss": 1.4086, + "theoretical_loss": 3.9230394651608496, + "tokens_seen": 524812288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004247241725175527, + "loss": 1.7152, + "theoretical_loss": 3.922987522661363, + "tokens_seen": 524877824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042471414242728186, + "loss": 1.5665, + "theoretical_loss": 3.922935588462675, + "tokens_seen": 524943360 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042470411233701104, + "loss": 1.401, + "theoretical_loss": 3.9228836625624206, + "tokens_seen": 525008896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246940822467402, + "loss": 1.5402, + "theoretical_loss": 3.9228317449582404, + "tokens_seen": 525074432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042468405215646945, + "loss": 1.5158, + "theoretical_loss": 3.9227798356477726, + "tokens_seen": 525139968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246740220661986, + "loss": 1.5477, + "theoretical_loss": 3.922727934628658, + "tokens_seen": 525205504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246639919759278, + "loss": 1.5831, + "theoretical_loss": 3.922676041898537, + "tokens_seen": 525271040 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424653961885657, + "loss": 1.6195, + "theoretical_loss": 3.9226241574550533, + "tokens_seen": 525336576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246439317953862, + "loss": 1.5942, + "theoretical_loss": 3.9225722812958486, + "tokens_seen": 525402112 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042463390170511536, + "loss": 1.6761, + "theoretical_loss": 3.922520413418568, + "tokens_seen": 525467648 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042462387161484454, + "loss": 1.5967, + "theoretical_loss": 3.922468553820856, + "tokens_seen": 525533184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004246138415245737, + "loss": 1.6083, + "theoretical_loss": 3.9224167025003593, + "tokens_seen": 525598720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042460381143430296, + "loss": 1.5904, + "theoretical_loss": 3.922364859454725, + "tokens_seen": 525664256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245937813440321, + "loss": 1.4963, + "theoretical_loss": 3.9223130246816007, + "tokens_seen": 525729792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245837512537613, + "loss": 1.5155, + "theoretical_loss": 3.9222611981786355, + "tokens_seen": 525795328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042457372116349044, + "loss": 1.5699, + "theoretical_loss": 3.92220937994348, + "tokens_seen": 525860864 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245636910732197, + "loss": 1.5017, + "theoretical_loss": 3.9221575699737836, + "tokens_seen": 525926400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042455366098294886, + "loss": 1.5662, + "theoretical_loss": 3.9221057682672, + "tokens_seen": 525991936 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042454363089267804, + "loss": 1.5877, + "theoretical_loss": 3.92205397482138, + "tokens_seen": 526057472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245336008024072, + "loss": 1.4476, + "theoretical_loss": 3.9220021896339796, + "tokens_seen": 526123008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245235707121364, + "loss": 1.624, + "theoretical_loss": 3.921950412702652, + "tokens_seen": 526188544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245135406218656, + "loss": 1.5433, + "theoretical_loss": 3.9218986440250534, + "tokens_seen": 526254080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004245035105315948, + "loss": 1.6154, + "theoretical_loss": 3.9218468835988403, + "tokens_seen": 526319616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042449348044132395, + "loss": 1.5965, + "theoretical_loss": 3.9217951314216704, + "tokens_seen": 526385152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244834503510532, + "loss": 1.6056, + "theoretical_loss": 3.9217433874912024, + "tokens_seen": 526450688 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042447342026078236, + "loss": 1.4896, + "theoretical_loss": 3.921691651805096, + "tokens_seen": 526516224 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042446339017051155, + "loss": 1.5685, + "theoretical_loss": 3.9216399243610116, + "tokens_seen": 526581760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244533600802407, + "loss": 1.5532, + "theoretical_loss": 3.9215882051566098, + "tokens_seen": 526647296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244433299899699, + "loss": 1.5568, + "theoretical_loss": 3.921536494189554, + "tokens_seen": 526712832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244332998996991, + "loss": 1.6239, + "theoretical_loss": 3.921484791457507, + "tokens_seen": 526778368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244232698094283, + "loss": 1.4922, + "theoretical_loss": 3.921433096958134, + "tokens_seen": 526843904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042441323971915745, + "loss": 1.5217, + "theoretical_loss": 3.921381410689099, + "tokens_seen": 526909440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004244032096288867, + "loss": 1.5687, + "theoretical_loss": 3.9213297326480694, + "tokens_seen": 526974976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243931795386158, + "loss": 1.4837, + "theoretical_loss": 3.9212780628327115, + "tokens_seen": 527040512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042438314944834505, + "loss": 1.6069, + "theoretical_loss": 3.921226401240693, + "tokens_seen": 527106048 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042437311935807423, + "loss": 1.49, + "theoretical_loss": 3.9211747478696846, + "tokens_seen": 527171584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243630892678034, + "loss": 1.5112, + "theoretical_loss": 3.9211231027173548, + "tokens_seen": 527237120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243530591775326, + "loss": 1.595, + "theoretical_loss": 3.921071465781375, + "tokens_seen": 527302656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243430290872618, + "loss": 1.5897, + "theoretical_loss": 3.9210198370594176, + "tokens_seen": 527368192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042433299899699095, + "loss": 1.529, + "theoretical_loss": 3.920968216549155, + "tokens_seen": 527433728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243229689067202, + "loss": 1.5616, + "theoretical_loss": 3.9209166042482613, + "tokens_seen": 527499264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 198271, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4213464260101318, + "objective/train/theoretical_loss": 3.9208650001544108, + "objective/train/tokens_used": 548024800, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004243129388164493, + "loss": 1.4976, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042430290872617855, + "loss": 1.5854, + "theoretical_loss": 3.920813404265279, + "tokens_seen": 527630336 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042429287863590773, + "loss": 1.5428, + "theoretical_loss": 3.920761816578543, + "tokens_seen": 527695872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242828485456369, + "loss": 1.5982, + "theoretical_loss": 3.920710237091881, + "tokens_seen": 527761408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242728184553661, + "loss": 1.4906, + "theoretical_loss": 3.9206586658029714, + "tokens_seen": 527826944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242627883650953, + "loss": 1.6426, + "theoretical_loss": 3.920607102709493, + "tokens_seen": 527892480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042425275827482446, + "loss": 1.5411, + "theoretical_loss": 3.9205555478091254, + "tokens_seen": 527958016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242427281845537, + "loss": 1.6091, + "theoretical_loss": 3.920504001099552, + "tokens_seen": 528023552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242326980942828, + "loss": 1.5442, + "theoretical_loss": 3.920452462578454, + "tokens_seen": 528089088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042422266800401206, + "loss": 1.5818, + "theoretical_loss": 3.920400932243515, + "tokens_seen": 528154624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242126379137412, + "loss": 1.4645, + "theoretical_loss": 3.920349410092419, + "tokens_seen": 528220160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004242026078234704, + "loss": 1.5695, + "theoretical_loss": 3.920297896122851, + "tokens_seen": 528285696 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241925777331996, + "loss": 1.5888, + "theoretical_loss": 3.9202463903324976, + "tokens_seen": 528351232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241825476429288, + "loss": 1.553, + "theoretical_loss": 3.920194892719045, + "tokens_seen": 528416768 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424172517552658, + "loss": 1.5544, + "theoretical_loss": 3.920143403280182, + "tokens_seen": 528482304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241624874623872, + "loss": 1.487, + "theoretical_loss": 3.9200919220135972, + "tokens_seen": 528547840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241524573721164, + "loss": 1.588, + "theoretical_loss": 3.920040448916981, + "tokens_seen": 528613376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042414242728184556, + "loss": 1.5738, + "theoretical_loss": 3.919988983988023, + "tokens_seen": 528678912 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042413239719157474, + "loss": 1.4709, + "theoretical_loss": 3.919937527224416, + "tokens_seen": 528744448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241223671013039, + "loss": 1.5319, + "theoretical_loss": 3.9198860786238523, + "tokens_seen": 528809984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042411233701103316, + "loss": 1.6694, + "theoretical_loss": 3.919834638184025, + "tokens_seen": 528875520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004241023069207623, + "loss": 1.5829, + "theoretical_loss": 3.91978320590263, + "tokens_seen": 528941056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240922768304915, + "loss": 1.5883, + "theoretical_loss": 3.919731781777361, + "tokens_seen": 529006592 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042408224674022065, + "loss": 1.5705, + "theoretical_loss": 3.919680365805916, + "tokens_seen": 529072128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240722166499499, + "loss": 1.4977, + "theoretical_loss": 3.9196289579859913, + "tokens_seen": 529137664 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042406218655967906, + "loss": 1.5246, + "theoretical_loss": 3.919577558315286, + "tokens_seen": 529203200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042405215646940824, + "loss": 1.5942, + "theoretical_loss": 3.9195261667914982, + "tokens_seen": 529268736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240421263791374, + "loss": 1.4939, + "theoretical_loss": 3.91947478341233, + "tokens_seen": 529334272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240320962888666, + "loss": 1.5451, + "theoretical_loss": 3.9194234081754797, + "tokens_seen": 529399808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004240220661985958, + "loss": 1.5226, + "theoretical_loss": 3.9193720410786517, + "tokens_seen": 529465344 + }, + { + "epoch": 0.16, + "learning_rate": 0.000424012036108325, + "loss": 1.7204, + "theoretical_loss": 3.9193206821195483, + "tokens_seen": 529530880 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042400200601805415, + "loss": 1.469, + "theoretical_loss": 3.9192693312958733, + "tokens_seen": 529596416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239919759277834, + "loss": 1.5676, + "theoretical_loss": 3.919217988605331, + "tokens_seen": 529661952 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042398194583751256, + "loss": 1.583, + "theoretical_loss": 3.919166654045628, + "tokens_seen": 529727488 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042397191574724175, + "loss": 1.611, + "theoretical_loss": 3.919115327614471, + "tokens_seen": 529793024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239618856569709, + "loss": 1.6209, + "theoretical_loss": 3.9190640093095666, + "tokens_seen": 529858560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239518555667001, + "loss": 1.6334, + "theoretical_loss": 3.9190126991286243, + "tokens_seen": 529924096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239418254764293, + "loss": 1.552, + "theoretical_loss": 3.918961397069353, + "tokens_seen": 529989632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239317953861585, + "loss": 1.6576, + "theoretical_loss": 3.9189101031294635, + "tokens_seen": 530055168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042392176529588765, + "loss": 1.4259, + "theoretical_loss": 3.918858817306667, + "tokens_seen": 530120704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004239117352056169, + "loss": 1.4252, + "theoretical_loss": 3.918807539598676, + "tokens_seen": 530186240 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423901705115346, + "loss": 1.5178, + "theoretical_loss": 3.918756270003203, + "tokens_seen": 530251776 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042389167502507525, + "loss": 1.563, + "theoretical_loss": 3.9187050085179633, + "tokens_seen": 530317312 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042388164493480443, + "loss": 1.5342, + "theoretical_loss": 3.918653755140671, + "tokens_seen": 530382848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238716148445336, + "loss": 1.4763, + "theoretical_loss": 3.918602509869042, + "tokens_seen": 530448384 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238615847542628, + "loss": 1.5642, + "theoretical_loss": 3.9185512727007934, + "tokens_seen": 530513920 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423851554663992, + "loss": 1.5455, + "theoretical_loss": 3.9185000436336432, + "tokens_seen": 530579456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042384152457372115, + "loss": 1.5813, + "theoretical_loss": 3.9184488226653102, + "tokens_seen": 530644992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238314944834504, + "loss": 1.546, + "theoretical_loss": 3.918397609793514, + "tokens_seen": 530710528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004238214643931795, + "loss": 1.534, + "theoretical_loss": 3.9183464050159746, + "tokens_seen": 530776064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 199409, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5682408809661865, + "objective/train/theoretical_loss": 3.918295208330415, + "objective/train/tokens_used": 551301600, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042381143430290875, + "loss": 1.5232, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042380140421263793, + "loss": 1.5652, + "theoretical_loss": 3.918244019734556, + "tokens_seen": 530907136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237913741223671, + "loss": 1.5359, + "theoretical_loss": 3.9181928392261214, + "tokens_seen": 530972672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237813440320963, + "loss": 1.441, + "theoretical_loss": 3.9181416668028355, + "tokens_seen": 531038208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237713139418255, + "loss": 1.4749, + "theoretical_loss": 3.918090502462425, + "tokens_seen": 531103744 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042376128385155466, + "loss": 1.6501, + "theoretical_loss": 3.9180393462026135, + "tokens_seen": 531169280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237512537612839, + "loss": 1.6137, + "theoretical_loss": 3.9179881980211295, + "tokens_seen": 531234816 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423741223671013, + "loss": 1.6057, + "theoretical_loss": 3.917937057915701, + "tokens_seen": 531300352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042373119358074226, + "loss": 1.5172, + "theoretical_loss": 3.9178859258840566, + "tokens_seen": 531365888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237211634904714, + "loss": 1.527, + "theoretical_loss": 3.9178348019239264, + "tokens_seen": 531431424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237111334002006, + "loss": 1.5952, + "theoretical_loss": 3.91778368603304, + "tokens_seen": 531496960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004237011033099298, + "loss": 1.6898, + "theoretical_loss": 3.9177325782091303, + "tokens_seen": 531562496 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423691073219659, + "loss": 1.5875, + "theoretical_loss": 3.9176814784499303, + "tokens_seen": 531628032 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042368104312938816, + "loss": 1.5623, + "theoretical_loss": 3.9176303867531717, + "tokens_seen": 531693568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236710130391174, + "loss": 1.6398, + "theoretical_loss": 3.9175793031165895, + "tokens_seen": 531759104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236609829488465, + "loss": 1.6449, + "theoretical_loss": 3.91752822753792, + "tokens_seen": 531824640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042365095285857576, + "loss": 1.4635, + "theoretical_loss": 3.917477160014898, + "tokens_seen": 531890176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236409227683049, + "loss": 1.5079, + "theoretical_loss": 3.9174261005452617, + "tokens_seen": 531955712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236308926780341, + "loss": 1.5723, + "theoretical_loss": 3.917375049126749, + "tokens_seen": 532021248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236208625877633, + "loss": 1.4704, + "theoretical_loss": 3.9173240057570986, + "tokens_seen": 532086784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004236108324974925, + "loss": 1.5043, + "theoretical_loss": 3.91727297043405, + "tokens_seen": 532152320 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042360080240722166, + "loss": 1.4368, + "theoretical_loss": 3.917221943155345, + "tokens_seen": 532217856 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042359077231695085, + "loss": 1.4995, + "theoretical_loss": 3.9171709239187247, + "tokens_seen": 532283392 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042358074222668, + "loss": 1.7532, + "theoretical_loss": 3.9171199127219314, + "tokens_seen": 532348928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042357071213640926, + "loss": 1.5334, + "theoretical_loss": 3.917068909562709, + "tokens_seen": 532414464 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004235606820461384, + "loss": 1.6178, + "theoretical_loss": 3.917017914438802, + "tokens_seen": 532480000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004235506519558676, + "loss": 1.6668, + "theoretical_loss": 3.9169669273479553, + "tokens_seen": 532545536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042354062186559675, + "loss": 1.6075, + "theoretical_loss": 3.916915948287916, + "tokens_seen": 532611072 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423530591775326, + "loss": 1.5483, + "theoretical_loss": 3.91686497725643, + "tokens_seen": 532676608 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042352056168505517, + "loss": 1.4589, + "theoretical_loss": 3.916814014251247, + "tokens_seen": 532742144 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042351053159478435, + "loss": 1.5757, + "theoretical_loss": 3.9167630592701146, + "tokens_seen": 532807680 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042350050150451353, + "loss": 1.6018, + "theoretical_loss": 3.9167121123107833, + "tokens_seen": 532873216 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042349047141424276, + "loss": 1.5502, + "theoretical_loss": 3.9166611733710037, + "tokens_seen": 532938752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234804413239719, + "loss": 1.5932, + "theoretical_loss": 3.9166102424485274, + "tokens_seen": 533004288 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042347041123370113, + "loss": 1.5468, + "theoretical_loss": 3.9165593195411077, + "tokens_seen": 533069824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042346038114343025, + "loss": 1.5906, + "theoretical_loss": 3.916508404646497, + "tokens_seen": 533135360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234503510531595, + "loss": 1.5446, + "theoretical_loss": 3.9164574977624507, + "tokens_seen": 533200896 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042344032096288867, + "loss": 1.5697, + "theoretical_loss": 3.916406598886724, + "tokens_seen": 533266432 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042343029087261785, + "loss": 1.5303, + "theoretical_loss": 3.916355708017073, + "tokens_seen": 533331968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234202607823471, + "loss": 1.5658, + "theoretical_loss": 3.916304825151254, + "tokens_seen": 533397504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004234102306920762, + "loss": 1.5521, + "theoretical_loss": 3.916253950287026, + "tokens_seen": 533463040 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042340020060180545, + "loss": 1.55, + "theoretical_loss": 3.916203083422148, + "tokens_seen": 533528576 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042339017051153463, + "loss": 1.5277, + "theoretical_loss": 3.91615222455438, + "tokens_seen": 533594112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233801404212638, + "loss": 1.6067, + "theoretical_loss": 3.9161013736814816, + "tokens_seen": 533659648 + }, + { + "epoch": 0.16, + "learning_rate": 0.000423370110330993, + "loss": 1.549, + "theoretical_loss": 3.916050530801215, + "tokens_seen": 533725184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233600802407222, + "loss": 1.5647, + "theoretical_loss": 3.915999695911343, + "tokens_seen": 533790720 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042335005015045135, + "loss": 1.5042, + "theoretical_loss": 3.91594886900963, + "tokens_seen": 533856256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233400200601806, + "loss": 1.6357, + "theoretical_loss": 3.9158980500938383, + "tokens_seen": 533921792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004233299899699097, + "loss": 1.5706, + "theoretical_loss": 3.9158472391617343, + "tokens_seen": 533987328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042331995987963895, + "loss": 1.5933, + "theoretical_loss": 3.9157964362110844, + "tokens_seen": 534052864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 200516, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6013188362121582, + "objective/train/theoretical_loss": 3.9157456412396554, + "objective/train/tokens_used": 554578400, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042330992978936813, + "loss": 1.5404, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232998996990973, + "loss": 1.6596, + "theoretical_loss": 3.9156948542452144, + "tokens_seen": 534183936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232898696088265, + "loss": 1.6456, + "theoretical_loss": 3.915644075225532, + "tokens_seen": 534249472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232798395185557, + "loss": 1.5683, + "theoretical_loss": 3.915593304178376, + "tokens_seen": 534315008 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042326980942828486, + "loss": 1.5816, + "theoretical_loss": 3.9155425411015186, + "tokens_seen": 534380544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232597793380141, + "loss": 1.6126, + "theoretical_loss": 3.9154917859927307, + "tokens_seen": 534446080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232497492477432, + "loss": 1.59, + "theoretical_loss": 3.9154410388497842, + "tokens_seen": 534511616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042323971915747246, + "loss": 1.575, + "theoretical_loss": 3.9153902996704533, + "tokens_seen": 534577152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232296890672016, + "loss": 1.4902, + "theoretical_loss": 3.915339568452512, + "tokens_seen": 534642688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004232196589769308, + "loss": 1.5407, + "theoretical_loss": 3.9152888451937358, + "tokens_seen": 534708224 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042320962888666, + "loss": 1.5464, + "theoretical_loss": 3.9152381298919, + "tokens_seen": 534773760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231995987963892, + "loss": 1.6761, + "theoretical_loss": 3.915187422544782, + "tokens_seen": 534839296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042318956870611836, + "loss": 1.4779, + "theoretical_loss": 3.915136723150159, + "tokens_seen": 534904832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231795386158476, + "loss": 1.5589, + "theoretical_loss": 3.91508603170581, + "tokens_seen": 534970368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231695085255767, + "loss": 1.5314, + "theoretical_loss": 3.915035348209514, + "tokens_seen": 535035904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042315947843530596, + "loss": 1.5176, + "theoretical_loss": 3.9149846726590534, + "tokens_seen": 535101440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231494483450351, + "loss": 1.5883, + "theoretical_loss": 3.914934005052208, + "tokens_seen": 535166976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231394182547643, + "loss": 1.6274, + "theoretical_loss": 3.91488334538676, + "tokens_seen": 535232512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231293881644935, + "loss": 1.512, + "theoretical_loss": 3.914832693660493, + "tokens_seen": 535298048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004231193580742227, + "loss": 1.4823, + "theoretical_loss": 3.9147820498711914, + "tokens_seen": 535363584 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042310932798395186, + "loss": 1.479, + "theoretical_loss": 3.91473141401664, + "tokens_seen": 535429120 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042309929789368105, + "loss": 1.5518, + "theoretical_loss": 3.9146807860946238, + "tokens_seen": 535494656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230892678034102, + "loss": 1.5202, + "theoretical_loss": 3.91463016610293, + "tokens_seen": 535560192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042307923771313946, + "loss": 1.5836, + "theoretical_loss": 3.9145795540393467, + "tokens_seen": 535625728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230692076228686, + "loss": 1.5517, + "theoretical_loss": 3.9145289499016616, + "tokens_seen": 535691264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230591775325978, + "loss": 1.5696, + "theoretical_loss": 3.9144783536876644, + "tokens_seen": 535756800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042304914744232695, + "loss": 1.4902, + "theoretical_loss": 3.9144277653951454, + "tokens_seen": 535822336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004230391173520562, + "loss": 1.6386, + "theoretical_loss": 3.914377185021896, + "tokens_seen": 535887872 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042302908726178537, + "loss": 1.6163, + "theoretical_loss": 3.914326612565708, + "tokens_seen": 535953408 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042301905717151455, + "loss": 1.5888, + "theoretical_loss": 3.914276048024374, + "tokens_seen": 536018944 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042300902708124373, + "loss": 1.5462, + "theoretical_loss": 3.9142254913956886, + "tokens_seen": 536084480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042299899699097297, + "loss": 1.5354, + "theoretical_loss": 3.914174942677446, + "tokens_seen": 536150016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229889669007021, + "loss": 1.506, + "theoretical_loss": 3.9141244018674413, + "tokens_seen": 536215552 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042297893681043133, + "loss": 1.5983, + "theoretical_loss": 3.9140738689634715, + "tokens_seen": 536281088 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042296890672016045, + "loss": 1.5354, + "theoretical_loss": 3.914023343963334, + "tokens_seen": 536346624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229588766298897, + "loss": 1.6333, + "theoretical_loss": 3.9139728268648266, + "tokens_seen": 536412160 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042294884653961887, + "loss": 1.5363, + "theoretical_loss": 3.913922317665749, + "tokens_seen": 536477696 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042293881644934805, + "loss": 1.4749, + "theoretical_loss": 3.9138718163639012, + "tokens_seen": 536543232 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042292878635907723, + "loss": 1.4632, + "theoretical_loss": 3.9138213229570837, + "tokens_seen": 536608768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229187562688064, + "loss": 1.5462, + "theoretical_loss": 3.9137708374430984, + "tokens_seen": 536674304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004229087261785356, + "loss": 1.6263, + "theoretical_loss": 3.913720359819748, + "tokens_seen": 536739840 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042289869608826483, + "loss": 1.5275, + "theoretical_loss": 3.913669890084835, + "tokens_seen": 536805376 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042288866599799396, + "loss": 1.5672, + "theoretical_loss": 3.9136194282361654, + "tokens_seen": 536870912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228786359077232, + "loss": 1.4644, + "theoretical_loss": 3.913568974271543, + "tokens_seen": 536936448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228686058174523, + "loss": 1.5171, + "theoretical_loss": 3.913518528188775, + "tokens_seen": 537001984 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042285857572718155, + "loss": 1.4966, + "theoretical_loss": 3.9134680899856686, + "tokens_seen": 537067520 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042284854563691074, + "loss": 1.5225, + "theoretical_loss": 3.913417659660031, + "tokens_seen": 537133056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228385155466399, + "loss": 1.4982, + "theoretical_loss": 3.913367237209671, + "tokens_seen": 537198592 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004228284854563691, + "loss": 1.5745, + "theoretical_loss": 3.9133168226323987, + "tokens_seen": 537264128 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042281845536609833, + "loss": 1.56, + "theoretical_loss": 3.913266415926024, + "tokens_seen": 537329664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 201641, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4612195491790771, + "objective/train/theoretical_loss": 3.9132160170883585, + "objective/train/tokens_used": 557855200, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042280842527582746, + "loss": 1.4637, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227983951855567, + "loss": 1.5777, + "theoretical_loss": 3.913165626117215, + "tokens_seen": 537460736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227883650952858, + "loss": 1.6045, + "theoretical_loss": 3.913115243010407, + "tokens_seen": 537526272 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042277833500501506, + "loss": 1.4689, + "theoretical_loss": 3.9130648677657476, + "tokens_seen": 537591808 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042276830491474424, + "loss": 1.5086, + "theoretical_loss": 3.9130145003810517, + "tokens_seen": 537657344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227582748244734, + "loss": 1.6041, + "theoretical_loss": 3.912964140854135, + "tokens_seen": 537722880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227482447342026, + "loss": 1.601, + "theoretical_loss": 3.9129137891828147, + "tokens_seen": 537788416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227382146439318, + "loss": 1.5575, + "theoretical_loss": 3.9128634453649083, + "tokens_seen": 537853952 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042272818455366096, + "loss": 1.5982, + "theoretical_loss": 3.9128131093982343, + "tokens_seen": 537919488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227181544633902, + "loss": 1.6205, + "theoretical_loss": 3.9127627812806116, + "tokens_seen": 537985024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004227081243731193, + "loss": 1.5685, + "theoretical_loss": 3.91271246100986, + "tokens_seen": 538050560 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042269809428284856, + "loss": 1.585, + "theoretical_loss": 3.9126621485838013, + "tokens_seen": 538116096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226880641925777, + "loss": 1.4407, + "theoretical_loss": 3.9126118440002573, + "tokens_seen": 538181632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226780341023069, + "loss": 1.5605, + "theoretical_loss": 3.9125615472570505, + "tokens_seen": 538247168 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042266800401203616, + "loss": 1.6605, + "theoretical_loss": 3.912511258352004, + "tokens_seen": 538312704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226579739217653, + "loss": 1.6448, + "theoretical_loss": 3.9124609772829437, + "tokens_seen": 538378240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226479438314945, + "loss": 1.4814, + "theoretical_loss": 3.9124107040476934, + "tokens_seen": 538443776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226379137412237, + "loss": 1.5203, + "theoretical_loss": 3.91236043864408, + "tokens_seen": 538509312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004226278836509529, + "loss": 1.5284, + "theoretical_loss": 3.912310181069931, + "tokens_seen": 538574848 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042261785356068206, + "loss": 1.6965, + "theoretical_loss": 3.9122599313230735, + "tokens_seen": 538640384 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042260782347041125, + "loss": 1.5123, + "theoretical_loss": 3.9122096894013376, + "tokens_seen": 538705920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225977933801404, + "loss": 1.6106, + "theoretical_loss": 3.9121594553025516, + "tokens_seen": 538771456 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042258776328986966, + "loss": 1.5321, + "theoretical_loss": 3.912109229024547, + "tokens_seen": 538836992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225777331995988, + "loss": 1.5334, + "theoretical_loss": 3.9120590105651547, + "tokens_seen": 538902528 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422567703109328, + "loss": 1.5892, + "theoretical_loss": 3.9120087999222073, + "tokens_seen": 538968064 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042255767301905715, + "loss": 1.4349, + "theoretical_loss": 3.9119585970935375, + "tokens_seen": 539033600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004225476429287864, + "loss": 1.4747, + "theoretical_loss": 3.91190840207698, + "tokens_seen": 539099136 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042253761283851557, + "loss": 1.518, + "theoretical_loss": 3.9118582148703696, + "tokens_seen": 539164672 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042252758274824475, + "loss": 1.6117, + "theoretical_loss": 3.9118080354715414, + "tokens_seen": 539230208 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042251755265797393, + "loss": 1.632, + "theoretical_loss": 3.911757863878333, + "tokens_seen": 539295744 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042250752256770317, + "loss": 1.6862, + "theoretical_loss": 3.9117077000885807, + "tokens_seen": 539361280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224974924774323, + "loss": 1.5505, + "theoretical_loss": 3.911657544100124, + "tokens_seen": 539426816 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042248746238716153, + "loss": 1.479, + "theoretical_loss": 3.911607395910801, + "tokens_seen": 539492352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042247743229689065, + "loss": 1.5564, + "theoretical_loss": 3.911557255518453, + "tokens_seen": 539557888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224674022066199, + "loss": 1.5805, + "theoretical_loss": 3.9115071229209195, + "tokens_seen": 539623424 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042245737211634907, + "loss": 1.5894, + "theoretical_loss": 3.9114569981160434, + "tokens_seen": 539688960 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042244734202607825, + "loss": 1.5456, + "theoretical_loss": 3.9114068811016662, + "tokens_seen": 539754496 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042243731193580743, + "loss": 1.6299, + "theoretical_loss": 3.9113567718756332, + "tokens_seen": 539820032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224272818455366, + "loss": 1.557, + "theoretical_loss": 3.9113066704357866, + "tokens_seen": 539885568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004224172517552658, + "loss": 1.6334, + "theoretical_loss": 3.9112565767799734, + "tokens_seen": 539951104 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042240722166499503, + "loss": 1.6197, + "theoretical_loss": 3.9112064909060393, + "tokens_seen": 540016640 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042239719157472416, + "loss": 1.5701, + "theoretical_loss": 3.91115641281183, + "tokens_seen": 540082176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223871614844534, + "loss": 1.4901, + "theoretical_loss": 3.911106342495195, + "tokens_seen": 540147712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223771313941825, + "loss": 1.7504, + "theoretical_loss": 3.911056279953981, + "tokens_seen": 540213248 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042236710130391176, + "loss": 1.4901, + "theoretical_loss": 3.91100622518604, + "tokens_seen": 540278784 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042235707121364094, + "loss": 1.6664, + "theoretical_loss": 3.9109561781892204, + "tokens_seen": 540344320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223470411233701, + "loss": 1.5887, + "theoretical_loss": 3.9109061389613737, + "tokens_seen": 540409856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223370110330993, + "loss": 1.4987, + "theoretical_loss": 3.9108561075003525, + "tokens_seen": 540475392 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042232698094282853, + "loss": 1.6487, + "theoretical_loss": 3.9108060838040095, + "tokens_seen": 540540928 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042231695085255766, + "loss": 1.4912, + "theoretical_loss": 3.9107560678701985, + "tokens_seen": 540606464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 202818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5054774284362793, + "objective/train/theoretical_loss": 3.910706059696774, + "objective/train/tokens_used": 561132000, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004223069207622869, + "loss": 1.5469, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422296890672016, + "loss": 1.619, + "theoretical_loss": 3.910656059281591, + "tokens_seen": 540737536 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042228686058174526, + "loss": 1.5203, + "theoretical_loss": 3.910606066622507, + "tokens_seen": 540803072 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042227683049147444, + "loss": 1.5296, + "theoretical_loss": 3.9105560817173783, + "tokens_seen": 540868608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222668004012036, + "loss": 1.5378, + "theoretical_loss": 3.910506104564063, + "tokens_seen": 540934144 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222567703109328, + "loss": 1.5189, + "theoretical_loss": 3.9104561351604206, + "tokens_seen": 540999680 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422246740220662, + "loss": 1.474, + "theoretical_loss": 3.9104061735043105, + "tokens_seen": 541065216 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042223671013039116, + "loss": 1.5159, + "theoretical_loss": 3.9103562195935933, + "tokens_seen": 541130752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222266800401204, + "loss": 1.675, + "theoretical_loss": 3.91030627342613, + "tokens_seen": 541196288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004222166499498495, + "loss": 1.6715, + "theoretical_loss": 3.910256334999783, + "tokens_seen": 541261824 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042220661985957876, + "loss": 1.6124, + "theoretical_loss": 3.910206404312416, + "tokens_seen": 541327360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221965897693079, + "loss": 1.5807, + "theoretical_loss": 3.9101564813618923, + "tokens_seen": 541392896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221865596790371, + "loss": 1.5733, + "theoretical_loss": 3.9101065661460774, + "tokens_seen": 541458432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221765295887663, + "loss": 1.5652, + "theoretical_loss": 3.9100566586628362, + "tokens_seen": 541523968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221664994984955, + "loss": 1.652, + "theoretical_loss": 3.9100067589100362, + "tokens_seen": 541589504 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042215646940822467, + "loss": 1.6588, + "theoretical_loss": 3.909956866885544, + "tokens_seen": 541655040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221464393179539, + "loss": 1.5535, + "theoretical_loss": 3.909906982587228, + "tokens_seen": 541720576 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042213640922768303, + "loss": 1.5869, + "theoretical_loss": 3.909857106012957, + "tokens_seen": 541786112 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042212637913741226, + "loss": 1.4648, + "theoretical_loss": 3.9098072371606016, + "tokens_seen": 541851648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221163490471414, + "loss": 1.6911, + "theoretical_loss": 3.9097573760280326, + "tokens_seen": 541917184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004221063189568706, + "loss": 1.5971, + "theoretical_loss": 3.90970752261312, + "tokens_seen": 541982720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220962888665998, + "loss": 1.588, + "theoretical_loss": 3.9096576769137386, + "tokens_seen": 542048256 + }, + { + "epoch": 0.16, + "learning_rate": 0.000422086258776329, + "loss": 1.4804, + "theoretical_loss": 3.90960783892776, + "tokens_seen": 542113792 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042207622868605817, + "loss": 1.5348, + "theoretical_loss": 3.9095580086530592, + "tokens_seen": 542179328 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042206619859578735, + "loss": 1.5914, + "theoretical_loss": 3.90950818608751, + "tokens_seen": 542244864 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042205616850551653, + "loss": 1.5664, + "theoretical_loss": 3.9094583712289896, + "tokens_seen": 542310400 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042204613841524577, + "loss": 1.5532, + "theoretical_loss": 3.909408564075374, + "tokens_seen": 542375936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220361083249749, + "loss": 1.5224, + "theoretical_loss": 3.9093587646245407, + "tokens_seen": 542441472 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042202607823470413, + "loss": 1.6452, + "theoretical_loss": 3.9093089728743684, + "tokens_seen": 542507008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220160481444333, + "loss": 1.6002, + "theoretical_loss": 3.9092591888227357, + "tokens_seen": 542572544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004220060180541625, + "loss": 1.5845, + "theoretical_loss": 3.909209412467523, + "tokens_seen": 542638080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004219959879638917, + "loss": 1.5996, + "theoretical_loss": 3.9091596438066114, + "tokens_seen": 542703616 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042198595787362085, + "loss": 1.5988, + "theoretical_loss": 3.909109882837882, + "tokens_seen": 542769152 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042197592778335004, + "loss": 1.5776, + "theoretical_loss": 3.9090601295592178, + "tokens_seen": 542834688 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042196589769307927, + "loss": 1.4319, + "theoretical_loss": 3.9090103839685013, + "tokens_seen": 542900224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004219558676028084, + "loss": 1.7546, + "theoretical_loss": 3.908960646063618, + "tokens_seen": 542965760 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042194583751253763, + "loss": 1.542, + "theoretical_loss": 3.908910915842452, + "tokens_seen": 543031296 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042193580742226676, + "loss": 1.4918, + "theoretical_loss": 3.90886119330289, + "tokens_seen": 543096832 + }, + { + "epoch": 0.16, + "learning_rate": 0.000421925777331996, + "loss": 1.4754, + "theoretical_loss": 3.908811478442818, + "tokens_seen": 543162368 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042191574724172523, + "loss": 1.5657, + "theoretical_loss": 3.908761771260124, + "tokens_seen": 543227904 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042190571715145436, + "loss": 1.4732, + "theoretical_loss": 3.908712071752696, + "tokens_seen": 543293440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218956870611836, + "loss": 1.5999, + "theoretical_loss": 3.908662379918424, + "tokens_seen": 543358976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218856569709127, + "loss": 1.583, + "theoretical_loss": 3.908612695755197, + "tokens_seen": 543424512 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042187562688064196, + "loss": 1.5797, + "theoretical_loss": 3.9085630192609067, + "tokens_seen": 543490048 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042186559679037114, + "loss": 1.5888, + "theoretical_loss": 3.908513350433444, + "tokens_seen": 543555584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218555667001003, + "loss": 1.5916, + "theoretical_loss": 3.908463689270703, + "tokens_seen": 543621120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218455366098295, + "loss": 1.5401, + "theoretical_loss": 3.9084140357705754, + "tokens_seen": 543686656 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042183550651955873, + "loss": 1.5957, + "theoretical_loss": 3.9083643899309566, + "tokens_seen": 543752192 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042182547642928786, + "loss": 1.6109, + "theoretical_loss": 3.9083147517497414, + "tokens_seen": 543817728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218154463390171, + "loss": 1.5178, + "theoretical_loss": 3.908265121224825, + "tokens_seen": 543883264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 204061, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6034040451049805, + "objective/train/theoretical_loss": 3.908215498354105, + "objective/train/tokens_used": 564408800, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004218054162487462, + "loss": 1.6154, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042179538615847546, + "loss": 1.5568, + "theoretical_loss": 3.9081658831354784, + "tokens_seen": 544014336 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042178535606820464, + "loss": 1.5658, + "theoretical_loss": 3.908116275566844, + "tokens_seen": 544079872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217753259779338, + "loss": 1.5529, + "theoretical_loss": 3.9080666756461007, + "tokens_seen": 544145408 + }, + { + "epoch": 0.16, + "learning_rate": 0.000421765295887663, + "loss": 1.6041, + "theoretical_loss": 3.908017083371149, + "tokens_seen": 544210944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217552657973922, + "loss": 1.5982, + "theoretical_loss": 3.9079674987398896, + "tokens_seen": 544276480 + }, + { + "epoch": 0.16, + "learning_rate": 0.00042174523570712136, + "loss": 1.6043, + "theoretical_loss": 3.907917921750224, + "tokens_seen": 544342016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217352056168506, + "loss": 1.3887, + "theoretical_loss": 3.9078683524000546, + "tokens_seen": 544407552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0004217251755265797, + "loss": 1.6197, + "theoretical_loss": 3.907818790687285, + "tokens_seen": 544473088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042171514543630896, + "loss": 1.5614, + "theoretical_loss": 3.90776923660982, + "tokens_seen": 544538624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004217051153460381, + "loss": 1.5018, + "theoretical_loss": 3.907719690165564, + "tokens_seen": 544604160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216950852557673, + "loss": 1.5123, + "theoretical_loss": 3.9076701513524226, + "tokens_seen": 544669696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216850551654965, + "loss": 1.5692, + "theoretical_loss": 3.9076206201683026, + "tokens_seen": 544735232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216750250752257, + "loss": 1.5249, + "theoretical_loss": 3.907571096611112, + "tokens_seen": 544800768 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042166499498495487, + "loss": 1.5547, + "theoretical_loss": 3.9075215806787584, + "tokens_seen": 544866304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216549648946841, + "loss": 1.5735, + "theoretical_loss": 3.907472072369152, + "tokens_seen": 544931840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042164493480441323, + "loss": 1.7223, + "theoretical_loss": 3.907422571680202, + "tokens_seen": 544997376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042163490471414246, + "loss": 1.5752, + "theoretical_loss": 3.907373078609819, + "tokens_seen": 545062912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216248746238716, + "loss": 1.6507, + "theoretical_loss": 3.9073235931559154, + "tokens_seen": 545128448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004216148445336008, + "loss": 1.5849, + "theoretical_loss": 3.907274115316403, + "tokens_seen": 545193984 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042160481444333, + "loss": 1.4948, + "theoretical_loss": 3.907224645089195, + "tokens_seen": 545259520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215947843530592, + "loss": 1.6905, + "theoretical_loss": 3.907175182472206, + "tokens_seen": 545325056 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042158475426278837, + "loss": 1.5254, + "theoretical_loss": 3.9071257274633506, + "tokens_seen": 545390592 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042157472417251755, + "loss": 1.6691, + "theoretical_loss": 3.9070762800605454, + "tokens_seen": 545456128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042156469408224673, + "loss": 1.6172, + "theoretical_loss": 3.9070268402617057, + "tokens_seen": 545521664 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042155466399197597, + "loss": 1.587, + "theoretical_loss": 3.9069774080647495, + "tokens_seen": 545587200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215446339017051, + "loss": 1.6625, + "theoretical_loss": 3.9069279834675945, + "tokens_seen": 545652736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042153460381143433, + "loss": 1.6148, + "theoretical_loss": 3.9068785664681602, + "tokens_seen": 545718272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215245737211635, + "loss": 1.5646, + "theoretical_loss": 3.906829157064367, + "tokens_seen": 545783808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215145436308927, + "loss": 1.5592, + "theoretical_loss": 3.906779755254134, + "tokens_seen": 545849344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004215045135406219, + "loss": 1.5757, + "theoretical_loss": 3.9067303610353847, + "tokens_seen": 545914880 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042149448345035105, + "loss": 1.6128, + "theoretical_loss": 3.9066809744060396, + "tokens_seen": 545980416 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042148445336008024, + "loss": 1.5317, + "theoretical_loss": 3.9066315953640225, + "tokens_seen": 546045952 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042147442326980947, + "loss": 1.6256, + "theoretical_loss": 3.9065822239072574, + "tokens_seen": 546111488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214643931795386, + "loss": 1.612, + "theoretical_loss": 3.9065328600336695, + "tokens_seen": 546177024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042145436308926783, + "loss": 1.5415, + "theoretical_loss": 3.9064835037411836, + "tokens_seen": 546242560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042144433299899696, + "loss": 1.5566, + "theoretical_loss": 3.9064341550277266, + "tokens_seen": 546308096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214343029087262, + "loss": 1.6336, + "theoretical_loss": 3.9063848138912247, + "tokens_seen": 546373632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004214242728184554, + "loss": 1.5429, + "theoretical_loss": 3.9063354803296075, + "tokens_seen": 546439168 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042141424272818456, + "loss": 1.5552, + "theoretical_loss": 3.906286154340802, + "tokens_seen": 546504704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042140421263791374, + "loss": 1.6011, + "theoretical_loss": 3.9062368359227397, + "tokens_seen": 546570240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213941825476429, + "loss": 1.5511, + "theoretical_loss": 3.90618752507335, + "tokens_seen": 546635776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213841524573721, + "loss": 1.6081, + "theoretical_loss": 3.906138221790565, + "tokens_seen": 546701312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042137412236710134, + "loss": 1.5121, + "theoretical_loss": 3.9060889260723153, + "tokens_seen": 546766848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042136409227683046, + "loss": 1.6071, + "theoretical_loss": 3.906039637916535, + "tokens_seen": 546832384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213540621865597, + "loss": 1.6133, + "theoretical_loss": 3.9059903573211576, + "tokens_seen": 546897920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213440320962889, + "loss": 1.5782, + "theoretical_loss": 3.905941084284117, + "tokens_seen": 546963456 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042133400200601806, + "loss": 1.3706, + "theoretical_loss": 3.9058918188033496, + "tokens_seen": 547028992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042132397191574724, + "loss": 1.6261, + "theoretical_loss": 3.905842560876791, + "tokens_seen": 547094528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213139418254764, + "loss": 1.5954, + "theoretical_loss": 3.9057933105023777, + "tokens_seen": 547160064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 204823, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4537506103515625, + "objective/train/theoretical_loss": 3.9057440676780484, + "objective/train/tokens_used": 567685600, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004213039117352056, + "loss": 1.5161, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042129388164493484, + "loss": 1.5423, + "theoretical_loss": 3.905694832401741, + "tokens_seen": 547291136 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042128385155466397, + "loss": 1.6366, + "theoretical_loss": 3.905645604671395, + "tokens_seen": 547356672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212738214643932, + "loss": 1.5296, + "theoretical_loss": 3.9055963844849506, + "tokens_seen": 547422208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042126379137412233, + "loss": 1.4606, + "theoretical_loss": 3.9055471718403494, + "tokens_seen": 547487744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042125376128385156, + "loss": 1.5466, + "theoretical_loss": 3.905497966735532, + "tokens_seen": 547553280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042124373119358075, + "loss": 1.5193, + "theoretical_loss": 3.9054487691684425, + "tokens_seen": 547618816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212337011033099, + "loss": 1.5192, + "theoretical_loss": 3.905399579137023, + "tokens_seen": 547684352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212236710130391, + "loss": 1.5437, + "theoretical_loss": 3.9053503966392187, + "tokens_seen": 547749888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004212136409227683, + "loss": 1.5519, + "theoretical_loss": 3.905301221672974, + "tokens_seen": 547815424 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042120361083249747, + "loss": 1.5382, + "theoretical_loss": 3.9052520542362346, + "tokens_seen": 547880960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211935807422267, + "loss": 1.6296, + "theoretical_loss": 3.9052028943269477, + "tokens_seen": 547946496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042118355065195583, + "loss": 1.4933, + "theoretical_loss": 3.905153741943061, + "tokens_seen": 548012032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042117352056168507, + "loss": 1.5737, + "theoretical_loss": 3.905104597082522, + "tokens_seen": 548077568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211634904714143, + "loss": 1.5936, + "theoretical_loss": 3.9050554597432807, + "tokens_seen": 548143104 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042115346038114343, + "loss": 1.6002, + "theoretical_loss": 3.905006329923286, + "tokens_seen": 548208640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042114343029087266, + "loss": 1.6265, + "theoretical_loss": 3.904957207620489, + "tokens_seen": 548274176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211334002006018, + "loss": 1.5553, + "theoretical_loss": 3.904908092832841, + "tokens_seen": 548339712 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042112337011033103, + "loss": 1.5162, + "theoretical_loss": 3.904858985558295, + "tokens_seen": 548405248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211133400200602, + "loss": 1.6292, + "theoretical_loss": 3.904809885794803, + "tokens_seen": 548470784 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004211033099297894, + "loss": 1.4599, + "theoretical_loss": 3.9047607935403197, + "tokens_seen": 548536320 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042109327983951857, + "loss": 1.4847, + "theoretical_loss": 3.9047117087927994, + "tokens_seen": 548601856 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042108324974924775, + "loss": 1.4091, + "theoretical_loss": 3.904662631550198, + "tokens_seen": 548667392 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042107321965897693, + "loss": 1.4015, + "theoretical_loss": 3.904613561810471, + "tokens_seen": 548732928 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042106318956870617, + "loss": 1.5605, + "theoretical_loss": 3.9045644995715767, + "tokens_seen": 548798464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210531594784353, + "loss": 1.6547, + "theoretical_loss": 3.9045154448314716, + "tokens_seen": 548864000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042104312938816453, + "loss": 1.5923, + "theoretical_loss": 3.9044663975881155, + "tokens_seen": 548929536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210330992978937, + "loss": 1.5552, + "theoretical_loss": 3.9044173578394674, + "tokens_seen": 548995072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210230692076229, + "loss": 1.5536, + "theoretical_loss": 3.904368325583488, + "tokens_seen": 549060608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004210130391173521, + "loss": 1.6015, + "theoretical_loss": 3.9043193008181376, + "tokens_seen": 549126144 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042100300902708125, + "loss": 1.5328, + "theoretical_loss": 3.9042702835413787, + "tokens_seen": 549191680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042099297893681044, + "loss": 1.5844, + "theoretical_loss": 3.9042212737511743, + "tokens_seen": 549257216 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042098294884653967, + "loss": 1.5848, + "theoretical_loss": 3.904172271445487, + "tokens_seen": 549322752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209729187562688, + "loss": 1.5692, + "theoretical_loss": 3.9041232766222813, + "tokens_seen": 549388288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042096288866599803, + "loss": 1.5404, + "theoretical_loss": 3.9040742892795226, + "tokens_seen": 549453824 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042095285857572716, + "loss": 1.5996, + "theoretical_loss": 3.9040253094151773, + "tokens_seen": 549519360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209428284854564, + "loss": 1.6798, + "theoretical_loss": 3.9039763370272107, + "tokens_seen": 549584896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209327983951856, + "loss": 1.4342, + "theoretical_loss": 3.903927372113592, + "tokens_seen": 549650432 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042092276830491476, + "loss": 1.5649, + "theoretical_loss": 3.9038784146722874, + "tokens_seen": 549715968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042091273821464394, + "loss": 1.6139, + "theoretical_loss": 3.9038294647012672, + "tokens_seen": 549781504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004209027081243731, + "loss": 1.5479, + "theoretical_loss": 3.9037805221985016, + "tokens_seen": 549847040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208926780341023, + "loss": 1.491, + "theoretical_loss": 3.90373158716196, + "tokens_seen": 549912576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042088264794383154, + "loss": 1.61, + "theoretical_loss": 3.903682659589615, + "tokens_seen": 549978112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042087261785356066, + "loss": 1.5066, + "theoretical_loss": 3.903633739479438, + "tokens_seen": 550043648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208625877632899, + "loss": 1.5242, + "theoretical_loss": 3.9035848268294027, + "tokens_seen": 550109184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208525576730191, + "loss": 1.4642, + "theoretical_loss": 3.9035359216374825, + "tokens_seen": 550174720 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042084252758274826, + "loss": 1.6741, + "theoretical_loss": 3.9034870239016524, + "tokens_seen": 550240256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042083249749247744, + "loss": 1.5825, + "theoretical_loss": 3.903438133619887, + "tokens_seen": 550305792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208224674022066, + "loss": 1.3974, + "theoretical_loss": 3.9033892507901635, + "tokens_seen": 550371328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004208124373119358, + "loss": 1.5958, + "theoretical_loss": 3.9033403754104583, + "tokens_seen": 550436864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 206347, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.65757155418396, + "objective/train/theoretical_loss": 3.9032915074787486, + "objective/train/tokens_used": 570962400, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042080240722166504, + "loss": 1.5843, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042079237713139417, + "loss": 1.5921, + "theoretical_loss": 3.903242646993014, + "tokens_seen": 550567936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207823470411234, + "loss": 1.5729, + "theoretical_loss": 3.9031937939512336, + "tokens_seen": 550633472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042077231695085253, + "loss": 1.525, + "theoretical_loss": 3.903144948351387, + "tokens_seen": 550699008 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042076228686058176, + "loss": 1.4598, + "theoretical_loss": 3.9030961101914556, + "tokens_seen": 550764544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042075225677031095, + "loss": 1.6102, + "theoretical_loss": 3.903047279469421, + "tokens_seen": 550830080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207422266800401, + "loss": 1.5922, + "theoretical_loss": 3.9029984561832656, + "tokens_seen": 550895616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207321965897693, + "loss": 1.5406, + "theoretical_loss": 3.9029496403309727, + "tokens_seen": 550961152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207221664994985, + "loss": 1.7136, + "theoretical_loss": 3.902900831910527, + "tokens_seen": 551026688 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042071213640922767, + "loss": 1.5988, + "theoretical_loss": 3.902852030919912, + "tokens_seen": 551092224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004207021063189569, + "loss": 1.5425, + "theoretical_loss": 3.902803237357115, + "tokens_seen": 551157760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042069207622868603, + "loss": 1.5829, + "theoretical_loss": 3.9027544512201207, + "tokens_seen": 551223296 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042068204613841527, + "loss": 1.5685, + "theoretical_loss": 3.902705672506918, + "tokens_seen": 551288832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042067201604814445, + "loss": 1.561, + "theoretical_loss": 3.9026569012154932, + "tokens_seen": 551354368 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042066198595787363, + "loss": 1.5456, + "theoretical_loss": 3.9026081373438366, + "tokens_seen": 551419904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206519558676028, + "loss": 1.5179, + "theoretical_loss": 3.9025593808899366, + "tokens_seen": 551485440 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420641925777332, + "loss": 1.6687, + "theoretical_loss": 3.9025106318517846, + "tokens_seen": 551550976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206318956870612, + "loss": 1.4832, + "theoretical_loss": 3.902461890227371, + "tokens_seen": 551616512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004206218655967904, + "loss": 1.4636, + "theoretical_loss": 3.902413156014688, + "tokens_seen": 551682048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042061183550651954, + "loss": 1.5451, + "theoretical_loss": 3.902364429211728, + "tokens_seen": 551747584 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042060180541624877, + "loss": 1.5667, + "theoretical_loss": 3.9023157098164853, + "tokens_seen": 551813120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205917753259779, + "loss": 1.5995, + "theoretical_loss": 3.902266997826953, + "tokens_seen": 551878656 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042058174523570713, + "loss": 1.4608, + "theoretical_loss": 3.902218293241127, + "tokens_seen": 551944192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205717151454363, + "loss": 1.4932, + "theoretical_loss": 3.9021695960570026, + "tokens_seen": 552009728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205616850551655, + "loss": 1.5708, + "theoretical_loss": 3.902120906272577, + "tokens_seen": 552075264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205516549648947, + "loss": 1.5305, + "theoretical_loss": 3.9020722238858467, + "tokens_seen": 552140800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205416248746239, + "loss": 1.5208, + "theoretical_loss": 3.902023548894811, + "tokens_seen": 552206336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042053159478435304, + "loss": 1.6749, + "theoretical_loss": 3.901974881297468, + "tokens_seen": 552271872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205215646940823, + "loss": 1.5422, + "theoretical_loss": 3.9019262210918173, + "tokens_seen": 552337408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004205115346038114, + "loss": 1.4243, + "theoretical_loss": 3.90187756827586, + "tokens_seen": 552402944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042050150451354064, + "loss": 1.5328, + "theoretical_loss": 3.9018289228475975, + "tokens_seen": 552468480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204914744232698, + "loss": 1.6587, + "theoretical_loss": 3.9017802848050307, + "tokens_seen": 552534016 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420481444332999, + "loss": 1.4564, + "theoretical_loss": 3.9017316541461637, + "tokens_seen": 552599552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204714142427282, + "loss": 1.6583, + "theoretical_loss": 3.9016830308689996, + "tokens_seen": 552665088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042046138415245736, + "loss": 1.5834, + "theoretical_loss": 3.901634414971543, + "tokens_seen": 552730624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042045135406218654, + "loss": 1.5044, + "theoretical_loss": 3.901585806451799, + "tokens_seen": 552796160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204413239719158, + "loss": 1.4579, + "theoretical_loss": 3.901537205307773, + "tokens_seen": 552861696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204312938816449, + "loss": 1.3961, + "theoretical_loss": 3.9014886115374723, + "tokens_seen": 552927232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042042126379137414, + "loss": 1.6138, + "theoretical_loss": 3.9014400251389043, + "tokens_seen": 552992768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204112337011033, + "loss": 1.5408, + "theoretical_loss": 3.901391446110077, + "tokens_seen": 553058304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004204012036108325, + "loss": 1.6263, + "theoretical_loss": 3.901342874449, + "tokens_seen": 553123840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042039117352056174, + "loss": 1.4783, + "theoretical_loss": 3.9012943101536823, + "tokens_seen": 553189376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042038114343029086, + "loss": 1.5495, + "theoretical_loss": 3.901245753222135, + "tokens_seen": 553254912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203711133400201, + "loss": 1.5825, + "theoretical_loss": 3.90119720365237, + "tokens_seen": 553320448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203610832497493, + "loss": 1.5265, + "theoretical_loss": 3.9011486614423982, + "tokens_seen": 553385984 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042035105315947846, + "loss": 1.5798, + "theoretical_loss": 3.901100126590233, + "tokens_seen": 553451520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042034102306920764, + "loss": 1.4984, + "theoretical_loss": 3.901051599093889, + "tokens_seen": 553517056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004203309929789368, + "loss": 1.6311, + "theoretical_loss": 3.901003078951379, + "tokens_seen": 553582592 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420320962888666, + "loss": 1.511, + "theoretical_loss": 3.9009545661607192, + "tokens_seen": 553648128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042031093279839524, + "loss": 1.6599, + "theoretical_loss": 3.900906060719925, + "tokens_seen": 553713664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 207589, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7764500379562378, + "objective/train/theoretical_loss": 3.9008575626270146, + "objective/train/tokens_used": 574239200, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042030090270812437, + "loss": 1.6069, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202908726178536, + "loss": 1.511, + "theoretical_loss": 3.900809071880004, + "tokens_seen": 553844736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042028084252758273, + "loss": 1.5882, + "theoretical_loss": 3.900760588476912, + "tokens_seen": 553910272 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042027081243731196, + "loss": 1.6452, + "theoretical_loss": 3.900712112415758, + "tokens_seen": 553975808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042026078234704115, + "loss": 1.5647, + "theoretical_loss": 3.900663643694561, + "tokens_seen": 554041344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202507522567703, + "loss": 1.539, + "theoretical_loss": 3.900615182311342, + "tokens_seen": 554106880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202407221664995, + "loss": 1.5483, + "theoretical_loss": 3.900566728264123, + "tokens_seen": 554172416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202306920762287, + "loss": 1.5102, + "theoretical_loss": 3.9005182815509256, + "tokens_seen": 554237952 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042022066198595787, + "loss": 1.4975, + "theoretical_loss": 3.9004698421697723, + "tokens_seen": 554303488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004202106318956871, + "loss": 1.6949, + "theoretical_loss": 3.900421410118688, + "tokens_seen": 554369024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042020060180541623, + "loss": 1.4607, + "theoretical_loss": 3.9003729853956957, + "tokens_seen": 554434560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042019057171514547, + "loss": 1.6326, + "theoretical_loss": 3.9003245679988217, + "tokens_seen": 554500096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042018054162487465, + "loss": 1.5879, + "theoretical_loss": 3.9002761579260907, + "tokens_seen": 554565632 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042017051153460383, + "loss": 1.5086, + "theoretical_loss": 3.9002277551755307, + "tokens_seen": 554631168 + }, + { + "epoch": 0.17, + "learning_rate": 0.000420160481444333, + "loss": 1.5341, + "theoretical_loss": 3.9001793597451684, + "tokens_seen": 554696704 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201504513540622, + "loss": 1.5667, + "theoretical_loss": 3.900130971633033, + "tokens_seen": 554762240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201404212637914, + "loss": 1.5292, + "theoretical_loss": 3.9000825908371524, + "tokens_seen": 554827776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201303911735206, + "loss": 1.4922, + "theoretical_loss": 3.9000342173555573, + "tokens_seen": 554893312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042012036108324974, + "loss": 1.6087, + "theoretical_loss": 3.899985851186277, + "tokens_seen": 554958848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042011033099297897, + "loss": 1.4813, + "theoretical_loss": 3.8999374923273447, + "tokens_seen": 555024384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004201003009027081, + "loss": 1.6093, + "theoretical_loss": 3.899889140776791, + "tokens_seen": 555089920 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042009027081243733, + "loss": 1.5267, + "theoretical_loss": 3.8998407965326494, + "tokens_seen": 555155456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200802407221665, + "loss": 1.4425, + "theoretical_loss": 3.899792459592953, + "tokens_seen": 555220992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200702106318957, + "loss": 1.4562, + "theoretical_loss": 3.899744129955737, + "tokens_seen": 555286528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200601805416249, + "loss": 1.5017, + "theoretical_loss": 3.899695807619035, + "tokens_seen": 555352064 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200501504513541, + "loss": 1.5836, + "theoretical_loss": 3.899647492580885, + "tokens_seen": 555417600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042004012036108324, + "loss": 1.6053, + "theoretical_loss": 3.899599184839322, + "tokens_seen": 555483136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200300902708125, + "loss": 1.4708, + "theoretical_loss": 3.899550884392384, + "tokens_seen": 555548672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004200200601805416, + "loss": 1.4251, + "theoretical_loss": 3.8995025912381087, + "tokens_seen": 555614208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042001003009027084, + "loss": 1.5171, + "theoretical_loss": 3.899454305374536, + "tokens_seen": 555679744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00042, + "loss": 1.5027, + "theoretical_loss": 3.899406026799704, + "tokens_seen": 555745280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199899699097292, + "loss": 1.5009, + "theoretical_loss": 3.899357755511655, + "tokens_seen": 555810816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199799398194584, + "loss": 1.6629, + "theoretical_loss": 3.899309491508429, + "tokens_seen": 555876352 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041996990972918756, + "loss": 1.6253, + "theoretical_loss": 3.899261234788068, + "tokens_seen": 555941888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041995987963891674, + "loss": 1.3213, + "theoretical_loss": 3.8992129853486155, + "tokens_seen": 556007424 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419949849548646, + "loss": 1.5332, + "theoretical_loss": 3.8991647431881145, + "tokens_seen": 556072960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199398194583751, + "loss": 1.6573, + "theoretical_loss": 3.899116508304609, + "tokens_seen": 556138496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041992978936810434, + "loss": 1.6323, + "theoretical_loss": 3.899068280696144, + "tokens_seen": 556204032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041991975927783347, + "loss": 1.4243, + "theoretical_loss": 3.8990200603607654, + "tokens_seen": 556269568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004199097291875627, + "loss": 1.6023, + "theoretical_loss": 3.8989718472965196, + "tokens_seen": 556335104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198996990972919, + "loss": 1.5731, + "theoretical_loss": 3.898923641501454, + "tokens_seen": 556400640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041988966900702106, + "loss": 1.5216, + "theoretical_loss": 3.898875442973617, + "tokens_seen": 556466176 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041987963891675025, + "loss": 1.4962, + "theoretical_loss": 3.8988272517110563, + "tokens_seen": 556531712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198696088264795, + "loss": 1.4863, + "theoretical_loss": 3.898779067711822, + "tokens_seen": 556597248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198595787362086, + "loss": 1.5111, + "theoretical_loss": 3.8987308909739644, + "tokens_seen": 556662784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041984954864593784, + "loss": 1.6661, + "theoretical_loss": 3.898682721495535, + "tokens_seen": 556728320 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041983951855566697, + "loss": 1.5631, + "theoretical_loss": 3.8986345592745844, + "tokens_seen": 556793856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198294884653962, + "loss": 1.5465, + "theoretical_loss": 3.898586404309166, + "tokens_seen": 556859392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004198194583751254, + "loss": 1.4707, + "theoretical_loss": 3.8985382565973326, + "tokens_seen": 556924928 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041980942828485457, + "loss": 1.4855, + "theoretical_loss": 3.898490116137139, + "tokens_seen": 556990464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 208740, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5904827117919922, + "objective/train/theoretical_loss": 3.898441982926639, + "objective/train/tokens_used": 577516000, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041979939819458375, + "loss": 1.4995, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041978936810431293, + "loss": 1.4357, + "theoretical_loss": 3.8983938569638883, + "tokens_seen": 557121536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004197793380140421, + "loss": 1.5928, + "theoretical_loss": 3.8983457382469435, + "tokens_seen": 557187072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041976930792377135, + "loss": 1.4884, + "theoretical_loss": 3.8982976267738616, + "tokens_seen": 557252608 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041975927783350047, + "loss": 1.4285, + "theoretical_loss": 3.8982495225427005, + "tokens_seen": 557318144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004197492477432297, + "loss": 1.4857, + "theoretical_loss": 3.8982014255515183, + "tokens_seen": 557383680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041973921765295883, + "loss": 1.5486, + "theoretical_loss": 3.8981533357983746, + "tokens_seen": 557449216 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041972918756268807, + "loss": 1.5234, + "theoretical_loss": 3.8981052532813294, + "tokens_seen": 557514752 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041971915747241725, + "loss": 1.514, + "theoretical_loss": 3.898057177998443, + "tokens_seen": 557580288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041970912738214643, + "loss": 1.476, + "theoretical_loss": 3.898009109947777, + "tokens_seen": 557645824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196990972918756, + "loss": 1.4411, + "theoretical_loss": 3.897961049127395, + "tokens_seen": 557711360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041968906720160485, + "loss": 1.5008, + "theoretical_loss": 3.8979129955353584, + "tokens_seen": 557776896 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419679037111334, + "loss": 1.5109, + "theoretical_loss": 3.8978649491697315, + "tokens_seen": 557842432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196690070210632, + "loss": 1.4921, + "theoretical_loss": 3.8978169100285784, + "tokens_seen": 557907968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196589769307924, + "loss": 1.472, + "theoretical_loss": 3.8977688781099653, + "tokens_seen": 557973504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196489468405216, + "loss": 1.5021, + "theoretical_loss": 3.8977208534119576, + "tokens_seen": 558039040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196389167502508, + "loss": 1.6118, + "theoretical_loss": 3.8976728359326223, + "tokens_seen": 558104576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041962888665997994, + "loss": 1.5115, + "theoretical_loss": 3.897624825670026, + "tokens_seen": 558170112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041961885656970917, + "loss": 1.6053, + "theoretical_loss": 3.8975768226222383, + "tokens_seen": 558235648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004196088264794383, + "loss": 1.5389, + "theoretical_loss": 3.897528826787327, + "tokens_seen": 558301184 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041959879638916753, + "loss": 1.5368, + "theoretical_loss": 3.8974808381633625, + "tokens_seen": 558366720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195887662988967, + "loss": 1.6761, + "theoretical_loss": 3.8974328567484147, + "tokens_seen": 558432256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195787362086259, + "loss": 1.5996, + "theoretical_loss": 3.8973848825405555, + "tokens_seen": 558497792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195687061183551, + "loss": 1.5337, + "theoretical_loss": 3.8973369155378563, + "tokens_seen": 558563328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195586760280843, + "loss": 1.5526, + "theoretical_loss": 3.89728895573839, + "tokens_seen": 558628864 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041954864593781344, + "loss": 1.5629, + "theoretical_loss": 3.8972410031402296, + "tokens_seen": 558694400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195386158475427, + "loss": 1.5878, + "theoretical_loss": 3.8971930577414495, + "tokens_seen": 558759936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195285857572718, + "loss": 1.5944, + "theoretical_loss": 3.8971451195401254, + "tokens_seen": 558825472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041951855566700104, + "loss": 1.5638, + "theoretical_loss": 3.897097188534332, + "tokens_seen": 558891008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004195085255767302, + "loss": 1.5937, + "theoretical_loss": 3.8970492647221464, + "tokens_seen": 558956544 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194984954864594, + "loss": 1.6083, + "theoretical_loss": 3.897001348101645, + "tokens_seen": 559022080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194884653961886, + "loss": 1.5013, + "theoretical_loss": 3.8969534386709053, + "tokens_seen": 559087616 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041947843530591776, + "loss": 1.6256, + "theoretical_loss": 3.896905536428007, + "tokens_seen": 559153152 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041946840521564694, + "loss": 1.6094, + "theoretical_loss": 3.8968576413710294, + "tokens_seen": 559218688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194583751253762, + "loss": 1.5515, + "theoretical_loss": 3.8968097534980517, + "tokens_seen": 559284224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194483450351053, + "loss": 1.4999, + "theoretical_loss": 3.8967618728071556, + "tokens_seen": 559349760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041943831494483454, + "loss": 1.5804, + "theoretical_loss": 3.896713999296422, + "tokens_seen": 559415296 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041942828485456367, + "loss": 1.5646, + "theoretical_loss": 3.8966661329639334, + "tokens_seen": 559480832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194182547642929, + "loss": 1.5926, + "theoretical_loss": 3.896618273807773, + "tokens_seen": 559546368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004194082246740221, + "loss": 1.5404, + "theoretical_loss": 3.8965704218260244, + "tokens_seen": 559611904 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041939819458375126, + "loss": 1.616, + "theoretical_loss": 3.896522577016772, + "tokens_seen": 559677440 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041938816449348045, + "loss": 1.5705, + "theoretical_loss": 3.896474739378102, + "tokens_seen": 559742976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193781344032097, + "loss": 1.5944, + "theoretical_loss": 3.896426908908099, + "tokens_seen": 559808512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193681043129388, + "loss": 1.534, + "theoretical_loss": 3.89637908560485, + "tokens_seen": 559874048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041935807422266804, + "loss": 1.5263, + "theoretical_loss": 3.896331269466443, + "tokens_seen": 559939584 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041934804413239717, + "loss": 1.47, + "theoretical_loss": 3.8962834604909657, + "tokens_seen": 560005120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193380140421264, + "loss": 1.4907, + "theoretical_loss": 3.8962356586765075, + "tokens_seen": 560070656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004193279839518556, + "loss": 1.4651, + "theoretical_loss": 3.896187864021157, + "tokens_seen": 560136192 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041931795386158477, + "loss": 1.58, + "theoretical_loss": 3.8961400765230065, + "tokens_seen": 560201728 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041930792377131395, + "loss": 1.5162, + "theoretical_loss": 3.8960922961801447, + "tokens_seen": 560267264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 209977, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4986798763275146, + "objective/train/theoretical_loss": 3.8960445229906657, + "objective/train/tokens_used": 580792800, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041929789368104313, + "loss": 1.4737, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192878635907723, + "loss": 1.5115, + "theoretical_loss": 3.895996756952661, + "tokens_seen": 560398336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041927783350050155, + "loss": 1.4989, + "theoretical_loss": 3.895948998064223, + "tokens_seen": 560463872 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041926780341023067, + "loss": 1.5245, + "theoretical_loss": 3.8959012463234473, + "tokens_seen": 560529408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192577733199599, + "loss": 1.3667, + "theoretical_loss": 3.8958535017284284, + "tokens_seen": 560594944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041924774322968904, + "loss": 1.4875, + "theoretical_loss": 3.895805764277261, + "tokens_seen": 560660480 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041923771313941827, + "loss": 1.5673, + "theoretical_loss": 3.895758033968042, + "tokens_seen": 560726016 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041922768304914745, + "loss": 1.4816, + "theoretical_loss": 3.895710310798868, + "tokens_seen": 560791552 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041921765295887663, + "loss": 1.5509, + "theoretical_loss": 3.895662594767837, + "tokens_seen": 560857088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004192076228686058, + "loss": 1.6063, + "theoretical_loss": 3.895614885873047, + "tokens_seen": 560922624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041919759277833505, + "loss": 1.5571, + "theoretical_loss": 3.895567184112598, + "tokens_seen": 560988160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191875626880642, + "loss": 1.6526, + "theoretical_loss": 3.8955194894845886, + "tokens_seen": 561053696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191775325977934, + "loss": 1.5105, + "theoretical_loss": 3.89547180198712, + "tokens_seen": 561119232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041916750250752254, + "loss": 1.4558, + "theoretical_loss": 3.8954241216182943, + "tokens_seen": 561184768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191574724172518, + "loss": 1.5593, + "theoretical_loss": 3.8953764483762123, + "tokens_seen": 561250304 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041914744232698095, + "loss": 1.4963, + "theoretical_loss": 3.8953287822589777, + "tokens_seen": 561315840 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041913741223671014, + "loss": 1.4298, + "theoretical_loss": 3.8952811232646933, + "tokens_seen": 561381376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191273821464393, + "loss": 1.5867, + "theoretical_loss": 3.895233471391464, + "tokens_seen": 561446912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191173520561685, + "loss": 1.4545, + "theoretical_loss": 3.895185826637394, + "tokens_seen": 561512448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004191073219658977, + "loss": 1.624, + "theoretical_loss": 3.89513818900059, + "tokens_seen": 561577984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190972918756269, + "loss": 1.5141, + "theoretical_loss": 3.8950905584791577, + "tokens_seen": 561643520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041908726178535604, + "loss": 1.5473, + "theoretical_loss": 3.8950429350712037, + "tokens_seen": 561709056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190772316950853, + "loss": 1.5132, + "theoretical_loss": 3.8949953187748374, + "tokens_seen": 561774592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190672016048144, + "loss": 1.5394, + "theoretical_loss": 3.894947709588166, + "tokens_seen": 561840128 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041905717151454364, + "loss": 1.422, + "theoretical_loss": 3.8949001075093, + "tokens_seen": 561905664 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190471414242728, + "loss": 1.5832, + "theoretical_loss": 3.8948525125363487, + "tokens_seen": 561971200 + }, + { + "epoch": 0.17, + "learning_rate": 0.000419037111334002, + "loss": 1.5745, + "theoretical_loss": 3.8948049246674223, + "tokens_seen": 562036736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190270812437312, + "loss": 1.5174, + "theoretical_loss": 3.894757343900633, + "tokens_seen": 562102272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004190170511534604, + "loss": 1.6155, + "theoretical_loss": 3.894709770234093, + "tokens_seen": 562167808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041900702106318954, + "loss": 1.5541, + "theoretical_loss": 3.8946622036659155, + "tokens_seen": 562233344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189969909729188, + "loss": 1.6732, + "theoretical_loss": 3.8946146441942138, + "tokens_seen": 562298880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189869608826479, + "loss": 1.5481, + "theoretical_loss": 3.894567091817102, + "tokens_seen": 562364416 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041897693079237714, + "loss": 1.4682, + "theoretical_loss": 3.8945195465326954, + "tokens_seen": 562429952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189669007021063, + "loss": 1.5198, + "theoretical_loss": 3.8944720083391093, + "tokens_seen": 562495488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189568706118355, + "loss": 1.5586, + "theoretical_loss": 3.8944244772344616, + "tokens_seen": 562561024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189468405215647, + "loss": 1.5045, + "theoretical_loss": 3.894376953216868, + "tokens_seen": 562626560 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041893681043129387, + "loss": 1.573, + "theoretical_loss": 3.8943294362844476, + "tokens_seen": 562692096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041892678034102305, + "loss": 1.5758, + "theoretical_loss": 3.8942819264353186, + "tokens_seen": 562757632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004189167502507523, + "loss": 1.5688, + "theoretical_loss": 3.8942344236676005, + "tokens_seen": 562823168 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041890672016048146, + "loss": 1.5607, + "theoretical_loss": 3.894186927979413, + "tokens_seen": 562888704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041889669007021065, + "loss": 1.6688, + "theoretical_loss": 3.8941394393688777, + "tokens_seen": 562954240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188866599799399, + "loss": 1.5392, + "theoretical_loss": 3.8940919578341155, + "tokens_seen": 563019776 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418876629889669, + "loss": 1.5315, + "theoretical_loss": 3.894044483373249, + "tokens_seen": 563085312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041886659979939824, + "loss": 1.6801, + "theoretical_loss": 3.893997015984401, + "tokens_seen": 563150848 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041885656970912737, + "loss": 1.5611, + "theoretical_loss": 3.893949555665696, + "tokens_seen": 563216384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188465396188566, + "loss": 1.4364, + "theoretical_loss": 3.893902102415257, + "tokens_seen": 563281920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004188365095285858, + "loss": 1.6739, + "theoretical_loss": 3.89385465623121, + "tokens_seen": 563347456 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041882647943831497, + "loss": 1.5617, + "theoretical_loss": 3.8938072171116804, + "tokens_seen": 563412992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041881644934804415, + "loss": 1.4844, + "theoretical_loss": 3.893759785054796, + "tokens_seen": 563478528 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041880641925777333, + "loss": 1.4888, + "theoretical_loss": 3.8937123600586823, + "tokens_seen": 563544064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 211161, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6598248481750488, + "objective/train/theoretical_loss": 3.8936649421214686, + "objective/train/tokens_used": 584069600, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187963891675025, + "loss": 1.5457, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041878635907723175, + "loss": 1.5758, + "theoretical_loss": 3.8936175312412833, + "tokens_seen": 563675136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187763289869609, + "loss": 1.5379, + "theoretical_loss": 3.8935701274162553, + "tokens_seen": 563740672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004187662988966901, + "loss": 1.4818, + "theoretical_loss": 3.8935227306445155, + "tokens_seen": 563806208 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041875626880641924, + "loss": 1.6127, + "theoretical_loss": 3.8934753409241942, + "tokens_seen": 563871744 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041874623871614847, + "loss": 1.6181, + "theoretical_loss": 3.8934279582534232, + "tokens_seen": 563937280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041873620862587765, + "loss": 1.5845, + "theoretical_loss": 3.8933805826303347, + "tokens_seen": 564002816 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041872617853560683, + "loss": 1.5021, + "theoretical_loss": 3.893333214053062, + "tokens_seen": 564068352 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418716148445336, + "loss": 1.6005, + "theoretical_loss": 3.8932858525197385, + "tokens_seen": 564133888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041870611835506525, + "loss": 1.5789, + "theoretical_loss": 3.8932384980284986, + "tokens_seen": 564199424 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186960882647944, + "loss": 1.47, + "theoretical_loss": 3.8931911505774766, + "tokens_seen": 564264960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186860581745236, + "loss": 1.5558, + "theoretical_loss": 3.89314381016481, + "tokens_seen": 564330496 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041867602808425274, + "loss": 1.549, + "theoretical_loss": 3.8930964767886347, + "tokens_seen": 564396032 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418665997993982, + "loss": 1.4684, + "theoretical_loss": 3.893049150447087, + "tokens_seen": 564461568 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041865596790371115, + "loss": 1.5208, + "theoretical_loss": 3.8930018311383066, + "tokens_seen": 564527104 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041864593781344034, + "loss": 1.4459, + "theoretical_loss": 3.8929545188604306, + "tokens_seen": 564592640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186359077231695, + "loss": 1.534, + "theoretical_loss": 3.8929072136115987, + "tokens_seen": 564658176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186258776328987, + "loss": 1.5551, + "theoretical_loss": 3.8928599153899515, + "tokens_seen": 564723712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186158475426279, + "loss": 1.4947, + "theoretical_loss": 3.8928126241936294, + "tokens_seen": 564789248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004186058174523571, + "loss": 1.4813, + "theoretical_loss": 3.8927653400207745, + "tokens_seen": 564854784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041859578736208624, + "loss": 1.4796, + "theoretical_loss": 3.892718062869528, + "tokens_seen": 564920320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185857572718155, + "loss": 1.5784, + "theoretical_loss": 3.8926707927380337, + "tokens_seen": 564985856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185757271815446, + "loss": 1.5971, + "theoretical_loss": 3.892623529624435, + "tokens_seen": 565051392 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041856569709127384, + "loss": 1.4777, + "theoretical_loss": 3.892576273526876, + "tokens_seen": 565116928 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418555667001003, + "loss": 1.5312, + "theoretical_loss": 3.892529024443502, + "tokens_seen": 565182464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185456369107322, + "loss": 1.6698, + "theoretical_loss": 3.892481782372458, + "tokens_seen": 565248000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185356068204614, + "loss": 1.4816, + "theoretical_loss": 3.8924345473118915, + "tokens_seen": 565313536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004185255767301906, + "loss": 1.5975, + "theoretical_loss": 3.892387319259949, + "tokens_seen": 565379072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041851554663991974, + "loss": 1.5666, + "theoretical_loss": 3.8923400982147784, + "tokens_seen": 565444608 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418505516549649, + "loss": 1.5587, + "theoretical_loss": 3.8922928841745286, + "tokens_seen": 565510144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184954864593781, + "loss": 1.6348, + "theoretical_loss": 3.892245677137349, + "tokens_seen": 565575680 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041848545636910734, + "loss": 1.5067, + "theoretical_loss": 3.8921984771013887, + "tokens_seen": 565641216 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184754262788365, + "loss": 1.5334, + "theoretical_loss": 3.8921512840647985, + "tokens_seen": 565706752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184653961885657, + "loss": 1.5075, + "theoretical_loss": 3.892104098025731, + "tokens_seen": 565772288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184553660982949, + "loss": 1.5249, + "theoretical_loss": 3.892056918982337, + "tokens_seen": 565837824 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041844533600802407, + "loss": 1.65, + "theoretical_loss": 3.8920097469327697, + "tokens_seen": 565903360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041843530591775325, + "loss": 1.4984, + "theoretical_loss": 3.8919625818751826, + "tokens_seen": 565968896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184252758274825, + "loss": 1.5221, + "theoretical_loss": 3.8919154238077294, + "tokens_seen": 566034432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004184152457372116, + "loss": 1.5036, + "theoretical_loss": 3.8918682727285665, + "tokens_seen": 566099968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041840521564694085, + "loss": 1.3004, + "theoretical_loss": 3.8918211286358475, + "tokens_seen": 566165504 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041839518555666997, + "loss": 1.5408, + "theoretical_loss": 3.89177399152773, + "tokens_seen": 566231040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183851554663992, + "loss": 1.5654, + "theoretical_loss": 3.89172686140237, + "tokens_seen": 566296576 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183751253761284, + "loss": 1.549, + "theoretical_loss": 3.8916797382579262, + "tokens_seen": 566362112 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041836509528585757, + "loss": 1.5328, + "theoretical_loss": 3.8916326220925566, + "tokens_seen": 566427648 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041835506519558675, + "loss": 1.4325, + "theoretical_loss": 3.8915855129044195, + "tokens_seen": 566493184 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418345035105316, + "loss": 1.52, + "theoretical_loss": 3.891538410691676, + "tokens_seen": 566558720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183350050150451, + "loss": 1.4941, + "theoretical_loss": 3.8914913154524857, + "tokens_seen": 566624256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041832497492477435, + "loss": 1.4724, + "theoretical_loss": 3.89144422718501, + "tokens_seen": 566689792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183149448345035, + "loss": 1.5358, + "theoretical_loss": 3.891397145887411, + "tokens_seen": 566755328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004183049147442327, + "loss": 1.5945, + "theoretical_loss": 3.8913500715578504, + "tokens_seen": 566820864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 212374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5385618209838867, + "objective/train/theoretical_loss": 3.8913030041944925, + "objective/train/tokens_used": 587346400, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182948846539619, + "loss": 1.5984, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182848545636911, + "loss": 1.5634, + "theoretical_loss": 3.891255943795501, + "tokens_seen": 566951936 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041827482447342025, + "loss": 1.5627, + "theoretical_loss": 3.8912088903590405, + "tokens_seen": 567017472 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041826479438314944, + "loss": 1.5036, + "theoretical_loss": 3.8911618438832756, + "tokens_seen": 567083008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182547642928786, + "loss": 1.5397, + "theoretical_loss": 3.8911148043663735, + "tokens_seen": 567148544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041824473420260785, + "loss": 1.4498, + "theoretical_loss": 3.8910677718065, + "tokens_seen": 567214080 + }, + { + "epoch": 0.17, + "learning_rate": 0.000418234704112337, + "loss": 1.6945, + "theoretical_loss": 3.8910207462018236, + "tokens_seen": 567279616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182246740220662, + "loss": 1.5721, + "theoretical_loss": 3.890973727550511, + "tokens_seen": 567345152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182146439317954, + "loss": 1.5251, + "theoretical_loss": 3.890926715850733, + "tokens_seen": 567410688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004182046138415246, + "loss": 1.7168, + "theoretical_loss": 3.890879711100657, + "tokens_seen": 567476224 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041819458375125376, + "loss": 1.5417, + "theoretical_loss": 3.8908327132984546, + "tokens_seen": 567541760 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041818455366098294, + "loss": 1.4895, + "theoretical_loss": 3.890785722442296, + "tokens_seen": 567607296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181745235707121, + "loss": 1.506, + "theoretical_loss": 3.890738738530353, + "tokens_seen": 567672832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041816449348044136, + "loss": 1.5614, + "theoretical_loss": 3.890691761560798, + "tokens_seen": 567738368 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041815446339017054, + "loss": 1.5725, + "theoretical_loss": 3.8906447915318036, + "tokens_seen": 567803904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181444332998997, + "loss": 1.3962, + "theoretical_loss": 3.8905978284415443, + "tokens_seen": 567869440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181344032096289, + "loss": 1.6267, + "theoretical_loss": 3.890550872288194, + "tokens_seen": 567934976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181243731193581, + "loss": 1.5362, + "theoretical_loss": 3.890503923069927, + "tokens_seen": 568000512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004181143430290873, + "loss": 1.6173, + "theoretical_loss": 3.8904569807849203, + "tokens_seen": 568066048 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041810431293881644, + "loss": 1.5599, + "theoretical_loss": 3.8904100454313495, + "tokens_seen": 568131584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180942828485457, + "loss": 1.6167, + "theoretical_loss": 3.890363117007392, + "tokens_seen": 568197120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180842527582748, + "loss": 1.5266, + "theoretical_loss": 3.890316195511226, + "tokens_seen": 568262656 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041807422266800404, + "loss": 1.5455, + "theoretical_loss": 3.8902692809410295, + "tokens_seen": 568328192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180641925777332, + "loss": 1.5529, + "theoretical_loss": 3.8902223732949817, + "tokens_seen": 568393728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180541624874624, + "loss": 1.6265, + "theoretical_loss": 3.8901754725712627, + "tokens_seen": 568459264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180441323971916, + "loss": 1.5508, + "theoretical_loss": 3.890128578768053, + "tokens_seen": 568524800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180341023069208, + "loss": 1.5442, + "theoretical_loss": 3.8900816918835335, + "tokens_seen": 568590336 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041802407221664994, + "loss": 1.5351, + "theoretical_loss": 3.8900348119158865, + "tokens_seen": 568655872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180140421263792, + "loss": 1.6435, + "theoretical_loss": 3.8899879388632947, + "tokens_seen": 568721408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004180040120361083, + "loss": 1.6069, + "theoretical_loss": 3.889941072723942, + "tokens_seen": 568786944 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041799398194583754, + "loss": 1.5581, + "theoretical_loss": 3.8898942134960106, + "tokens_seen": 568852480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179839518555667, + "loss": 1.5831, + "theoretical_loss": 3.8898473611776865, + "tokens_seen": 568918016 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179739217652959, + "loss": 1.4758, + "theoretical_loss": 3.8898005157671554, + "tokens_seen": 568983552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179638916750251, + "loss": 1.4469, + "theoretical_loss": 3.889753677262603, + "tokens_seen": 569049088 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041795386158475427, + "loss": 1.6384, + "theoretical_loss": 3.8897068456622153, + "tokens_seen": 569114624 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041794383149448345, + "loss": 1.5027, + "theoretical_loss": 3.8896600209641807, + "tokens_seen": 569180160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179338014042127, + "loss": 1.5995, + "theoretical_loss": 3.889613203166687, + "tokens_seen": 569245696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004179237713139418, + "loss": 1.6531, + "theoretical_loss": 3.8895663922679224, + "tokens_seen": 569311232 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041791374122367105, + "loss": 1.5091, + "theoretical_loss": 3.8895195882660776, + "tokens_seen": 569376768 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041790371113340017, + "loss": 1.551, + "theoretical_loss": 3.8894727911593416, + "tokens_seen": 569442304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178936810431294, + "loss": 1.687, + "theoretical_loss": 3.889426000945906, + "tokens_seen": 569507840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178836509528586, + "loss": 1.5233, + "theoretical_loss": 3.889379217623962, + "tokens_seen": 569573376 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041787362086258777, + "loss": 1.452, + "theoretical_loss": 3.8893324411917014, + "tokens_seen": 569638912 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041786359077231695, + "loss": 1.6002, + "theoretical_loss": 3.8892856716473183, + "tokens_seen": 569704448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178535606820462, + "loss": 1.6722, + "theoretical_loss": 3.8892389089890056, + "tokens_seen": 569769984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178435305917753, + "loss": 1.5446, + "theoretical_loss": 3.8891921532149576, + "tokens_seen": 569835520 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041783350050150455, + "loss": 1.5458, + "theoretical_loss": 3.8891454043233686, + "tokens_seen": 569901056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178234704112337, + "loss": 1.5174, + "theoretical_loss": 3.8890986623124353, + "tokens_seen": 569966592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178134403209629, + "loss": 1.6917, + "theoretical_loss": 3.889051927180353, + "tokens_seen": 570032128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004178034102306921, + "loss": 1.4804, + "theoretical_loss": 3.88900519892532, + "tokens_seen": 570097664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 213119, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4728543758392334, + "objective/train/theoretical_loss": 3.8889584775455326, + "objective/train/tokens_used": 590623200, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177933801404213, + "loss": 1.5645, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041778335005015045, + "loss": 1.5078, + "theoretical_loss": 3.8889117630391894, + "tokens_seen": 570228736 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041777331995987964, + "loss": 1.4843, + "theoretical_loss": 3.888865055404491, + "tokens_seen": 570294272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177632898696088, + "loss": 1.4644, + "theoretical_loss": 3.8888183546396347, + "tokens_seen": 570359808 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041775325977933805, + "loss": 1.5443, + "theoretical_loss": 3.8887716607428224, + "tokens_seen": 570425344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177432296890672, + "loss": 1.464, + "theoretical_loss": 3.8887249737122547, + "tokens_seen": 570490880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177331995987964, + "loss": 1.4662, + "theoretical_loss": 3.8886782935461333, + "tokens_seen": 570556416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177231695085256, + "loss": 1.6238, + "theoretical_loss": 3.888631620242661, + "tokens_seen": 570621952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004177131394182548, + "loss": 1.5825, + "theoretical_loss": 3.8885849538000397, + "tokens_seen": 570687488 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041770310932798396, + "loss": 1.5631, + "theoretical_loss": 3.8885382942164752, + "tokens_seen": 570753024 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041769307923771314, + "loss": 1.4684, + "theoretical_loss": 3.8884916414901705, + "tokens_seen": 570818560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176830491474423, + "loss": 1.5735, + "theoretical_loss": 3.8884449956193308, + "tokens_seen": 570884096 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041767301905717156, + "loss": 1.703, + "theoretical_loss": 3.8883983566021625, + "tokens_seen": 570949632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176629889669007, + "loss": 1.442, + "theoretical_loss": 3.8883517244368715, + "tokens_seen": 571015168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176529588766299, + "loss": 1.5897, + "theoretical_loss": 3.8883050991216654, + "tokens_seen": 571080704 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041764292878635904, + "loss": 1.5325, + "theoretical_loss": 3.888258480654751, + "tokens_seen": 571146240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176328986960883, + "loss": 1.4552, + "theoretical_loss": 3.888211869034339, + "tokens_seen": 571211776 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041762286860581746, + "loss": 1.5991, + "theoretical_loss": 3.888165264258636, + "tokens_seen": 571277312 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041761283851554664, + "loss": 1.631, + "theoretical_loss": 3.888118666325853, + "tokens_seen": 571342848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004176028084252758, + "loss": 1.5604, + "theoretical_loss": 3.888072075234201, + "tokens_seen": 571408384 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417592778335005, + "loss": 1.4775, + "theoretical_loss": 3.8880254909818905, + "tokens_seen": 571473920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175827482447342, + "loss": 1.7107, + "theoretical_loss": 3.8879789135671334, + "tokens_seen": 571539456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175727181544634, + "loss": 1.519, + "theoretical_loss": 3.8879323429881425, + "tokens_seen": 571604992 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041756268806419255, + "loss": 1.5448, + "theoretical_loss": 3.887885779243131, + "tokens_seen": 571670528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175526579739218, + "loss": 1.5976, + "theoretical_loss": 3.887839222330313, + "tokens_seen": 571736064 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041754262788365096, + "loss": 1.6828, + "theoretical_loss": 3.887792672247902, + "tokens_seen": 571801600 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041753259779338015, + "loss": 1.5068, + "theoretical_loss": 3.887746128994114, + "tokens_seen": 571867136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175225677031093, + "loss": 1.6907, + "theoretical_loss": 3.8876995925671656, + "tokens_seen": 571932672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175125376128385, + "loss": 1.4601, + "theoretical_loss": 3.8876530629652724, + "tokens_seen": 571998208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004175025075225677, + "loss": 1.6105, + "theoretical_loss": 3.8876065401866513, + "tokens_seen": 572063744 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174924774322969, + "loss": 1.5684, + "theoretical_loss": 3.887560024229521, + "tokens_seen": 572129280 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041748244734202605, + "loss": 1.5362, + "theoretical_loss": 3.8875135150921, + "tokens_seen": 572194816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174724172517553, + "loss": 1.6183, + "theoretical_loss": 3.887467012772607, + "tokens_seen": 572260352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174623871614844, + "loss": 1.6805, + "theoretical_loss": 3.8874205172692626, + "tokens_seen": 572325888 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041745235707121365, + "loss": 1.5239, + "theoretical_loss": 3.887374028580287, + "tokens_seen": 572391424 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041744232698094283, + "loss": 1.6844, + "theoretical_loss": 3.8873275467039017, + "tokens_seen": 572456960 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417432296890672, + "loss": 1.5191, + "theoretical_loss": 3.887281071638328, + "tokens_seen": 572522496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174222668004012, + "loss": 1.5156, + "theoretical_loss": 3.8872346033817893, + "tokens_seen": 572588032 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041741223671013037, + "loss": 1.6615, + "theoretical_loss": 3.887188141932508, + "tokens_seen": 572653568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004174022066198596, + "loss": 1.5984, + "theoretical_loss": 3.8871416872887083, + "tokens_seen": 572719104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173921765295888, + "loss": 1.5672, + "theoretical_loss": 3.8870952394486156, + "tokens_seen": 572784640 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041738214643931797, + "loss": 1.4527, + "theoretical_loss": 3.8870487984104543, + "tokens_seen": 572850176 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041737211634904715, + "loss": 1.3989, + "theoretical_loss": 3.8870023641724503, + "tokens_seen": 572915712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173620862587764, + "loss": 1.6239, + "theoretical_loss": 3.886955936732831, + "tokens_seen": 572981248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173520561685055, + "loss": 1.5027, + "theoretical_loss": 3.886909516089822, + "tokens_seen": 573046784 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041734202607823475, + "loss": 1.6833, + "theoretical_loss": 3.8868631022416533, + "tokens_seen": 573112320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173319959879639, + "loss": 1.6345, + "theoretical_loss": 3.886816695186552, + "tokens_seen": 573177856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173219658976931, + "loss": 1.5474, + "theoretical_loss": 3.886770294922748, + "tokens_seen": 573243392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173119358074223, + "loss": 1.4757, + "theoretical_loss": 3.8867239014484705, + "tokens_seen": 573308928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004173019057171515, + "loss": 1.6575, + "theoretical_loss": 3.886677514761951, + "tokens_seen": 573374464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 214228, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5797169208526611, + "objective/train/theoretical_loss": 3.8866311348614198, + "objective/train/tokens_used": 593900000, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041729187562688065, + "loss": 1.5166, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041728184553660984, + "loss": 1.4566, + "theoretical_loss": 3.8865847617451097, + "tokens_seen": 573505536 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417271815446339, + "loss": 1.6076, + "theoretical_loss": 3.8865383954112525, + "tokens_seen": 573571072 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041726178535606825, + "loss": 1.648, + "theoretical_loss": 3.8864920358580823, + "tokens_seen": 573636608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172517552657974, + "loss": 1.5234, + "theoretical_loss": 3.8864456830838314, + "tokens_seen": 573702144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172417251755266, + "loss": 1.5904, + "theoretical_loss": 3.8863993370867362, + "tokens_seen": 573767680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004172316950852558, + "loss": 1.4768, + "theoretical_loss": 3.886352997865031, + "tokens_seen": 573833216 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417221664994985, + "loss": 1.472, + "theoretical_loss": 3.8863066654169516, + "tokens_seen": 573898752 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041721163490471416, + "loss": 1.6119, + "theoretical_loss": 3.886260339740735, + "tokens_seen": 573964288 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041720160481444334, + "loss": 1.6434, + "theoretical_loss": 3.8862140208346174, + "tokens_seen": 574029824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171915747241725, + "loss": 1.5521, + "theoretical_loss": 3.8861677086968376, + "tokens_seen": 574095360 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041718154463390176, + "loss": 1.4816, + "theoretical_loss": 3.8861214033256344, + "tokens_seen": 574160896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171715145436309, + "loss": 1.5598, + "theoretical_loss": 3.8860751047192457, + "tokens_seen": 574226432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171614844533601, + "loss": 1.6319, + "theoretical_loss": 3.886028812875912, + "tokens_seen": 574291968 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041715145436308924, + "loss": 1.5108, + "theoretical_loss": 3.885982527793874, + "tokens_seen": 574357504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171414242728185, + "loss": 1.6171, + "theoretical_loss": 3.8859362494713725, + "tokens_seen": 574423040 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041713139418254766, + "loss": 1.6125, + "theoretical_loss": 3.8858899779066496, + "tokens_seen": 574488576 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041712136409227684, + "loss": 1.5399, + "theoretical_loss": 3.8858437130979473, + "tokens_seen": 574554112 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417111334002006, + "loss": 1.6706, + "theoretical_loss": 3.8857974550435097, + "tokens_seen": 574619648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004171013039117352, + "loss": 1.5906, + "theoretical_loss": 3.885751203741579, + "tokens_seen": 574685184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170912738214644, + "loss": 1.5329, + "theoretical_loss": 3.8857049591904014, + "tokens_seen": 574750720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170812437311936, + "loss": 1.3944, + "theoretical_loss": 3.8856587213882205, + "tokens_seen": 574816256 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041707121364092275, + "loss": 1.4804, + "theoretical_loss": 3.8856124903332825, + "tokens_seen": 574881792 + }, + { + "epoch": 0.17, + "learning_rate": 0.000417061183550652, + "loss": 1.6557, + "theoretical_loss": 3.885566266023834, + "tokens_seen": 574947328 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041705115346038116, + "loss": 1.5969, + "theoretical_loss": 3.8855200484581225, + "tokens_seen": 575012864 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041704112337011035, + "loss": 1.5381, + "theoretical_loss": 3.885473837634395, + "tokens_seen": 575078400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170310932798395, + "loss": 1.4673, + "theoretical_loss": 3.8854276335508997, + "tokens_seen": 575143936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170210631895687, + "loss": 1.5633, + "theoretical_loss": 3.885381436205886, + "tokens_seen": 575209472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170110330992979, + "loss": 1.5944, + "theoretical_loss": 3.885335245597604, + "tokens_seen": 575275008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004170010030090271, + "loss": 1.5713, + "theoretical_loss": 3.885289061724303, + "tokens_seen": 575340544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041699097291875625, + "loss": 1.5978, + "theoretical_loss": 3.885242884584235, + "tokens_seen": 575406080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169809428284855, + "loss": 1.4786, + "theoretical_loss": 3.8851967141756507, + "tokens_seen": 575471616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169709127382146, + "loss": 1.4178, + "theoretical_loss": 3.885150550496803, + "tokens_seen": 575537152 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041696088264794385, + "loss": 1.5602, + "theoretical_loss": 3.8851043935459453, + "tokens_seen": 575602688 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041695085255767303, + "loss": 1.6361, + "theoretical_loss": 3.88505824332133, + "tokens_seen": 575668224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169408224674022, + "loss": 1.5458, + "theoretical_loss": 3.8850120998212123, + "tokens_seen": 575733760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169307923771314, + "loss": 1.7101, + "theoretical_loss": 3.884965963043847, + "tokens_seen": 575799296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004169207622868606, + "loss": 1.5005, + "theoretical_loss": 3.884919832987489, + "tokens_seen": 575864832 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041691073219658975, + "loss": 1.5493, + "theoretical_loss": 3.884873709650395, + "tokens_seen": 575930368 + }, + { + "epoch": 0.17, + "learning_rate": 0.000416900702106319, + "loss": 1.5055, + "theoretical_loss": 3.8848275930308223, + "tokens_seen": 575995904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168906720160481, + "loss": 1.3831, + "theoretical_loss": 3.884781483127027, + "tokens_seen": 576061440 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041688064192577735, + "loss": 1.4587, + "theoretical_loss": 3.8847353799372684, + "tokens_seen": 576126976 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041687061183550653, + "loss": 1.5696, + "theoretical_loss": 3.8846892834598052, + "tokens_seen": 576192512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168605817452357, + "loss": 1.5854, + "theoretical_loss": 3.884643193692896, + "tokens_seen": 576258048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168505516549649, + "loss": 1.5782, + "theoretical_loss": 3.884597110634802, + "tokens_seen": 576323584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168405215646941, + "loss": 1.4746, + "theoretical_loss": 3.884551034283784, + "tokens_seen": 576389120 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041683049147442326, + "loss": 1.4483, + "theoretical_loss": 3.8845049646381025, + "tokens_seen": 576454656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168204613841525, + "loss": 1.5359, + "theoretical_loss": 3.8844589016960196, + "tokens_seen": 576520192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004168104312938816, + "loss": 1.4345, + "theoretical_loss": 3.884412845455799, + "tokens_seen": 576585728 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041680040120361085, + "loss": 1.5103, + "theoretical_loss": 3.8843667959157027, + "tokens_seen": 576651264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 215374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5996088981628418, + "objective/train/theoretical_loss": 3.884320753073996, + "objective/train/tokens_used": 597176800, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041679037111334, + "loss": 1.5453, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167803410230692, + "loss": 1.577, + "theoretical_loss": 3.884274716928942, + "tokens_seen": 576782336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167703109327984, + "loss": 1.5001, + "theoretical_loss": 3.8842286874788075, + "tokens_seen": 576847872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167602808425276, + "loss": 1.5851, + "theoretical_loss": 3.8841826647218576, + "tokens_seen": 576913408 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041675025075225676, + "loss": 1.5635, + "theoretical_loss": 3.8841366486563595, + "tokens_seen": 576978944 + }, + { + "epoch": 0.17, + "learning_rate": 0.000416740220661986, + "loss": 1.5443, + "theoretical_loss": 3.8840906392805796, + "tokens_seen": 577044480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167301905717151, + "loss": 1.5819, + "theoretical_loss": 3.8840446365927863, + "tokens_seen": 577110016 + }, + { + "epoch": 0.17, + "learning_rate": 0.00041672016048144436, + "loss": 1.5951, + "theoretical_loss": 3.8839986405912477, + "tokens_seen": 577175552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167101303911735, + "loss": 1.606, + "theoretical_loss": 3.8839526512742335, + "tokens_seen": 577241088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004167001003009027, + "loss": 1.5442, + "theoretical_loss": 3.883906668640013, + "tokens_seen": 577306624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004166900702106319, + "loss": 1.6966, + "theoretical_loss": 3.8838606926868575, + "tokens_seen": 577372160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0004166800401203611, + "loss": 1.4823, + "theoretical_loss": 3.8838147234130367, + "tokens_seen": 577437696 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041667001003009026, + "loss": 1.4034, + "theoretical_loss": 3.8837687608168237, + "tokens_seen": 577503232 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041665997993981944, + "loss": 1.5194, + "theoretical_loss": 3.8837228048964896, + "tokens_seen": 577568768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166499498495487, + "loss": 1.561, + "theoretical_loss": 3.8836768556503083, + "tokens_seen": 577634304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041663991975927786, + "loss": 1.5852, + "theoretical_loss": 3.883630913076553, + "tokens_seen": 577699840 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041662988966900704, + "loss": 1.602, + "theoretical_loss": 3.883584977173499, + "tokens_seen": 577765376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166198595787362, + "loss": 1.5851, + "theoretical_loss": 3.8835390479394203, + "tokens_seen": 577830912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004166098294884654, + "loss": 1.4518, + "theoretical_loss": 3.883493125372592, + "tokens_seen": 577896448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165997993981946, + "loss": 1.5728, + "theoretical_loss": 3.8834472094712913, + "tokens_seen": 577961984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165897693079238, + "loss": 1.6439, + "theoretical_loss": 3.8834013002337953, + "tokens_seen": 578027520 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041657973921765295, + "loss": 1.5251, + "theoretical_loss": 3.883355397658381, + "tokens_seen": 578093056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165697091273822, + "loss": 1.5943, + "theoretical_loss": 3.8833095017433257, + "tokens_seen": 578158592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041655967903711136, + "loss": 1.5388, + "theoretical_loss": 3.8832636124869095, + "tokens_seen": 578224128 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041654964894684055, + "loss": 1.5718, + "theoretical_loss": 3.883217729887411, + "tokens_seen": 578289664 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165396188565697, + "loss": 1.5769, + "theoretical_loss": 3.883171853943111, + "tokens_seen": 578355200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165295887662989, + "loss": 1.6842, + "theoretical_loss": 3.8831259846522896, + "tokens_seen": 578420736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165195586760281, + "loss": 1.4667, + "theoretical_loss": 3.8830801220132285, + "tokens_seen": 578486272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004165095285857573, + "loss": 1.5493, + "theoretical_loss": 3.883034266024209, + "tokens_seen": 578551808 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041649949849548645, + "loss": 1.4489, + "theoretical_loss": 3.8829884166835154, + "tokens_seen": 578617344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164894684052157, + "loss": 1.5625, + "theoretical_loss": 3.882942573989429, + "tokens_seen": 578682880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164794383149448, + "loss": 1.5648, + "theoretical_loss": 3.882896737940235, + "tokens_seen": 578748416 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041646940822467405, + "loss": 1.4817, + "theoretical_loss": 3.882850908534217, + "tokens_seen": 578813952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041645937813440323, + "loss": 1.6169, + "theoretical_loss": 3.882805085769661, + "tokens_seen": 578879488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164493480441324, + "loss": 1.6288, + "theoretical_loss": 3.8827592696448523, + "tokens_seen": 578945024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164393179538616, + "loss": 1.5648, + "theoretical_loss": 3.8827134601580777, + "tokens_seen": 579010560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164292878635908, + "loss": 1.5079, + "theoretical_loss": 3.8826676573076235, + "tokens_seen": 579076096 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041641925777331995, + "loss": 1.5607, + "theoretical_loss": 3.8826218610917786, + "tokens_seen": 579141632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004164092276830492, + "loss": 1.5921, + "theoretical_loss": 3.8825760715088307, + "tokens_seen": 579207168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163991975927783, + "loss": 1.4809, + "theoretical_loss": 3.882530288557069, + "tokens_seen": 579272704 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041638916750250755, + "loss": 1.5928, + "theoretical_loss": 3.8824845122347824, + "tokens_seen": 579338240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041637913741223673, + "loss": 1.5355, + "theoretical_loss": 3.8824387425402618, + "tokens_seen": 579403776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163691073219659, + "loss": 1.508, + "theoretical_loss": 3.8823929794717986, + "tokens_seen": 579469312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163590772316951, + "loss": 1.6645, + "theoretical_loss": 3.882347223027683, + "tokens_seen": 579534848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163490471414243, + "loss": 1.5475, + "theoretical_loss": 3.882301473206209, + "tokens_seen": 579600384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041633901705115346, + "loss": 1.5668, + "theoretical_loss": 3.8822557300056673, + "tokens_seen": 579665920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163289869608827, + "loss": 1.527, + "theoretical_loss": 3.882209993424353, + "tokens_seen": 579731456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004163189568706118, + "loss": 1.5318, + "theoretical_loss": 3.8821642634605587, + "tokens_seen": 579796992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041630892678034105, + "loss": 1.5975, + "theoretical_loss": 3.8821185401125806, + "tokens_seen": 579862528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162988966900702, + "loss": 1.5123, + "theoretical_loss": 3.8820728233787136, + "tokens_seen": 579928064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 216668, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6755932569503784, + "objective/train/theoretical_loss": 3.8820271132572524, + "objective/train/tokens_used": 600453600, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162888665997994, + "loss": 1.6197, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162788365095286, + "loss": 1.6428, + "theoretical_loss": 3.881981409746495, + "tokens_seen": 580059136 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162688064192578, + "loss": 1.5372, + "theoretical_loss": 3.8819357128447383, + "tokens_seen": 580124672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041625877632898696, + "loss": 1.4963, + "theoretical_loss": 3.88189002255028, + "tokens_seen": 580190208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162487462387162, + "loss": 1.5713, + "theoretical_loss": 3.8818443388614186, + "tokens_seen": 580255744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162387161484453, + "loss": 1.5405, + "theoretical_loss": 3.881798661776453, + "tokens_seen": 580321280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041622868605817456, + "loss": 1.6068, + "theoretical_loss": 3.8817529912936832, + "tokens_seen": 580386816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162186559679037, + "loss": 1.5531, + "theoretical_loss": 3.88170732741141, + "tokens_seen": 580452352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004162086258776329, + "loss": 1.6163, + "theoretical_loss": 3.8816616701279334, + "tokens_seen": 580517888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161985957873621, + "loss": 1.5572, + "theoretical_loss": 3.881616019441556, + "tokens_seen": 580583424 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161885656970913, + "loss": 1.4744, + "theoretical_loss": 3.881570375350579, + "tokens_seen": 580648960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041617853560682046, + "loss": 1.6072, + "theoretical_loss": 3.881524737853306, + "tokens_seen": 580714496 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041616850551654964, + "loss": 1.6137, + "theoretical_loss": 3.881479106948041, + "tokens_seen": 580780032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161584754262788, + "loss": 1.4817, + "theoretical_loss": 3.8814334826330876, + "tokens_seen": 580845568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041614844533600806, + "loss": 1.5968, + "theoretical_loss": 3.88138786490675, + "tokens_seen": 580911104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161384152457372, + "loss": 1.5403, + "theoretical_loss": 3.8813422537673343, + "tokens_seen": 580976640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161283851554664, + "loss": 1.571, + "theoretical_loss": 3.881296649213146, + "tokens_seen": 581042176 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041611835506519555, + "loss": 1.5253, + "theoretical_loss": 3.8812510512424927, + "tokens_seen": 581107712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004161083249749248, + "loss": 1.7042, + "theoretical_loss": 3.881205459853681, + "tokens_seen": 581173248 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041609829488465397, + "loss": 1.5438, + "theoretical_loss": 3.8811598750450185, + "tokens_seen": 581238784 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041608826479438315, + "loss": 1.5868, + "theoretical_loss": 3.8811142968148142, + "tokens_seen": 581304320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041607823470411233, + "loss": 1.5059, + "theoretical_loss": 3.881068725161377, + "tokens_seen": 581369856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041606820461384156, + "loss": 1.5025, + "theoretical_loss": 3.881023160083018, + "tokens_seen": 581435392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160581745235707, + "loss": 1.5889, + "theoretical_loss": 3.8809776015780453, + "tokens_seen": 581500928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160481444332999, + "loss": 1.5099, + "theoretical_loss": 3.880932049644771, + "tokens_seen": 581566464 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041603811434302905, + "loss": 1.5388, + "theoretical_loss": 3.8808865042815075, + "tokens_seen": 581632000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004160280842527583, + "loss": 1.5148, + "theoretical_loss": 3.880840965486566, + "tokens_seen": 581697536 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041601805416248747, + "loss": 1.4971, + "theoretical_loss": 3.8807954332582595, + "tokens_seen": 581763072 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041600802407221665, + "loss": 1.626, + "theoretical_loss": 3.880749907594902, + "tokens_seen": 581828608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041599799398194583, + "loss": 1.58, + "theoretical_loss": 3.8807043884948085, + "tokens_seen": 581894144 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415987963891675, + "loss": 1.5812, + "theoretical_loss": 3.880658875956292, + "tokens_seen": 581959680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159779338014042, + "loss": 1.5244, + "theoretical_loss": 3.8806133699776684, + "tokens_seen": 582025216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041596790371113343, + "loss": 1.5446, + "theoretical_loss": 3.880567870557254, + "tokens_seen": 582090752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041595787362086256, + "loss": 1.6739, + "theoretical_loss": 3.8805223776933655, + "tokens_seen": 582156288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159478435305918, + "loss": 1.711, + "theoretical_loss": 3.88047689138432, + "tokens_seen": 582221824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159378134403209, + "loss": 1.5112, + "theoretical_loss": 3.880431411628436, + "tokens_seen": 582287360 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041592778335005015, + "loss": 1.6038, + "theoretical_loss": 3.8803859384240305, + "tokens_seen": 582352896 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041591775325977934, + "loss": 1.5057, + "theoretical_loss": 3.880340471769424, + "tokens_seen": 582418432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004159077231695085, + "loss": 1.6025, + "theoretical_loss": 3.880295011662936, + "tokens_seen": 582483968 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041589769307923775, + "loss": 1.5363, + "theoretical_loss": 3.8802495581028857, + "tokens_seen": 582549504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041588766298896693, + "loss": 1.7327, + "theoretical_loss": 3.880204111087596, + "tokens_seen": 582615040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158776328986961, + "loss": 1.5552, + "theoretical_loss": 3.880158670615387, + "tokens_seen": 582680576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158676028084253, + "loss": 1.5334, + "theoretical_loss": 3.880113236684582, + "tokens_seen": 582746112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158575727181545, + "loss": 1.6321, + "theoretical_loss": 3.880067809293503, + "tokens_seen": 582811648 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041584754262788366, + "loss": 1.6736, + "theoretical_loss": 3.880022388440474, + "tokens_seen": 582877184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158375125376129, + "loss": 1.5741, + "theoretical_loss": 3.879976974123818, + "tokens_seen": 582942720 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415827482447342, + "loss": 1.5948, + "theoretical_loss": 3.8799315663418614, + "tokens_seen": 583008256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041581745235707126, + "loss": 1.4289, + "theoretical_loss": 3.8798861650929277, + "tokens_seen": 583073792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004158074222668004, + "loss": 1.5972, + "theoretical_loss": 3.8798407703753446, + "tokens_seen": 583139328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157973921765296, + "loss": 1.5913, + "theoretical_loss": 3.879795382187438, + "tokens_seen": 583204864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 217803, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6284210681915283, + "objective/train/theoretical_loss": 3.879750000527534, + "objective/train/tokens_used": 603730400, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157873620862588, + "loss": 1.598, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415777331995988, + "loss": 1.56, + "theoretical_loss": 3.8797046253939618, + "tokens_seen": 583335936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041576730190571716, + "loss": 1.606, + "theoretical_loss": 3.8796592567850485, + "tokens_seen": 583401472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157572718154464, + "loss": 1.56, + "theoretical_loss": 3.8796138946991245, + "tokens_seen": 583467008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157472417251755, + "loss": 1.6742, + "theoretical_loss": 3.8795685391345183, + "tokens_seen": 583532544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041573721163490476, + "loss": 1.5776, + "theoretical_loss": 3.879523190089561, + "tokens_seen": 583598080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157271815446339, + "loss": 1.5554, + "theoretical_loss": 3.8794778475625833, + "tokens_seen": 583663616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157171514543631, + "loss": 1.6388, + "theoretical_loss": 3.879432511551916, + "tokens_seen": 583729152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004157071213640923, + "loss": 1.4938, + "theoretical_loss": 3.8793871820558916, + "tokens_seen": 583794688 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156970912738215, + "loss": 1.6061, + "theoretical_loss": 3.8793418590728423, + "tokens_seen": 583860224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041568706118355066, + "loss": 1.54, + "theoretical_loss": 3.879296542601103, + "tokens_seen": 583925760 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041567703109327984, + "loss": 1.5004, + "theoretical_loss": 3.879251232639006, + "tokens_seen": 583991296 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415667001003009, + "loss": 1.5915, + "theoretical_loss": 3.879205929184886, + "tokens_seen": 584056832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041565697091273826, + "loss": 1.4949, + "theoretical_loss": 3.8791606322370793, + "tokens_seen": 584122368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156469408224674, + "loss": 1.5213, + "theoretical_loss": 3.8791153417939204, + "tokens_seen": 584187904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004156369107321966, + "loss": 1.5018, + "theoretical_loss": 3.8790700578537463, + "tokens_seen": 584253440 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041562688064192575, + "loss": 1.674, + "theoretical_loss": 3.879024780414894, + "tokens_seen": 584318976 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415616850551655, + "loss": 1.5753, + "theoretical_loss": 3.878979509475701, + "tokens_seen": 584384512 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041560682046138417, + "loss": 1.5818, + "theoretical_loss": 3.8789342450345052, + "tokens_seen": 584450048 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041559679037111335, + "loss": 1.5462, + "theoretical_loss": 3.8788889870896455, + "tokens_seen": 584515584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041558676028084253, + "loss": 1.5196, + "theoretical_loss": 3.878843735639462, + "tokens_seen": 584581120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041557673019057176, + "loss": 1.4895, + "theoretical_loss": 3.8787984906822945, + "tokens_seen": 584646656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155667001003009, + "loss": 1.5543, + "theoretical_loss": 3.878753252216483, + "tokens_seen": 584712192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155566700100301, + "loss": 1.6333, + "theoretical_loss": 3.8787080202403694, + "tokens_seen": 584777728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041554663991975925, + "loss": 1.5353, + "theoretical_loss": 3.8786627947522954, + "tokens_seen": 584843264 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004155366098294885, + "loss": 1.6012, + "theoretical_loss": 3.878617575750603, + "tokens_seen": 584908800 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041552657973921767, + "loss": 1.5803, + "theoretical_loss": 3.8785723632336366, + "tokens_seen": 584974336 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041551654964894685, + "loss": 1.4692, + "theoretical_loss": 3.8785271571997386, + "tokens_seen": 585039872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041550651955867603, + "loss": 1.646, + "theoretical_loss": 3.878481957647254, + "tokens_seen": 585105408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154964894684052, + "loss": 1.5006, + "theoretical_loss": 3.878436764574527, + "tokens_seen": 585170944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154864593781344, + "loss": 1.5791, + "theoretical_loss": 3.8783915779799036, + "tokens_seen": 585236480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041547642928786363, + "loss": 1.6243, + "theoretical_loss": 3.8783463978617303, + "tokens_seen": 585302016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041546639919759276, + "loss": 1.5505, + "theoretical_loss": 3.878301224218353, + "tokens_seen": 585367552 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415456369107322, + "loss": 1.5083, + "theoretical_loss": 3.8782560570481204, + "tokens_seen": 585433088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154463390170511, + "loss": 1.526, + "theoretical_loss": 3.8782108963493793, + "tokens_seen": 585498624 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041543630892678035, + "loss": 1.6231, + "theoretical_loss": 3.878165742120478, + "tokens_seen": 585564160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041542627883650954, + "loss": 1.523, + "theoretical_loss": 3.8781205943597667, + "tokens_seen": 585629696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154162487462387, + "loss": 1.5883, + "theoretical_loss": 3.8780754530655948, + "tokens_seen": 585695232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004154062186559679, + "loss": 1.4855, + "theoretical_loss": 3.8780303182363123, + "tokens_seen": 585760768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041539618856569713, + "loss": 1.567, + "theoretical_loss": 3.877985189870271, + "tokens_seen": 585826304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041538615847542626, + "loss": 1.5345, + "theoretical_loss": 3.877940067965821, + "tokens_seen": 585891840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153761283851555, + "loss": 1.4442, + "theoretical_loss": 3.8778949525213164, + "tokens_seen": 585957376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153660982948846, + "loss": 1.5314, + "theoretical_loss": 3.8778498435351088, + "tokens_seen": 586022912 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041535606820461386, + "loss": 1.6421, + "theoretical_loss": 3.8778047410055514, + "tokens_seen": 586088448 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041534603811434304, + "loss": 1.5255, + "theoretical_loss": 3.8777596449309994, + "tokens_seen": 586153984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153360080240722, + "loss": 1.6509, + "theoretical_loss": 3.8777145553098062, + "tokens_seen": 586219520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153259779338014, + "loss": 1.5729, + "theoretical_loss": 3.8776694721403278, + "tokens_seen": 586285056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004153159478435306, + "loss": 1.5137, + "theoretical_loss": 3.877624395420919, + "tokens_seen": 586350592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041530591775325976, + "loss": 1.5099, + "theoretical_loss": 3.8775793251499375, + "tokens_seen": 586416128 + }, + { + "epoch": 0.18, + "learning_rate": 0.000415295887662989, + "loss": 1.5723, + "theoretical_loss": 3.8775342613257395, + "tokens_seen": 586481664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 218927, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7090811729431152, + "objective/train/theoretical_loss": 3.8774892039466833, + "objective/train/tokens_used": 607007200, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152858575727181, + "loss": 1.6711, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041527582748244736, + "loss": 1.5107, + "theoretical_loss": 3.877444153011126, + "tokens_seen": 586612736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152657973921765, + "loss": 1.5179, + "theoretical_loss": 3.8773991085174284, + "tokens_seen": 586678272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152557673019057, + "loss": 1.4997, + "theoretical_loss": 3.8773540704639475, + "tokens_seen": 586743808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152457372116349, + "loss": 1.5859, + "theoretical_loss": 3.8773090388490448, + "tokens_seen": 586809344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152357071213641, + "loss": 1.4881, + "theoretical_loss": 3.8772640136710805, + "tokens_seen": 586874880 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041522567703109327, + "loss": 1.5451, + "theoretical_loss": 3.877218994928416, + "tokens_seen": 586940416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004152156469408225, + "loss": 1.662, + "theoretical_loss": 3.8771739826194134, + "tokens_seen": 587005952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041520561685055163, + "loss": 1.5637, + "theoretical_loss": 3.8771289767424344, + "tokens_seen": 587071488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041519558676028086, + "loss": 1.6109, + "theoretical_loss": 3.8770839772958428, + "tokens_seen": 587137024 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041518555667001, + "loss": 1.5428, + "theoretical_loss": 3.877038984278002, + "tokens_seen": 587202560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151755265797392, + "loss": 1.5548, + "theoretical_loss": 3.8769939976872756, + "tokens_seen": 587268096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151654964894684, + "loss": 1.506, + "theoretical_loss": 3.8769490175220294, + "tokens_seen": 587333632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151554663991976, + "loss": 1.4761, + "theoretical_loss": 3.876904043780628, + "tokens_seen": 587399168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151454363089268, + "loss": 1.5793, + "theoretical_loss": 3.8768590764614377, + "tokens_seen": 587464704 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041513540621865595, + "loss": 1.5357, + "theoretical_loss": 3.8768141155628255, + "tokens_seen": 587530240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004151253761283852, + "loss": 1.5523, + "theoretical_loss": 3.8767691610831587, + "tokens_seen": 587595776 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041511534603811437, + "loss": 1.5397, + "theoretical_loss": 3.876724213020804, + "tokens_seen": 587661312 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041510531594784355, + "loss": 1.5114, + "theoretical_loss": 3.876679271374131, + "tokens_seen": 587726848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041509528585757273, + "loss": 1.4844, + "theoretical_loss": 3.8766343361415085, + "tokens_seen": 587792384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041508525576730196, + "loss": 1.5901, + "theoretical_loss": 3.876589407321305, + "tokens_seen": 587857920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150752256770311, + "loss": 1.5283, + "theoretical_loss": 3.8765444849118924, + "tokens_seen": 587923456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150651955867603, + "loss": 1.4869, + "theoretical_loss": 3.8764995689116404, + "tokens_seen": 587988992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041505516549648945, + "loss": 1.5216, + "theoretical_loss": 3.8764546593189206, + "tokens_seen": 588054528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150451354062187, + "loss": 1.5791, + "theoretical_loss": 3.876409756132105, + "tokens_seen": 588120064 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041503510531594787, + "loss": 1.5514, + "theoretical_loss": 3.8763648593495663, + "tokens_seen": 588185600 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041502507522567705, + "loss": 1.6035, + "theoretical_loss": 3.8763199689696783, + "tokens_seen": 588251136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041501504513540623, + "loss": 1.4976, + "theoretical_loss": 3.876275084990813, + "tokens_seen": 588316672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004150050150451354, + "loss": 1.3901, + "theoretical_loss": 3.876230207411347, + "tokens_seen": 588382208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149949849548646, + "loss": 1.4355, + "theoretical_loss": 3.8761853362296534, + "tokens_seen": 588447744 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041498495486459383, + "loss": 1.4812, + "theoretical_loss": 3.8761404714441086, + "tokens_seen": 588513280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041497492477432296, + "loss": 1.4845, + "theoretical_loss": 3.8760956130530886, + "tokens_seen": 588578816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149648946840522, + "loss": 1.6174, + "theoretical_loss": 3.8760507610549704, + "tokens_seen": 588644352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149548645937813, + "loss": 1.5942, + "theoretical_loss": 3.876005915448131, + "tokens_seen": 588709888 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041494483450351055, + "loss": 1.5002, + "theoretical_loss": 3.8759610762309484, + "tokens_seen": 588775424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041493480441323974, + "loss": 1.6294, + "theoretical_loss": 3.875916243401801, + "tokens_seen": 588840960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149247743229689, + "loss": 1.5705, + "theoretical_loss": 3.8758714169590682, + "tokens_seen": 588906496 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004149147442326981, + "loss": 1.6387, + "theoretical_loss": 3.87582659690113, + "tokens_seen": 588972032 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041490471414242733, + "loss": 1.613, + "theoretical_loss": 3.875781783226366, + "tokens_seen": 589037568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041489468405215646, + "loss": 1.5534, + "theoretical_loss": 3.8757369759331572, + "tokens_seen": 589103104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148846539618857, + "loss": 1.4693, + "theoretical_loss": 3.8756921750198856, + "tokens_seen": 589168640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148746238716148, + "loss": 1.4701, + "theoretical_loss": 3.8756473804849323, + "tokens_seen": 589234176 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041486459378134406, + "loss": 1.5529, + "theoretical_loss": 3.8756025923266813, + "tokens_seen": 589299712 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041485456369107324, + "loss": 1.6067, + "theoretical_loss": 3.8755578105435147, + "tokens_seen": 589365248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148445336008024, + "loss": 1.5353, + "theoretical_loss": 3.8755130351338165, + "tokens_seen": 589430784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148345035105316, + "loss": 1.6998, + "theoretical_loss": 3.8754682660959716, + "tokens_seen": 589496320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148244734202608, + "loss": 1.503, + "theoretical_loss": 3.8754235034283653, + "tokens_seen": 589561856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041481444332998996, + "loss": 1.5063, + "theoretical_loss": 3.875378747129382, + "tokens_seen": 589627392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004148044132397192, + "loss": 1.5657, + "theoretical_loss": 3.8753339971974086, + "tokens_seen": 589692928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147943831494483, + "loss": 1.538, + "theoretical_loss": 3.8752892536308323, + "tokens_seen": 589758464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 220165, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.611093521118164, + "objective/train/theoretical_loss": 3.87524451642804, + "objective/train/tokens_used": 610284000, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041478435305917756, + "loss": 1.6287, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147743229689067, + "loss": 1.5164, + "theoretical_loss": 3.8751997855874194, + "tokens_seen": 589889536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147642928786359, + "loss": 1.5688, + "theoretical_loss": 3.8751550611073595, + "tokens_seen": 589955072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147542627883651, + "loss": 1.5044, + "theoretical_loss": 3.875110342986249, + "tokens_seen": 590020608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147442326980943, + "loss": 1.5611, + "theoretical_loss": 3.875065631222478, + "tokens_seen": 590086144 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041473420260782347, + "loss": 1.6015, + "theoretical_loss": 3.875020925814437, + "tokens_seen": 590151680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004147241725175527, + "loss": 1.7216, + "theoretical_loss": 3.8749762267605163, + "tokens_seen": 590217216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041471414242728183, + "loss": 1.5556, + "theoretical_loss": 3.8749315340591073, + "tokens_seen": 590282752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041470411233701106, + "loss": 1.6051, + "theoretical_loss": 3.874886847708603, + "tokens_seen": 590348288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146940822467402, + "loss": 1.5354, + "theoretical_loss": 3.8748421677073956, + "tokens_seen": 590413824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146840521564694, + "loss": 1.5323, + "theoretical_loss": 3.874797494053878, + "tokens_seen": 590479360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146740220661986, + "loss": 1.5312, + "theoretical_loss": 3.8747528267464437, + "tokens_seen": 590544896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146639919759278, + "loss": 1.3478, + "theoretical_loss": 3.8747081657834874, + "tokens_seen": 590610432 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041465396188565697, + "loss": 1.5646, + "theoretical_loss": 3.874663511163405, + "tokens_seen": 590675968 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041464393179538615, + "loss": 1.3896, + "theoretical_loss": 3.8746188628845912, + "tokens_seen": 590741504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041463390170511533, + "loss": 1.5422, + "theoretical_loss": 3.8745742209454423, + "tokens_seen": 590807040 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041462387161484457, + "loss": 1.5003, + "theoretical_loss": 3.874529585344355, + "tokens_seen": 590872576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004146138415245737, + "loss": 1.5367, + "theoretical_loss": 3.874484956079727, + "tokens_seen": 590938112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041460381143430293, + "loss": 1.5244, + "theoretical_loss": 3.8744403331499555, + "tokens_seen": 591003648 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145937813440321, + "loss": 1.5708, + "theoretical_loss": 3.8743957165534395, + "tokens_seen": 591069184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145837512537613, + "loss": 1.667, + "theoretical_loss": 3.8743511062885783, + "tokens_seen": 591134720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145737211634905, + "loss": 1.4601, + "theoretical_loss": 3.874306502353771, + "tokens_seen": 591200256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041456369107321965, + "loss": 1.5867, + "theoretical_loss": 3.8742619047474176, + "tokens_seen": 591265792 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041455366098294884, + "loss": 1.5217, + "theoretical_loss": 3.87421731346792, + "tokens_seen": 591331328 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041454363089267807, + "loss": 1.5563, + "theoretical_loss": 3.874172728513678, + "tokens_seen": 591396864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145336008024072, + "loss": 1.5239, + "theoretical_loss": 3.874128149883095, + "tokens_seen": 591462400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041452357071213643, + "loss": 1.587, + "theoretical_loss": 3.8740835775745732, + "tokens_seen": 591527936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041451354062186556, + "loss": 1.5009, + "theoretical_loss": 3.8740390115865155, + "tokens_seen": 591593472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004145035105315948, + "loss": 1.6688, + "theoretical_loss": 3.873994451917326, + "tokens_seen": 591659008 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414493480441324, + "loss": 1.3954, + "theoretical_loss": 3.8739498985654084, + "tokens_seen": 591724544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041448345035105316, + "loss": 1.5345, + "theoretical_loss": 3.873905351529168, + "tokens_seen": 591790080 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041447342026078234, + "loss": 1.5581, + "theoretical_loss": 3.8738608108070096, + "tokens_seen": 591855616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144633901705115, + "loss": 1.4622, + "theoretical_loss": 3.8738162763973407, + "tokens_seen": 591921152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144533600802407, + "loss": 1.5426, + "theoretical_loss": 3.8737717482985667, + "tokens_seen": 591986688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041444332998996994, + "loss": 1.4854, + "theoretical_loss": 3.8737272265090947, + "tokens_seen": 592052224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041443329989969906, + "loss": 1.4819, + "theoretical_loss": 3.8736827110273335, + "tokens_seen": 592117760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004144232698094283, + "loss": 1.5554, + "theoretical_loss": 3.8736382018516906, + "tokens_seen": 592183296 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041441323971915753, + "loss": 1.5103, + "theoretical_loss": 3.8735936989805744, + "tokens_seen": 592248832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041440320962888666, + "loss": 1.6052, + "theoretical_loss": 3.873549202412396, + "tokens_seen": 592314368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143931795386159, + "loss": 1.5911, + "theoretical_loss": 3.8735047121455644, + "tokens_seen": 592379904 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414383149448345, + "loss": 1.6185, + "theoretical_loss": 3.873460228178491, + "tokens_seen": 592445440 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041437311935807426, + "loss": 1.4381, + "theoretical_loss": 3.8734157505095856, + "tokens_seen": 592510976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041436308926780344, + "loss": 1.5358, + "theoretical_loss": 3.8733712791372614, + "tokens_seen": 592576512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143530591775326, + "loss": 1.5374, + "theoretical_loss": 3.8733268140599306, + "tokens_seen": 592642048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143430290872618, + "loss": 1.485, + "theoretical_loss": 3.8732823552760056, + "tokens_seen": 592707584 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414332998996991, + "loss": 1.5333, + "theoretical_loss": 3.8732379027839, + "tokens_seen": 592773120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041432296890672016, + "loss": 1.5354, + "theoretical_loss": 3.8731934565820287, + "tokens_seen": 592838656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143129388164494, + "loss": 1.4998, + "theoretical_loss": 3.8731490166688056, + "tokens_seen": 592904192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004143029087261785, + "loss": 1.4788, + "theoretical_loss": 3.873104583042646, + "tokens_seen": 592969728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041429287863590776, + "loss": 1.5883, + "theoretical_loss": 3.8730601557019657, + "tokens_seen": 593035264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 221319, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5837016105651855, + "objective/train/theoretical_loss": 3.873015734645182, + "objective/train/tokens_used": 613560800, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142828485456369, + "loss": 1.5478, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142728184553661, + "loss": 1.5399, + "theoretical_loss": 3.872971319870711, + "tokens_seen": 593166336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142627883650953, + "loss": 1.3673, + "theoretical_loss": 3.872926911376971, + "tokens_seen": 593231872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142527582748245, + "loss": 1.5435, + "theoretical_loss": 3.8728825091623786, + "tokens_seen": 593297408 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041424272818455367, + "loss": 1.5358, + "theoretical_loss": 3.872838113225354, + "tokens_seen": 593362944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142326980942829, + "loss": 1.5488, + "theoretical_loss": 3.8727937235643157, + "tokens_seen": 593428480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041422266800401203, + "loss": 1.4744, + "theoretical_loss": 3.872749340177685, + "tokens_seen": 593494016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041421263791374126, + "loss": 1.5856, + "theoretical_loss": 3.87270496306388, + "tokens_seen": 593559552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004142026078234704, + "loss": 1.6172, + "theoretical_loss": 3.8726605922213233, + "tokens_seen": 593625088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141925777331996, + "loss": 1.5334, + "theoretical_loss": 3.872616227648436, + "tokens_seen": 593690624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141825476429288, + "loss": 1.4793, + "theoretical_loss": 3.8725718693436404, + "tokens_seen": 593756160 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414172517552658, + "loss": 1.4931, + "theoretical_loss": 3.8725275173053593, + "tokens_seen": 593821696 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041416248746238717, + "loss": 1.4671, + "theoretical_loss": 3.872483171532015, + "tokens_seen": 593887232 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041415245737211635, + "loss": 1.5337, + "theoretical_loss": 3.8724388320220333, + "tokens_seen": 593952768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041414242728184553, + "loss": 1.5315, + "theoretical_loss": 3.872394498773837, + "tokens_seen": 594018304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041413239719157477, + "loss": 1.6037, + "theoretical_loss": 3.8723501717858513, + "tokens_seen": 594083840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141223671013039, + "loss": 1.4238, + "theoretical_loss": 3.872305851056502, + "tokens_seen": 594149376 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041411233701103313, + "loss": 1.5569, + "theoretical_loss": 3.8722615365842157, + "tokens_seen": 594214912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004141023069207623, + "loss": 1.6153, + "theoretical_loss": 3.8722172283674183, + "tokens_seen": 594280448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140922768304915, + "loss": 1.5132, + "theoretical_loss": 3.8721729264045375, + "tokens_seen": 594345984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140822467402207, + "loss": 1.4844, + "theoretical_loss": 3.872128630694001, + "tokens_seen": 594411520 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041407221664994985, + "loss": 1.6893, + "theoretical_loss": 3.8720843412342374, + "tokens_seen": 594477056 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041406218655967904, + "loss": 1.5711, + "theoretical_loss": 3.872040058023676, + "tokens_seen": 594542592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041405215646940827, + "loss": 1.5472, + "theoretical_loss": 3.871995781060745, + "tokens_seen": 594608128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140421263791374, + "loss": 1.4503, + "theoretical_loss": 3.8719515103438757, + "tokens_seen": 594673664 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041403209628886663, + "loss": 1.535, + "theoretical_loss": 3.871907245871499, + "tokens_seen": 594739200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041402206619859576, + "loss": 1.6557, + "theoretical_loss": 3.871862987642045, + "tokens_seen": 594804736 + }, + { + "epoch": 0.18, + "learning_rate": 0.000414012036108325, + "loss": 1.414, + "theoretical_loss": 3.871818735653947, + "tokens_seen": 594870272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004140020060180542, + "loss": 1.4242, + "theoretical_loss": 3.8717744899056354, + "tokens_seen": 594935808 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041399197592778336, + "loss": 1.5539, + "theoretical_loss": 3.8717302503955446, + "tokens_seen": 595001344 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041398194583751254, + "loss": 1.5321, + "theoretical_loss": 3.8716860171221086, + "tokens_seen": 595066880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139719157472417, + "loss": 1.5258, + "theoretical_loss": 3.8716417900837596, + "tokens_seen": 595132416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139618856569709, + "loss": 1.5636, + "theoretical_loss": 3.8715975692789337, + "tokens_seen": 595197952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041395185556670014, + "loss": 1.5185, + "theoretical_loss": 3.871553354706066, + "tokens_seen": 595263488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041394182547642926, + "loss": 1.6653, + "theoretical_loss": 3.871509146363592, + "tokens_seen": 595329024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139317953861585, + "loss": 1.5389, + "theoretical_loss": 3.8714649442499476, + "tokens_seen": 595394560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004139217652958877, + "loss": 1.4952, + "theoretical_loss": 3.8714207483635703, + "tokens_seen": 595460096 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041391173520561686, + "loss": 1.4978, + "theoretical_loss": 3.8713765587028974, + "tokens_seen": 595525632 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041390170511534604, + "loss": 1.5303, + "theoretical_loss": 3.871332375266367, + "tokens_seen": 595591168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138916750250752, + "loss": 1.5476, + "theoretical_loss": 3.871288198052418, + "tokens_seen": 595656704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138816449348044, + "loss": 1.568, + "theoretical_loss": 3.871244027059489, + "tokens_seen": 595722240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041387161484453364, + "loss": 1.4072, + "theoretical_loss": 3.8711998622860193, + "tokens_seen": 595787776 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041386158475426277, + "loss": 1.6786, + "theoretical_loss": 3.8711557037304507, + "tokens_seen": 595853312 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413851554663992, + "loss": 1.5269, + "theoretical_loss": 3.8711115513912224, + "tokens_seen": 595918848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041384152457372113, + "loss": 1.4454, + "theoretical_loss": 3.8710674052667766, + "tokens_seen": 595984384 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041383149448345036, + "loss": 1.5877, + "theoretical_loss": 3.8710232653555554, + "tokens_seen": 596049920 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041382146439317954, + "loss": 1.4906, + "theoretical_loss": 3.870979131656001, + "tokens_seen": 596115456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138114343029087, + "loss": 1.6314, + "theoretical_loss": 3.870935004166557, + "tokens_seen": 596180992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004138014042126379, + "loss": 1.4448, + "theoretical_loss": 3.870890882885667, + "tokens_seen": 596246528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137913741223671, + "loss": 1.5058, + "theoretical_loss": 3.8708467678117744, + "tokens_seen": 596312064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 222574, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4792110919952393, + "objective/train/theoretical_loss": 3.870802658943324, + "objective/train/tokens_used": 616837600, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041378134403209627, + "loss": 1.4595, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137713139418255, + "loss": 1.4722, + "theoretical_loss": 3.8707585562787625, + "tokens_seen": 596443136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041376128385155463, + "loss": 1.6859, + "theoretical_loss": 3.8707144598165346, + "tokens_seen": 596508672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041375125376128387, + "loss": 1.6512, + "theoretical_loss": 3.8706703695550875, + "tokens_seen": 596574208 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041374122367101305, + "loss": 1.3966, + "theoretical_loss": 3.8706262854928672, + "tokens_seen": 596639744 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041373119358074223, + "loss": 1.5831, + "theoretical_loss": 3.8705822076283223, + "tokens_seen": 596705280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137211634904714, + "loss": 1.4628, + "theoretical_loss": 3.870538135959901, + "tokens_seen": 596770816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004137111334002006, + "loss": 1.4412, + "theoretical_loss": 3.8704940704860507, + "tokens_seen": 596836352 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041370110330992977, + "loss": 1.6042, + "theoretical_loss": 3.8704500112052216, + "tokens_seen": 596901888 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413691073219659, + "loss": 1.5478, + "theoretical_loss": 3.870405958115864, + "tokens_seen": 596967424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041368104312938813, + "loss": 1.6623, + "theoretical_loss": 3.870361911216427, + "tokens_seen": 597032960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041367101303911737, + "loss": 1.6151, + "theoretical_loss": 3.8703178705053625, + "tokens_seen": 597098496 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041366098294884655, + "loss": 1.5549, + "theoretical_loss": 3.8702738359811217, + "tokens_seen": 597164032 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041365095285857573, + "loss": 1.5494, + "theoretical_loss": 3.8702298076421564, + "tokens_seen": 597229568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041364092276830497, + "loss": 1.5413, + "theoretical_loss": 3.87018578548692, + "tokens_seen": 597295104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136308926780341, + "loss": 1.4135, + "theoretical_loss": 3.8701417695138645, + "tokens_seen": 597360640 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041362086258776333, + "loss": 1.5345, + "theoretical_loss": 3.8700977597214443, + "tokens_seen": 597426176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136108324974925, + "loss": 1.5172, + "theoretical_loss": 3.870053756108114, + "tokens_seen": 597491712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004136008024072217, + "loss": 1.4219, + "theoretical_loss": 3.8700097586723277, + "tokens_seen": 597557248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135907723169509, + "loss": 1.603, + "theoretical_loss": 3.869965767412541, + "tokens_seen": 597622784 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041358074222668005, + "loss": 1.5437, + "theoretical_loss": 3.8699217823272103, + "tokens_seen": 597688320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041357071213640924, + "loss": 1.6641, + "theoretical_loss": 3.8698778034147914, + "tokens_seen": 597753856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041356068204613847, + "loss": 1.5517, + "theoretical_loss": 3.869833830673742, + "tokens_seen": 597819392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135506519558676, + "loss": 1.5628, + "theoretical_loss": 3.8697898641025192, + "tokens_seen": 597884928 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041354062186559683, + "loss": 1.6777, + "theoretical_loss": 3.8697459036995814, + "tokens_seen": 597950464 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041353059177532596, + "loss": 1.4593, + "theoretical_loss": 3.869701949463387, + "tokens_seen": 598016000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135205616850552, + "loss": 1.566, + "theoretical_loss": 3.869658001392396, + "tokens_seen": 598081536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004135105315947844, + "loss": 1.6169, + "theoretical_loss": 3.869614059485068, + "tokens_seen": 598147072 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041350050150451356, + "loss": 1.616, + "theoretical_loss": 3.8695701237398623, + "tokens_seen": 598212608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041349047141424274, + "loss": 1.6409, + "theoretical_loss": 3.869526194155241, + "tokens_seen": 598278144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134804413239719, + "loss": 1.4779, + "theoretical_loss": 3.869482270729665, + "tokens_seen": 598343680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134704112337011, + "loss": 1.6197, + "theoretical_loss": 3.8694383534615966, + "tokens_seen": 598409216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041346038114343034, + "loss": 1.5556, + "theoretical_loss": 3.8693944423494986, + "tokens_seen": 598474752 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041345035105315946, + "loss": 1.5433, + "theoretical_loss": 3.869350537391834, + "tokens_seen": 598540288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134403209628887, + "loss": 1.4545, + "theoretical_loss": 3.869306638587066, + "tokens_seen": 598605824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134302908726179, + "loss": 1.5754, + "theoretical_loss": 3.8692627459336597, + "tokens_seen": 598671360 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041342026078234706, + "loss": 1.5602, + "theoretical_loss": 3.8692188594300787, + "tokens_seen": 598736896 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041341023069207624, + "loss": 1.499, + "theoretical_loss": 3.8691749790747894, + "tokens_seen": 598802432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004134002006018054, + "loss": 1.4641, + "theoretical_loss": 3.8691311048662573, + "tokens_seen": 598867968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133901705115346, + "loss": 1.5313, + "theoretical_loss": 3.869087236802949, + "tokens_seen": 598933504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041338014042126384, + "loss": 1.5562, + "theoretical_loss": 3.869043374883331, + "tokens_seen": 598999040 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041337011033099297, + "loss": 1.5373, + "theoretical_loss": 3.8689995191058717, + "tokens_seen": 599064576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133600802407222, + "loss": 1.5855, + "theoretical_loss": 3.868955669469039, + "tokens_seen": 599130112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041335005015045133, + "loss": 1.5857, + "theoretical_loss": 3.868911825971301, + "tokens_seen": 599195648 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041334002006018056, + "loss": 1.6275, + "theoretical_loss": 3.868867988611127, + "tokens_seen": 599261184 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041332998996990975, + "loss": 1.6784, + "theoretical_loss": 3.868824157386987, + "tokens_seen": 599326720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133199598796389, + "loss": 1.4761, + "theoretical_loss": 3.8687803322973506, + "tokens_seen": 599392256 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004133099297893681, + "loss": 1.4833, + "theoretical_loss": 3.86873651334069, + "tokens_seen": 599457792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132998996990973, + "loss": 1.5333, + "theoretical_loss": 3.8686927005154756, + "tokens_seen": 599523328 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041328986960882647, + "loss": 1.478, + "theoretical_loss": 3.8686488938201795, + "tokens_seen": 599588864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 223680, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.61704683303833, + "objective/train/theoretical_loss": 3.868605093253274, + "objective/train/tokens_used": 620114400, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132798395185557, + "loss": 1.5179, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041326980942828483, + "loss": 1.5805, + "theoretical_loss": 3.8685612988132334, + "tokens_seen": 599719936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041325977933801407, + "loss": 1.6116, + "theoretical_loss": 3.8685175104985294, + "tokens_seen": 599785472 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041324974924774325, + "loss": 1.5125, + "theoretical_loss": 3.868473728307637, + "tokens_seen": 599851008 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041323971915747243, + "loss": 1.6054, + "theoretical_loss": 3.868429952239031, + "tokens_seen": 599916544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132296890672016, + "loss": 1.595, + "theoretical_loss": 3.8683861822911867, + "tokens_seen": 599982080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004132196589769308, + "loss": 1.5103, + "theoretical_loss": 3.8683424184625794, + "tokens_seen": 600047616 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041320962888665997, + "loss": 1.5152, + "theoretical_loss": 3.8682986607516856, + "tokens_seen": 600113152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131995987963892, + "loss": 1.6053, + "theoretical_loss": 3.8682549091569824, + "tokens_seen": 600178688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041318956870611833, + "loss": 1.5202, + "theoretical_loss": 3.868211163676947, + "tokens_seen": 600244224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041317953861584757, + "loss": 1.6938, + "theoretical_loss": 3.868167424310057, + "tokens_seen": 600309760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131695085255767, + "loss": 1.6065, + "theoretical_loss": 3.868123691054792, + "tokens_seen": 600375296 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041315947843530593, + "loss": 1.5155, + "theoretical_loss": 3.8680799639096297, + "tokens_seen": 600440832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131494483450351, + "loss": 1.561, + "theoretical_loss": 3.8680362428730506, + "tokens_seen": 600506368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131394182547643, + "loss": 1.6046, + "theoretical_loss": 3.8679925279435343, + "tokens_seen": 600571904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131293881644935, + "loss": 1.5377, + "theoretical_loss": 3.8679488191195617, + "tokens_seen": 600637440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004131193580742227, + "loss": 1.592, + "theoretical_loss": 3.8679051163996143, + "tokens_seen": 600702976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041310932798395184, + "loss": 1.49, + "theoretical_loss": 3.867861419782173, + "tokens_seen": 600768512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130992978936811, + "loss": 1.6598, + "theoretical_loss": 3.867817729265721, + "tokens_seen": 600834048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130892678034102, + "loss": 1.5961, + "theoretical_loss": 3.8677740448487414, + "tokens_seen": 600899584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041307923771313944, + "loss": 1.6248, + "theoretical_loss": 3.867730366529716, + "tokens_seen": 600965120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130692076228686, + "loss": 1.5048, + "theoretical_loss": 3.86768669430713, + "tokens_seen": 601030656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130591775325978, + "loss": 1.5711, + "theoretical_loss": 3.867643028179468, + "tokens_seen": 601096192 + }, + { + "epoch": 0.18, + "learning_rate": 0.000413049147442327, + "loss": 1.5905, + "theoretical_loss": 3.867599368145214, + "tokens_seen": 601161728 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041303911735205616, + "loss": 1.6255, + "theoretical_loss": 3.8675557142028545, + "tokens_seen": 601227264 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041302908726178534, + "loss": 1.5346, + "theoretical_loss": 3.8675120663508755, + "tokens_seen": 601292800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130190571715146, + "loss": 1.6173, + "theoretical_loss": 3.867468424587763, + "tokens_seen": 601358336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004130090270812437, + "loss": 1.668, + "theoretical_loss": 3.867424788912005, + "tokens_seen": 601423872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041299899699097294, + "loss": 1.5773, + "theoretical_loss": 3.8673811593220884, + "tokens_seen": 601489408 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041298896690070207, + "loss": 1.4577, + "theoretical_loss": 3.8673375358165014, + "tokens_seen": 601554944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129789368104313, + "loss": 1.5258, + "theoretical_loss": 3.867293918393734, + "tokens_seen": 601620480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129689067201605, + "loss": 1.4627, + "theoretical_loss": 3.8672503070522746, + "tokens_seen": 601686016 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041295887662988966, + "loss": 1.5673, + "theoretical_loss": 3.8672067017906135, + "tokens_seen": 601751552 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041294884653961884, + "loss": 1.4458, + "theoretical_loss": 3.8671631026072406, + "tokens_seen": 601817088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129388164493481, + "loss": 1.6144, + "theoretical_loss": 3.8671195095006468, + "tokens_seen": 601882624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129287863590772, + "loss": 1.6374, + "theoretical_loss": 3.8670759224693247, + "tokens_seen": 601948160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041291875626880644, + "loss": 1.5093, + "theoretical_loss": 3.867032341511765, + "tokens_seen": 602013696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004129087261785356, + "loss": 1.5422, + "theoretical_loss": 3.866988766626461, + "tokens_seen": 602079232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128986960882648, + "loss": 1.6095, + "theoretical_loss": 3.866945197811906, + "tokens_seen": 602144768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041288866599799404, + "loss": 1.4056, + "theoretical_loss": 3.866901635066593, + "tokens_seen": 602210304 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041287863590772317, + "loss": 1.5838, + "theoretical_loss": 3.8668580783890167, + "tokens_seen": 602275840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128686058174524, + "loss": 1.5198, + "theoretical_loss": 3.866814527777671, + "tokens_seen": 602341376 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041285857572718153, + "loss": 1.5903, + "theoretical_loss": 3.8667709832310524, + "tokens_seen": 602406912 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041284854563691076, + "loss": 1.6393, + "theoretical_loss": 3.866727444747656, + "tokens_seen": 602472448 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041283851554663995, + "loss": 1.4781, + "theoretical_loss": 3.8666839123259784, + "tokens_seen": 602537984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128284854563691, + "loss": 1.508, + "theoretical_loss": 3.8666403859645158, + "tokens_seen": 602603520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128184553660983, + "loss": 1.6183, + "theoretical_loss": 3.8665968656617666, + "tokens_seen": 602669056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004128084252758275, + "loss": 1.4983, + "theoretical_loss": 3.8665533514162282, + "tokens_seen": 602734592 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041279839518555667, + "loss": 1.515, + "theoretical_loss": 3.8665098432263987, + "tokens_seen": 602800128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127883650952859, + "loss": 1.5376, + "theoretical_loss": 3.866466341090778, + "tokens_seen": 602865664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 224736, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4998328685760498, + "objective/train/theoretical_loss": 3.866422845007865, + "objective/train/tokens_used": 623391200, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041277833500501503, + "loss": 1.4957, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041276830491474427, + "loss": 1.5461, + "theoretical_loss": 3.86637935497616, + "tokens_seen": 602996736 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041275827482447345, + "loss": 1.4511, + "theoretical_loss": 3.8663358709941633, + "tokens_seen": 603062272 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041274824473420263, + "loss": 1.493, + "theoretical_loss": 3.8662923930603768, + "tokens_seen": 603127808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127382146439318, + "loss": 1.3544, + "theoretical_loss": 3.8662489211733018, + "tokens_seen": 603193344 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412728184553661, + "loss": 1.6272, + "theoretical_loss": 3.8662054553314404, + "tokens_seen": 603258880 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041271815446339017, + "loss": 1.5349, + "theoretical_loss": 3.866161995533295, + "tokens_seen": 603324416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004127081243731194, + "loss": 1.5545, + "theoretical_loss": 3.86611854177737, + "tokens_seen": 603389952 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041269809428284854, + "loss": 1.5397, + "theoretical_loss": 3.866075094062168, + "tokens_seen": 603455488 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041268806419257777, + "loss": 1.4867, + "theoretical_loss": 3.8660316523861944, + "tokens_seen": 603521024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126780341023069, + "loss": 1.4631, + "theoretical_loss": 3.8659882167479527, + "tokens_seen": 603586560 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041266800401203613, + "loss": 1.5874, + "theoretical_loss": 3.86594478714595, + "tokens_seen": 603652096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126579739217653, + "loss": 1.4546, + "theoretical_loss": 3.8659013635786907, + "tokens_seen": 603717632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126479438314945, + "loss": 1.5814, + "theoretical_loss": 3.8658579460446827, + "tokens_seen": 603783168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126379137412237, + "loss": 1.6103, + "theoretical_loss": 3.865814534542432, + "tokens_seen": 603848704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126278836509529, + "loss": 1.5571, + "theoretical_loss": 3.8657711290704473, + "tokens_seen": 603914240 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041261785356068204, + "loss": 1.6203, + "theoretical_loss": 3.8657277296272348, + "tokens_seen": 603979776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004126078234704113, + "loss": 1.4943, + "theoretical_loss": 3.865684336211305, + "tokens_seen": 604045312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125977933801404, + "loss": 1.5266, + "theoretical_loss": 3.8656409488211656, + "tokens_seen": 604110848 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041258776328986964, + "loss": 1.57, + "theoretical_loss": 3.8655975674553273, + "tokens_seen": 604176384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125777331995988, + "loss": 1.5108, + "theoretical_loss": 3.8655541921123, + "tokens_seen": 604241920 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412567703109328, + "loss": 1.4566, + "theoretical_loss": 3.8655108227905943, + "tokens_seen": 604307456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125576730190572, + "loss": 1.6463, + "theoretical_loss": 3.8654674594887215, + "tokens_seen": 604372992 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041254764292878636, + "loss": 1.6293, + "theoretical_loss": 3.8654241022051927, + "tokens_seen": 604438528 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041253761283851554, + "loss": 1.5728, + "theoretical_loss": 3.865380750938522, + "tokens_seen": 604504064 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125275827482448, + "loss": 1.6034, + "theoretical_loss": 3.86533740568722, + "tokens_seen": 604569600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004125175526579739, + "loss": 1.5707, + "theoretical_loss": 3.8652940664498026, + "tokens_seen": 604635136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041250752256770314, + "loss": 1.5162, + "theoretical_loss": 3.865250733224781, + "tokens_seen": 604700672 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041249749247743227, + "loss": 1.5166, + "theoretical_loss": 3.865207406010672, + "tokens_seen": 604766208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124874623871615, + "loss": 1.5561, + "theoretical_loss": 3.865164084805989, + "tokens_seen": 604831744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124774322968907, + "loss": 1.5542, + "theoretical_loss": 3.865120769609248, + "tokens_seen": 604897280 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041246740220661986, + "loss": 1.5237, + "theoretical_loss": 3.8650774604189655, + "tokens_seen": 604962816 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041245737211634904, + "loss": 1.6143, + "theoretical_loss": 3.8650341572336573, + "tokens_seen": 605028352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124473420260783, + "loss": 1.4153, + "theoretical_loss": 3.8649908600518406, + "tokens_seen": 605093888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004124373119358074, + "loss": 1.5328, + "theoretical_loss": 3.8649475688720334, + "tokens_seen": 605159424 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041242728184553664, + "loss": 1.6768, + "theoretical_loss": 3.864904283692754, + "tokens_seen": 605224960 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041241725175526577, + "loss": 1.5446, + "theoretical_loss": 3.8648610045125196, + "tokens_seen": 605290496 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412407221664995, + "loss": 1.6098, + "theoretical_loss": 3.8648177313298513, + "tokens_seen": 605356032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123971915747242, + "loss": 1.5902, + "theoretical_loss": 3.8647744641432675, + "tokens_seen": 605421568 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041238716148445337, + "loss": 1.5207, + "theoretical_loss": 3.8647312029512886, + "tokens_seen": 605487104 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041237713139418255, + "loss": 1.5066, + "theoretical_loss": 3.864687947752436, + "tokens_seen": 605552640 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041236710130391173, + "loss": 1.6606, + "theoretical_loss": 3.8646446985452307, + "tokens_seen": 605618176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123570712136409, + "loss": 1.5512, + "theoretical_loss": 3.864601455328194, + "tokens_seen": 605683712 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041234704112337015, + "loss": 1.5064, + "theoretical_loss": 3.864558218099849, + "tokens_seen": 605749248 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041233701103309927, + "loss": 1.5335, + "theoretical_loss": 3.864514986858718, + "tokens_seen": 605814784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004123269809428285, + "loss": 1.5685, + "theoretical_loss": 3.8644717616033244, + "tokens_seen": 605880320 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041231695085255763, + "loss": 1.5271, + "theoretical_loss": 3.8644285423321927, + "tokens_seen": 605945856 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041230692076228687, + "loss": 1.4832, + "theoretical_loss": 3.864385329043847, + "tokens_seen": 606011392 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041229689067201605, + "loss": 1.5858, + "theoretical_loss": 3.8643421217368115, + "tokens_seen": 606076928 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041228686058174523, + "loss": 1.456, + "theoretical_loss": 3.8642989204096123, + "tokens_seen": 606142464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 225863, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.529518723487854, + "objective/train/theoretical_loss": 3.8642557250607754, + "objective/train/tokens_used": 626668000, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122768304914744, + "loss": 1.4988, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041226680040120365, + "loss": 1.5859, + "theoretical_loss": 3.8642125356888277, + "tokens_seen": 606273536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122567703109328, + "loss": 1.537, + "theoretical_loss": 3.8641693522922953, + "tokens_seen": 606339072 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412246740220662, + "loss": 1.6126, + "theoretical_loss": 3.8641261748697064, + "tokens_seen": 606404608 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041223671013039114, + "loss": 1.5647, + "theoretical_loss": 3.8640830034195894, + "tokens_seen": 606470144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004122266800401204, + "loss": 1.6295, + "theoretical_loss": 3.864039837940472, + "tokens_seen": 606535680 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041221664994984955, + "loss": 1.5353, + "theoretical_loss": 3.863996678430884, + "tokens_seen": 606601216 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041220661985957874, + "loss": 1.612, + "theoretical_loss": 3.8639535248893555, + "tokens_seen": 606666752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121965897693079, + "loss": 1.4911, + "theoretical_loss": 3.863910377314415, + "tokens_seen": 606732288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121865596790371, + "loss": 1.5708, + "theoretical_loss": 3.863867235704595, + "tokens_seen": 606797824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121765295887663, + "loss": 1.537, + "theoretical_loss": 3.8638241000584257, + "tokens_seen": 606863360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121664994984955, + "loss": 1.4912, + "theoretical_loss": 3.863780970374439, + "tokens_seen": 606928896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121564694082247, + "loss": 1.4059, + "theoretical_loss": 3.863737846651167, + "tokens_seen": 606994432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121464393179539, + "loss": 1.5513, + "theoretical_loss": 3.8636947288871424, + "tokens_seen": 607059968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121364092276831, + "loss": 1.5564, + "theoretical_loss": 3.863651617080899, + "tokens_seen": 607125504 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041212637913741224, + "loss": 1.4525, + "theoretical_loss": 3.8636085112309715, + "tokens_seen": 607191040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121163490471415, + "loss": 1.5242, + "theoretical_loss": 3.863565411335892, + "tokens_seen": 607256576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004121063189568706, + "loss": 1.584, + "theoretical_loss": 3.8635223173941964, + "tokens_seen": 607322112 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041209628886659984, + "loss": 1.5013, + "theoretical_loss": 3.8634792294044202, + "tokens_seen": 607387648 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412086258776329, + "loss": 1.472, + "theoretical_loss": 3.8634361473650993, + "tokens_seen": 607453184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120762286860582, + "loss": 1.6205, + "theoretical_loss": 3.8633930712747704, + "tokens_seen": 607518720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120661985957874, + "loss": 1.5386, + "theoretical_loss": 3.8633500011319692, + "tokens_seen": 607584256 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041205616850551656, + "loss": 1.64, + "theoretical_loss": 3.8633069369352344, + "tokens_seen": 607649792 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041204613841524574, + "loss": 1.5673, + "theoretical_loss": 3.8632638786831035, + "tokens_seen": 607715328 + }, + { + "epoch": 0.18, + "learning_rate": 0.000412036108324975, + "loss": 1.6616, + "theoretical_loss": 3.863220826374115, + "tokens_seen": 607780864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004120260782347041, + "loss": 1.5977, + "theoretical_loss": 3.863177780006808, + "tokens_seen": 607846400 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041201604814443334, + "loss": 1.5865, + "theoretical_loss": 3.8631347395797215, + "tokens_seen": 607911936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041200601805416247, + "loss": 1.5049, + "theoretical_loss": 3.8630917050913958, + "tokens_seen": 607977472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119959879638917, + "loss": 1.5543, + "theoretical_loss": 3.863048676540372, + "tokens_seen": 608043008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119859578736209, + "loss": 1.4178, + "theoretical_loss": 3.8630056539251902, + "tokens_seen": 608108544 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041197592778335006, + "loss": 1.5302, + "theoretical_loss": 3.862962637244393, + "tokens_seen": 608174080 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041196589769307924, + "loss": 1.5276, + "theoretical_loss": 3.862919626496521, + "tokens_seen": 608239616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119558676028085, + "loss": 1.7171, + "theoretical_loss": 3.862876621680119, + "tokens_seen": 608305152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119458375125376, + "loss": 1.5624, + "theoretical_loss": 3.862833622793728, + "tokens_seen": 608370688 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041193580742226684, + "loss": 1.5391, + "theoretical_loss": 3.8627906298358923, + "tokens_seen": 608436224 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041192577733199597, + "loss": 1.4349, + "theoretical_loss": 3.862747642805157, + "tokens_seen": 608501760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119157472417252, + "loss": 1.6364, + "theoretical_loss": 3.8627046617000653, + "tokens_seen": 608567296 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004119057171514544, + "loss": 1.4522, + "theoretical_loss": 3.862661686519163, + "tokens_seen": 608632832 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041189568706118357, + "loss": 1.5338, + "theoretical_loss": 3.862618717260996, + "tokens_seen": 608698368 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041188565697091275, + "loss": 1.517, + "theoretical_loss": 3.8625757539241103, + "tokens_seen": 608763904 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041187562688064193, + "loss": 1.6044, + "theoretical_loss": 3.862532796507053, + "tokens_seen": 608829440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004118655967903711, + "loss": 1.4994, + "theoretical_loss": 3.8624898450083704, + "tokens_seen": 608894976 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041185556670010035, + "loss": 1.4935, + "theoretical_loss": 3.862446899426611, + "tokens_seen": 608960512 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041184553660982947, + "loss": 1.5721, + "theoretical_loss": 3.8624039597603232, + "tokens_seen": 609026048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004118355065195587, + "loss": 1.5691, + "theoretical_loss": 3.862361026008055, + "tokens_seen": 609091584 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041182547642928783, + "loss": 1.4592, + "theoretical_loss": 3.8623180981683563, + "tokens_seen": 609157120 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041181544633901707, + "loss": 1.3912, + "theoretical_loss": 3.8622751762397765, + "tokens_seen": 609222656 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041180541624874625, + "loss": 1.5729, + "theoretical_loss": 3.862232260220866, + "tokens_seen": 609288192 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041179538615847543, + "loss": 1.5127, + "theoretical_loss": 3.8621893501101763, + "tokens_seen": 609353728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117853560682046, + "loss": 1.5558, + "theoretical_loss": 3.8621464459062573, + "tokens_seen": 609419264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 227108, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.483397364616394, + "objective/train/theoretical_loss": 3.8621035476076617, + "objective/train/tokens_used": 629944800, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041177532597793385, + "loss": 1.4896, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.000411765295887663, + "loss": 1.5542, + "theoretical_loss": 3.862060655212942, + "tokens_seen": 609550336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117552657973922, + "loss": 1.4749, + "theoretical_loss": 3.8620177687206514, + "tokens_seen": 609615872 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041174523570712134, + "loss": 1.4884, + "theoretical_loss": 3.861974888129342, + "tokens_seen": 609681408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117352056168506, + "loss": 1.5238, + "theoretical_loss": 3.8619320134375688, + "tokens_seen": 609746944 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041172517552657975, + "loss": 1.5592, + "theoretical_loss": 3.861889144643885, + "tokens_seen": 609812480 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041171514543630894, + "loss": 1.3563, + "theoretical_loss": 3.861846281746847, + "tokens_seen": 609878016 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004117051153460381, + "loss": 1.5397, + "theoretical_loss": 3.861803424745009, + "tokens_seen": 609943552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116950852557673, + "loss": 1.5422, + "theoretical_loss": 3.8617605736369276, + "tokens_seen": 610009088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116850551654965, + "loss": 1.697, + "theoretical_loss": 3.8617177284211595, + "tokens_seen": 610074624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116750250752257, + "loss": 1.4742, + "theoretical_loss": 3.8616748890962604, + "tokens_seen": 610140160 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041166499498495484, + "loss": 1.5222, + "theoretical_loss": 3.8616320556607886, + "tokens_seen": 610205696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116549648946841, + "loss": 1.6333, + "theoretical_loss": 3.861589228113302, + "tokens_seen": 610271232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116449348044132, + "loss": 1.4788, + "theoretical_loss": 3.8615464064523586, + "tokens_seen": 610336768 + }, + { + "epoch": 0.18, + "learning_rate": 0.00041163490471414244, + "loss": 1.5677, + "theoretical_loss": 3.861503590676518, + "tokens_seen": 610402304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0004116248746238716, + "loss": 1.526, + "theoretical_loss": 3.861460780784339, + "tokens_seen": 610467840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004116148445336008, + "loss": 1.6027, + "theoretical_loss": 3.8614179767743826, + "tokens_seen": 610533376 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041160481444333, + "loss": 1.4132, + "theoretical_loss": 3.861375178645208, + "tokens_seen": 610598912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115947843530592, + "loss": 1.6569, + "theoretical_loss": 3.861332386395377, + "tokens_seen": 610664448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041158475426278834, + "loss": 1.5329, + "theoretical_loss": 3.8612896000234516, + "tokens_seen": 610729984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115747241725176, + "loss": 1.5276, + "theoretical_loss": 3.861246819527992, + "tokens_seen": 610795520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115646940822467, + "loss": 1.57, + "theoretical_loss": 3.861204044907563, + "tokens_seen": 610861056 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041155466399197594, + "loss": 1.6136, + "theoretical_loss": 3.8611612761607255, + "tokens_seen": 610926592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115446339017051, + "loss": 1.4352, + "theoretical_loss": 3.861118513286044, + "tokens_seen": 610992128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115346038114343, + "loss": 1.4986, + "theoretical_loss": 3.8610757562820828, + "tokens_seen": 611057664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004115245737211635, + "loss": 1.4835, + "theoretical_loss": 3.861033005147406, + "tokens_seen": 611123200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041151454363089267, + "loss": 1.4645, + "theoretical_loss": 3.8609902598805785, + "tokens_seen": 611188736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041150451354062185, + "loss": 1.4418, + "theoretical_loss": 3.860947520480167, + "tokens_seen": 611254272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114944834503511, + "loss": 1.5001, + "theoretical_loss": 3.8609047869447357, + "tokens_seen": 611319808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114844533600802, + "loss": 1.54, + "theoretical_loss": 3.8608620592728524, + "tokens_seen": 611385344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041147442326980944, + "loss": 1.5299, + "theoretical_loss": 3.860819337463084, + "tokens_seen": 611450880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041146439317953857, + "loss": 1.6165, + "theoretical_loss": 3.860776621513998, + "tokens_seen": 611516416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114543630892678, + "loss": 1.5986, + "theoretical_loss": 3.8607339114241617, + "tokens_seen": 611581952 + }, + { + "epoch": 0.19, + "learning_rate": 0.000411444332998997, + "loss": 1.5593, + "theoretical_loss": 3.860691207192145, + "tokens_seen": 611647488 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041143430290872617, + "loss": 1.4777, + "theoretical_loss": 3.860648508816516, + "tokens_seen": 611713024 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041142427281845535, + "loss": 1.4023, + "theoretical_loss": 3.8606058162958443, + "tokens_seen": 611778560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004114142427281846, + "loss": 1.4435, + "theoretical_loss": 3.8605631296287006, + "tokens_seen": 611844096 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041140421263791377, + "loss": 1.4885, + "theoretical_loss": 3.860520448813655, + "tokens_seen": 611909632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041139418254764295, + "loss": 1.6235, + "theoretical_loss": 3.8604777738492784, + "tokens_seen": 611975168 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041138415245737213, + "loss": 1.5142, + "theoretical_loss": 3.860435104734143, + "tokens_seen": 612040704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004113741223671013, + "loss": 1.6285, + "theoretical_loss": 3.86039244146682, + "tokens_seen": 612106240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041136409227683055, + "loss": 1.4043, + "theoretical_loss": 3.860349784045883, + "tokens_seen": 612171776 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041135406218655967, + "loss": 1.4691, + "theoretical_loss": 3.8603071324699045, + "tokens_seen": 612237312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004113440320962889, + "loss": 1.4928, + "theoretical_loss": 3.8602644867374583, + "tokens_seen": 612302848 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041133400200601803, + "loss": 1.5582, + "theoretical_loss": 3.860221846847118, + "tokens_seen": 612368384 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041132397191574727, + "loss": 1.6094, + "theoretical_loss": 3.8601792127974583, + "tokens_seen": 612433920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041131394182547645, + "loss": 1.3304, + "theoretical_loss": 3.860136584587055, + "tokens_seen": 612499456 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041130391173520563, + "loss": 1.5085, + "theoretical_loss": 3.860093962214483, + "tokens_seen": 612564992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112938816449348, + "loss": 1.443, + "theoretical_loss": 3.860051345678319, + "tokens_seen": 612630528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041128385155466405, + "loss": 1.5676, + "theoretical_loss": 3.860008734977139, + "tokens_seen": 612696064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 228263, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6966383457183838, + "objective/train/theoretical_loss": 3.8599661301095196, + "objective/train/tokens_used": 633221600, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112738214643932, + "loss": 1.6119, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112637913741224, + "loss": 1.4303, + "theoretical_loss": 3.85992353107404, + "tokens_seen": 612827136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041125376128385154, + "loss": 1.4846, + "theoretical_loss": 3.8598809378692764, + "tokens_seen": 612892672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112437311935808, + "loss": 1.4249, + "theoretical_loss": 3.859838350493809, + "tokens_seen": 612958208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041123370110330995, + "loss": 1.5315, + "theoretical_loss": 3.8597957689462157, + "tokens_seen": 613023744 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041122367101303914, + "loss": 1.5616, + "theoretical_loss": 3.859753193225077, + "tokens_seen": 613089280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112136409227683, + "loss": 1.3509, + "theoretical_loss": 3.859710623328972, + "tokens_seen": 613154816 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004112036108324975, + "loss": 1.5367, + "theoretical_loss": 3.8596680592564816, + "tokens_seen": 613220352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111935807422267, + "loss": 1.5954, + "theoretical_loss": 3.859625501006188, + "tokens_seen": 613285888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111835506519559, + "loss": 1.55, + "theoretical_loss": 3.859582948576671, + "tokens_seen": 613351424 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041117352056168504, + "loss": 1.5641, + "theoretical_loss": 3.8595404019665143, + "tokens_seen": 613416960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111634904714143, + "loss": 1.4884, + "theoretical_loss": 3.859497861174299, + "tokens_seen": 613482496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111534603811434, + "loss": 1.5545, + "theoretical_loss": 3.859455326198609, + "tokens_seen": 613548032 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041114343029087264, + "loss": 1.5605, + "theoretical_loss": 3.8594127970380274, + "tokens_seen": 613613568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111334002006018, + "loss": 1.4533, + "theoretical_loss": 3.859370273691139, + "tokens_seen": 613679104 + }, + { + "epoch": 0.19, + "learning_rate": 0.000411123370110331, + "loss": 1.4172, + "theoretical_loss": 3.8593277561565276, + "tokens_seen": 613744640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111133400200602, + "loss": 1.5242, + "theoretical_loss": 3.8592852444327788, + "tokens_seen": 613810176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004111033099297894, + "loss": 1.4438, + "theoretical_loss": 3.8592427385184775, + "tokens_seen": 613875712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041109327983951854, + "loss": 1.5658, + "theoretical_loss": 3.85920023841221, + "tokens_seen": 613941248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110832497492478, + "loss": 1.4799, + "theoretical_loss": 3.859157744112563, + "tokens_seen": 614006784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110732196589769, + "loss": 1.5555, + "theoretical_loss": 3.859115255618123, + "tokens_seen": 614072320 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041106318956870614, + "loss": 1.5405, + "theoretical_loss": 3.8590727729274787, + "tokens_seen": 614137856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110531594784353, + "loss": 1.417, + "theoretical_loss": 3.859030296039217, + "tokens_seen": 614203392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110431293881645, + "loss": 1.6403, + "theoretical_loss": 3.8589878249519263, + "tokens_seen": 614268928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110330992978937, + "loss": 1.5458, + "theoretical_loss": 3.8589453596641965, + "tokens_seen": 614334464 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041102306920762287, + "loss": 1.4988, + "theoretical_loss": 3.858902900174617, + "tokens_seen": 614400000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041101303911735205, + "loss": 1.3992, + "theoretical_loss": 3.858860446481777, + "tokens_seen": 614465536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004110030090270813, + "loss": 1.5478, + "theoretical_loss": 3.8588179985842674, + "tokens_seen": 614531072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109929789368104, + "loss": 1.4053, + "theoretical_loss": 3.8587755564806794, + "tokens_seen": 614596608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041098294884653965, + "loss": 1.5611, + "theoretical_loss": 3.8587331201696045, + "tokens_seen": 614662144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041097291875626877, + "loss": 1.4972, + "theoretical_loss": 3.8586906896496336, + "tokens_seen": 614727680 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410962888665998, + "loss": 1.4472, + "theoretical_loss": 3.8586482649193607, + "tokens_seen": 614793216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109528585757272, + "loss": 1.4133, + "theoretical_loss": 3.8586058459773778, + "tokens_seen": 614858752 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041094282848545637, + "loss": 1.4299, + "theoretical_loss": 3.8585634328222778, + "tokens_seen": 614924288 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041093279839518555, + "loss": 1.4911, + "theoretical_loss": 3.8585210254526565, + "tokens_seen": 614989824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109227683049148, + "loss": 1.4219, + "theoretical_loss": 3.8584786238671063, + "tokens_seen": 615055360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004109127382146439, + "loss": 1.5437, + "theoretical_loss": 3.8584362280642233, + "tokens_seen": 615120896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041090270812437315, + "loss": 1.4964, + "theoretical_loss": 3.858393838042603, + "tokens_seen": 615186432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108926780341023, + "loss": 1.5527, + "theoretical_loss": 3.8583514538008403, + "tokens_seen": 615251968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108826479438315, + "loss": 1.4608, + "theoretical_loss": 3.858309075337532, + "tokens_seen": 615317504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108726178535607, + "loss": 1.3874, + "theoretical_loss": 3.8582667026512754, + "tokens_seen": 615383040 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041086258776328987, + "loss": 1.464, + "theoretical_loss": 3.858224335740668, + "tokens_seen": 615448576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041085255767301905, + "loss": 1.5295, + "theoretical_loss": 3.8581819746043067, + "tokens_seen": 615514112 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041084252758274823, + "loss": 1.5858, + "theoretical_loss": 3.8581396192407906, + "tokens_seen": 615579648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108324974924774, + "loss": 1.4763, + "theoretical_loss": 3.8580972696487184, + "tokens_seen": 615645184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041082246740220665, + "loss": 1.4612, + "theoretical_loss": 3.858054925826689, + "tokens_seen": 615710720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004108124373119358, + "loss": 1.5831, + "theoretical_loss": 3.8580125877733025, + "tokens_seen": 615776256 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410802407221665, + "loss": 1.4487, + "theoretical_loss": 3.8579702554871593, + "tokens_seen": 615841792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107923771313942, + "loss": 1.4158, + "theoretical_loss": 3.8579279289668604, + "tokens_seen": 615907328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107823470411234, + "loss": 1.5008, + "theoretical_loss": 3.8578856082110065, + "tokens_seen": 615972864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 229528, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4698405265808105, + "objective/train/theoretical_loss": 3.8578432932182, + "objective/train/tokens_used": 636498400, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041077231695085256, + "loss": 1.5329, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041076228686058174, + "loss": 1.4912, + "theoretical_loss": 3.8578009839870426, + "tokens_seen": 616103936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107522567703109, + "loss": 1.5182, + "theoretical_loss": 3.8577586805161372, + "tokens_seen": 616169472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041074222668004015, + "loss": 1.4818, + "theoretical_loss": 3.857716382804087, + "tokens_seen": 616235008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107321965897693, + "loss": 1.5309, + "theoretical_loss": 3.8576740908494958, + "tokens_seen": 616300544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107221664994985, + "loss": 1.5905, + "theoretical_loss": 3.8576318046509677, + "tokens_seen": 616366080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041071213640922764, + "loss": 1.4615, + "theoretical_loss": 3.857589524207108, + "tokens_seen": 616431616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004107021063189569, + "loss": 1.4804, + "theoretical_loss": 3.8575472495165206, + "tokens_seen": 616497152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041069207622868606, + "loss": 1.5294, + "theoretical_loss": 3.8575049805778123, + "tokens_seen": 616562688 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041068204613841524, + "loss": 1.4832, + "theoretical_loss": 3.8574627173895886, + "tokens_seen": 616628224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106720160481444, + "loss": 1.6258, + "theoretical_loss": 3.8574204599504567, + "tokens_seen": 616693760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106619859578736, + "loss": 1.3455, + "theoretical_loss": 3.857378208259023, + "tokens_seen": 616759296 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041065195586760284, + "loss": 1.4487, + "theoretical_loss": 3.8573359623138956, + "tokens_seen": 616824832 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410641925777332, + "loss": 1.6164, + "theoretical_loss": 3.857293722113683, + "tokens_seen": 616890368 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106318956870612, + "loss": 1.6785, + "theoretical_loss": 3.8572514876569923, + "tokens_seen": 616955904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106218655967904, + "loss": 1.5273, + "theoretical_loss": 3.857209258942434, + "tokens_seen": 617021440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004106118355065196, + "loss": 1.5283, + "theoretical_loss": 3.857167035968617, + "tokens_seen": 617086976 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041060180541624874, + "loss": 1.6064, + "theoretical_loss": 3.8571248187341522, + "tokens_seen": 617152512 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410591775325978, + "loss": 1.475, + "theoretical_loss": 3.8570826072376487, + "tokens_seen": 617218048 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105817452357071, + "loss": 1.5605, + "theoretical_loss": 3.8570404014777178, + "tokens_seen": 617283584 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041057171514543634, + "loss": 1.5256, + "theoretical_loss": 3.8569982014529716, + "tokens_seen": 617349120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105616850551655, + "loss": 1.5259, + "theoretical_loss": 3.8569560071620224, + "tokens_seen": 617414656 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105516549648947, + "loss": 1.4737, + "theoretical_loss": 3.8569138186034815, + "tokens_seen": 617480192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105416248746239, + "loss": 1.5335, + "theoretical_loss": 3.8568716357759625, + "tokens_seen": 617545728 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041053159478435307, + "loss": 1.6339, + "theoretical_loss": 3.856829458678079, + "tokens_seen": 617611264 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041052156469408225, + "loss": 1.5064, + "theoretical_loss": 3.856787287308444, + "tokens_seen": 617676800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105115346038115, + "loss": 1.5642, + "theoretical_loss": 3.8567451216656723, + "tokens_seen": 617742336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004105015045135406, + "loss": 1.5622, + "theoretical_loss": 3.8567029617483795, + "tokens_seen": 617807872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041049147442326985, + "loss": 1.4261, + "theoretical_loss": 3.8566608075551803, + "tokens_seen": 617873408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041048144433299897, + "loss": 1.5603, + "theoretical_loss": 3.8566186590846905, + "tokens_seen": 617938944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104714142427282, + "loss": 1.5577, + "theoretical_loss": 3.856576516335526, + "tokens_seen": 618004480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104613841524574, + "loss": 1.5365, + "theoretical_loss": 3.8565343793063045, + "tokens_seen": 618070016 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041045135406218657, + "loss": 1.5905, + "theoretical_loss": 3.856492247995643, + "tokens_seen": 618135552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041044132397191575, + "loss": 1.4846, + "theoretical_loss": 3.856450122402159, + "tokens_seen": 618201088 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410431293881645, + "loss": 1.6593, + "theoretical_loss": 3.8564080025244705, + "tokens_seen": 618266624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104212637913741, + "loss": 1.5387, + "theoretical_loss": 3.8563658883611964, + "tokens_seen": 618332160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041041123370110335, + "loss": 1.4755, + "theoretical_loss": 3.856323779910956, + "tokens_seen": 618397696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004104012036108325, + "loss": 1.6108, + "theoretical_loss": 3.856281677172369, + "tokens_seen": 618463232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103911735205617, + "loss": 1.4725, + "theoretical_loss": 3.856239580144056, + "tokens_seen": 618528768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103811434302909, + "loss": 1.4606, + "theoretical_loss": 3.8561974888246366, + "tokens_seen": 618594304 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041037111334002007, + "loss": 1.5525, + "theoretical_loss": 3.8561554032127323, + "tokens_seen": 618659840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041036108324974925, + "loss": 1.4933, + "theoretical_loss": 3.856113323306965, + "tokens_seen": 618725376 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041035105315947844, + "loss": 1.5525, + "theoretical_loss": 3.8560712491059563, + "tokens_seen": 618790912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103410230692076, + "loss": 1.3763, + "theoretical_loss": 3.8560291806083296, + "tokens_seen": 618856448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041033099297893685, + "loss": 1.477, + "theoretical_loss": 3.855987117812707, + "tokens_seen": 618921984 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410320962888666, + "loss": 1.609, + "theoretical_loss": 3.8559450607177124, + "tokens_seen": 618987520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103109327983952, + "loss": 1.5747, + "theoretical_loss": 3.8559030093219695, + "tokens_seen": 619053056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004103009027081244, + "loss": 1.5916, + "theoretical_loss": 3.855860963624103, + "tokens_seen": 619118592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102908726178536, + "loss": 1.5297, + "theoretical_loss": 3.855818923622738, + "tokens_seen": 619184128 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041028084252758276, + "loss": 1.4639, + "theoretical_loss": 3.855776889316499, + "tokens_seen": 619249664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 230719, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6862338781356812, + "objective/train/theoretical_loss": 3.855734860704014, + "objective/train/tokens_used": 639775200, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041027081243731194, + "loss": 1.6374, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102607823470411, + "loss": 1.5161, + "theoretical_loss": 3.8556928377839066, + "tokens_seen": 619380736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041025075225677035, + "loss": 1.6367, + "theoretical_loss": 3.8556508205548057, + "tokens_seen": 619446272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102407221664995, + "loss": 1.6394, + "theoretical_loss": 3.8556088090153375, + "tokens_seen": 619511808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102306920762287, + "loss": 1.535, + "theoretical_loss": 3.8555668031641304, + "tokens_seen": 619577344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041022066198595784, + "loss": 1.6068, + "theoretical_loss": 3.8555248029998124, + "tokens_seen": 619642880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004102106318956871, + "loss": 1.4317, + "theoretical_loss": 3.855482808521012, + "tokens_seen": 619708416 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041020060180541626, + "loss": 1.5066, + "theoretical_loss": 3.855440819726359, + "tokens_seen": 619773952 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041019057171514544, + "loss": 1.562, + "theoretical_loss": 3.8553988366144827, + "tokens_seen": 619839488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101805416248746, + "loss": 1.5802, + "theoretical_loss": 3.8553568591840133, + "tokens_seen": 619905024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101705115346038, + "loss": 1.5647, + "theoretical_loss": 3.8553148874335816, + "tokens_seen": 619970560 + }, + { + "epoch": 0.19, + "learning_rate": 0.000410160481444333, + "loss": 1.5439, + "theoretical_loss": 3.855272921361819, + "tokens_seen": 620036096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101504513540622, + "loss": 1.5867, + "theoretical_loss": 3.855230960967356, + "tokens_seen": 620101632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041014042126379135, + "loss": 1.5529, + "theoretical_loss": 3.8551890062488265, + "tokens_seen": 620167168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101303911735206, + "loss": 1.5378, + "theoretical_loss": 3.855147057204861, + "tokens_seen": 620232704 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041012036108324976, + "loss": 1.5493, + "theoretical_loss": 3.8551051138340937, + "tokens_seen": 620298240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041011033099297894, + "loss": 1.5876, + "theoretical_loss": 3.8550631761351575, + "tokens_seen": 620363776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004101003009027081, + "loss": 1.5709, + "theoretical_loss": 3.8550212441066876, + "tokens_seen": 620429312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100902708124373, + "loss": 1.4909, + "theoretical_loss": 3.8549793177473166, + "tokens_seen": 620494848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100802407221665, + "loss": 1.5356, + "theoretical_loss": 3.8549373970556813, + "tokens_seen": 620560384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100702106318957, + "loss": 1.4671, + "theoretical_loss": 3.8548954820304155, + "tokens_seen": 620625920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041006018054162485, + "loss": 1.542, + "theoretical_loss": 3.854853572670156, + "tokens_seen": 620691456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100501504513541, + "loss": 1.4723, + "theoretical_loss": 3.8548116689735386, + "tokens_seen": 620756992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100401203610832, + "loss": 1.4777, + "theoretical_loss": 3.8547697709392006, + "tokens_seen": 620822528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041003009027081245, + "loss": 1.5717, + "theoretical_loss": 3.854727878565779, + "tokens_seen": 620888064 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041002006018054163, + "loss": 1.6342, + "theoretical_loss": 3.8546859918519107, + "tokens_seen": 620953600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004100100300902708, + "loss": 1.5132, + "theoretical_loss": 3.8546441107962357, + "tokens_seen": 621019136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00041, + "loss": 1.6414, + "theoretical_loss": 3.854602235397391, + "tokens_seen": 621084672 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040998996990972917, + "loss": 1.5102, + "theoretical_loss": 3.8545603656540175, + "tokens_seen": 621150208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040997993981945835, + "loss": 1.4873, + "theoretical_loss": 3.854518501564753, + "tokens_seen": 621215744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099699097291876, + "loss": 1.4934, + "theoretical_loss": 3.8544766431282387, + "tokens_seen": 621281280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099598796389167, + "loss": 1.56, + "theoretical_loss": 3.854434790343115, + "tokens_seen": 621346816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040994984954864595, + "loss": 1.361, + "theoretical_loss": 3.8543929432080226, + "tokens_seen": 621412352 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040993981945837513, + "loss": 1.5659, + "theoretical_loss": 3.8543511017216034, + "tokens_seen": 621477888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099297893681043, + "loss": 1.4709, + "theoretical_loss": 3.8543092658825, + "tokens_seen": 621543424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099197592778335, + "loss": 1.5116, + "theoretical_loss": 3.8542674356893527, + "tokens_seen": 621608960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004099097291875627, + "loss": 1.4868, + "theoretical_loss": 3.8542256111408064, + "tokens_seen": 621674496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098996990972919, + "loss": 1.5327, + "theoretical_loss": 3.854183792235504, + "tokens_seen": 621740032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098896690070211, + "loss": 1.4494, + "theoretical_loss": 3.8541419789720894, + "tokens_seen": 621805568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098796389167503, + "loss": 1.5225, + "theoretical_loss": 3.854100171349206, + "tokens_seen": 621871104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040986960882647945, + "loss": 1.6273, + "theoretical_loss": 3.8540583693655, + "tokens_seen": 621936640 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040985957873620864, + "loss": 1.5218, + "theoretical_loss": 3.854016573019616, + "tokens_seen": 622002176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098495486459378, + "loss": 1.5631, + "theoretical_loss": 3.8539747823101997, + "tokens_seen": 622067712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040983951855566705, + "loss": 1.5025, + "theoretical_loss": 3.853932997235897, + "tokens_seen": 622133248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098294884653962, + "loss": 1.5727, + "theoretical_loss": 3.8538912177953555, + "tokens_seen": 622198784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098194583751254, + "loss": 1.4604, + "theoretical_loss": 3.8538494439872215, + "tokens_seen": 622264320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004098094282848546, + "loss": 1.397, + "theoretical_loss": 3.8538076758101427, + "tokens_seen": 622329856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097993981945838, + "loss": 1.6526, + "theoretical_loss": 3.8537659132627677, + "tokens_seen": 622395392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040978936810431296, + "loss": 1.6309, + "theoretical_loss": 3.853724156343744, + "tokens_seen": 622460928 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040977933801404214, + "loss": 1.5423, + "theoretical_loss": 3.853682405051722, + "tokens_seen": 622526464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 231894, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6124556064605713, + "objective/train/theoretical_loss": 3.8536406593853503, + "objective/train/tokens_used": 643052000, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097693079237713, + "loss": 1.5513, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040975927783350055, + "loss": 1.5439, + "theoretical_loss": 3.8535989193432787, + "tokens_seen": 622657536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097492477432297, + "loss": 1.592, + "theoretical_loss": 3.8535571849241577, + "tokens_seen": 622723072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097392176529589, + "loss": 1.4164, + "theoretical_loss": 3.8535154561266385, + "tokens_seen": 622788608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040972918756268804, + "loss": 1.5064, + "theoretical_loss": 3.8534737329493725, + "tokens_seen": 622854144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004097191574724173, + "loss": 1.5385, + "theoretical_loss": 3.853432015391011, + "tokens_seen": 622919680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040970912738214646, + "loss": 1.4276, + "theoretical_loss": 3.8533903034502064, + "tokens_seen": 622985216 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040969909729187564, + "loss": 1.5278, + "theoretical_loss": 3.8533485971256125, + "tokens_seen": 623050752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096890672016048, + "loss": 1.5162, + "theoretical_loss": 3.8533068964158805, + "tokens_seen": 623116288 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409679037111334, + "loss": 1.4838, + "theoretical_loss": 3.8532652013196653, + "tokens_seen": 623181824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096690070210632, + "loss": 1.732, + "theoretical_loss": 3.853223511835621, + "tokens_seen": 623247360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096589769307924, + "loss": 1.5047, + "theoretical_loss": 3.853181827962402, + "tokens_seen": 623312896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040964894684052155, + "loss": 1.5093, + "theoretical_loss": 3.853140149698664, + "tokens_seen": 623378432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096389167502508, + "loss": 1.4967, + "theoretical_loss": 3.8530984770430607, + "tokens_seen": 623443968 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040962888665997996, + "loss": 1.4751, + "theoretical_loss": 3.85305680999425, + "tokens_seen": 623509504 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040961885656970914, + "loss": 1.5119, + "theoretical_loss": 3.8530151485508872, + "tokens_seen": 623575040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004096088264794383, + "loss": 1.5748, + "theoretical_loss": 3.85297349271163, + "tokens_seen": 623640576 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095987963891675, + "loss": 1.5259, + "theoretical_loss": 3.852931842475135, + "tokens_seen": 623706112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095887662988967, + "loss": 1.4692, + "theoretical_loss": 3.8528901978400603, + "tokens_seen": 623771648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095787362086259, + "loss": 1.5154, + "theoretical_loss": 3.852848558805065, + "tokens_seen": 623837184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040956870611835505, + "loss": 1.4865, + "theoretical_loss": 3.8528069253688066, + "tokens_seen": 623902720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095586760280843, + "loss": 1.4494, + "theoretical_loss": 3.852765297529945, + "tokens_seen": 623968256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095486459378134, + "loss": 1.608, + "theoretical_loss": 3.8527236752871397, + "tokens_seen": 624033792 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040953861584754265, + "loss": 1.609, + "theoretical_loss": 3.8526820586390507, + "tokens_seen": 624099328 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040952858575727183, + "loss": 1.5279, + "theoretical_loss": 3.8526404475843385, + "tokens_seen": 624164864 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409518555667001, + "loss": 1.6099, + "theoretical_loss": 3.8525988421216653, + "tokens_seen": 624230400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004095085255767302, + "loss": 1.4701, + "theoretical_loss": 3.852557242249691, + "tokens_seen": 624295936 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040949849548645937, + "loss": 1.6442, + "theoretical_loss": 3.8525156479670786, + "tokens_seen": 624361472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040948846539618855, + "loss": 1.564, + "theoretical_loss": 3.8524740592724904, + "tokens_seen": 624427008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094784353059178, + "loss": 1.6782, + "theoretical_loss": 3.8524324761645894, + "tokens_seen": 624492544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094684052156469, + "loss": 1.4992, + "theoretical_loss": 3.8523908986420383, + "tokens_seen": 624558080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040945837512537615, + "loss": 1.5223, + "theoretical_loss": 3.8523493267035014, + "tokens_seen": 624623616 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040944834503510533, + "loss": 1.4754, + "theoretical_loss": 3.852307760347643, + "tokens_seen": 624689152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094383149448345, + "loss": 1.5367, + "theoretical_loss": 3.852266199573128, + "tokens_seen": 624754688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094282848545637, + "loss": 1.5887, + "theoretical_loss": 3.852224644378621, + "tokens_seen": 624820224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004094182547642929, + "loss": 1.5494, + "theoretical_loss": 3.852183094762789, + "tokens_seen": 624885760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040940822467402206, + "loss": 1.5694, + "theoretical_loss": 3.852141550724297, + "tokens_seen": 624951296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093981945837513, + "loss": 1.3718, + "theoretical_loss": 3.852100012261811, + "tokens_seen": 625016832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093881644934804, + "loss": 1.4438, + "theoretical_loss": 3.8520584793739996, + "tokens_seen": 625082368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040937813440320965, + "loss": 1.6937, + "theoretical_loss": 3.852016952059529, + "tokens_seen": 625147904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093681043129388, + "loss": 1.5037, + "theoretical_loss": 3.8519754303170677, + "tokens_seen": 625213440 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409358074222668, + "loss": 1.5657, + "theoretical_loss": 3.8519339141452846, + "tokens_seen": 625278976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093480441323972, + "loss": 1.4864, + "theoretical_loss": 3.8518924035428475, + "tokens_seen": 625344512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093380140421264, + "loss": 1.4708, + "theoretical_loss": 3.8518508985084265, + "tokens_seen": 625410048 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040932798395185556, + "loss": 1.5378, + "theoretical_loss": 3.8518093990406914, + "tokens_seen": 625475584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093179538615848, + "loss": 1.5047, + "theoretical_loss": 3.8517679051383116, + "tokens_seen": 625541120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004093079237713139, + "loss": 1.5606, + "theoretical_loss": 3.851726416799959, + "tokens_seen": 625606656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040929789368104316, + "loss": 1.4873, + "theoretical_loss": 3.8516849340243042, + "tokens_seen": 625672192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092878635907723, + "loss": 1.5066, + "theoretical_loss": 3.8516434568100184, + "tokens_seen": 625737728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092778335005015, + "loss": 1.6046, + "theoretical_loss": 3.8516019851557743, + "tokens_seen": 625803264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 233149, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4033749103546143, + "objective/train/theoretical_loss": 3.8515605190602447, + "objective/train/tokens_used": 646328800, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092678034102307, + "loss": 1.4947, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092577733199599, + "loss": 1.4407, + "theoretical_loss": 3.8515190585221015, + "tokens_seen": 625934336 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040924774322968906, + "loss": 1.5307, + "theoretical_loss": 3.851477603540019, + "tokens_seen": 625999872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040923771313941824, + "loss": 1.5423, + "theoretical_loss": 3.8514361541126703, + "tokens_seen": 626065408 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092276830491474, + "loss": 1.5555, + "theoretical_loss": 3.85139471023873, + "tokens_seen": 626130944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040921765295887666, + "loss": 1.6246, + "theoretical_loss": 3.8513532719168744, + "tokens_seen": 626196480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004092076228686058, + "loss": 1.5544, + "theoretical_loss": 3.851311839145777, + "tokens_seen": 626262016 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409197592778335, + "loss": 1.5222, + "theoretical_loss": 3.851270411924114, + "tokens_seen": 626327552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040918756268806415, + "loss": 1.6214, + "theoretical_loss": 3.8512289902505605, + "tokens_seen": 626393088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091775325977934, + "loss": 1.4353, + "theoretical_loss": 3.8511875741237955, + "tokens_seen": 626458624 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040916750250752257, + "loss": 1.6915, + "theoretical_loss": 3.851146163542495, + "tokens_seen": 626524160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040915747241725175, + "loss": 1.4728, + "theoretical_loss": 3.8511047585053353, + "tokens_seen": 626589696 + }, + { + "epoch": 0.19, + "learning_rate": 0.000409147442326981, + "loss": 1.6507, + "theoretical_loss": 3.851063359010996, + "tokens_seen": 626655232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040913741223671016, + "loss": 1.4618, + "theoretical_loss": 3.851021965058155, + "tokens_seen": 626720768 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040912738214643934, + "loss": 1.4648, + "theoretical_loss": 3.850980576645491, + "tokens_seen": 626786304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091173520561685, + "loss": 1.452, + "theoretical_loss": 3.8509391937716835, + "tokens_seen": 626851840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004091073219658977, + "loss": 1.5453, + "theoretical_loss": 3.8508978164354124, + "tokens_seen": 626917376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090972918756269, + "loss": 1.5244, + "theoretical_loss": 3.8508564446353573, + "tokens_seen": 626982912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090872617853561, + "loss": 1.6728, + "theoretical_loss": 3.8508150783702, + "tokens_seen": 627048448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040907723169508525, + "loss": 1.4743, + "theoretical_loss": 3.850773717638621, + "tokens_seen": 627113984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090672016048145, + "loss": 1.5988, + "theoretical_loss": 3.850732362439302, + "tokens_seen": 627179520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090571715145436, + "loss": 1.6318, + "theoretical_loss": 3.850691012770925, + "tokens_seen": 627245056 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040904714142427285, + "loss": 1.5136, + "theoretical_loss": 3.8506496686321725, + "tokens_seen": 627310592 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040903711133400203, + "loss": 1.5752, + "theoretical_loss": 3.8506083300217275, + "tokens_seen": 627376128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090270812437312, + "loss": 1.5358, + "theoretical_loss": 3.850566996938274, + "tokens_seen": 627441664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004090170511534604, + "loss": 1.6553, + "theoretical_loss": 3.8505256693804943, + "tokens_seen": 627507200 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040900702106318957, + "loss": 1.5308, + "theoretical_loss": 3.8504843473470745, + "tokens_seen": 627572736 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040899699097291875, + "loss": 1.6318, + "theoretical_loss": 3.850443030836698, + "tokens_seen": 627638272 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408986960882648, + "loss": 1.433, + "theoretical_loss": 3.850401719848051, + "tokens_seen": 627703808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089769307923771, + "loss": 1.4843, + "theoretical_loss": 3.8503604143798187, + "tokens_seen": 627769344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040896690070210635, + "loss": 1.5792, + "theoretical_loss": 3.850319114430687, + "tokens_seen": 627834880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040895687061183553, + "loss": 1.4692, + "theoretical_loss": 3.8502778199993433, + "tokens_seen": 627900416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089468405215647, + "loss": 1.5674, + "theoretical_loss": 3.850236531084473, + "tokens_seen": 627965952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089368104312939, + "loss": 1.4369, + "theoretical_loss": 3.8501952476847654, + "tokens_seen": 628031488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089267803410231, + "loss": 1.553, + "theoretical_loss": 3.8501539697989076, + "tokens_seen": 628097024 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040891675025075226, + "loss": 1.6535, + "theoretical_loss": 3.8501126974255877, + "tokens_seen": 628162560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004089067201604815, + "loss": 1.591, + "theoretical_loss": 3.850071430563495, + "tokens_seen": 628228096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088966900702106, + "loss": 1.528, + "theoretical_loss": 3.850030169211319, + "tokens_seen": 628293632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040888665997993985, + "loss": 1.5608, + "theoretical_loss": 3.8499889133677483, + "tokens_seen": 628359168 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408876629889669, + "loss": 1.4439, + "theoretical_loss": 3.8499476630314735, + "tokens_seen": 628424704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088665997993982, + "loss": 1.5547, + "theoretical_loss": 3.849906418201186, + "tokens_seen": 628490240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088565697091274, + "loss": 1.5379, + "theoretical_loss": 3.849865178875576, + "tokens_seen": 628555776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088465396188566, + "loss": 1.6028, + "theoretical_loss": 3.8498239450533354, + "tokens_seen": 628621312 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040883650952858576, + "loss": 1.4615, + "theoretical_loss": 3.8497827167331558, + "tokens_seen": 628686848 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408826479438315, + "loss": 1.4297, + "theoretical_loss": 3.84974149391373, + "tokens_seen": 628752384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004088164493480441, + "loss": 1.4442, + "theoretical_loss": 3.849700276593751, + "tokens_seen": 628817920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040880641925777336, + "loss": 1.4824, + "theoretical_loss": 3.8496590647719113, + "tokens_seen": 628883456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087963891675025, + "loss": 1.6137, + "theoretical_loss": 3.849617858446905, + "tokens_seen": 628948992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087863590772317, + "loss": 1.5795, + "theoretical_loss": 3.8495766576174266, + "tokens_seen": 629014528 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087763289869609, + "loss": 1.5533, + "theoretical_loss": 3.8495354622821703, + "tokens_seen": 629080064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 234355, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4306466579437256, + "objective/train/theoretical_loss": 3.8494942724398316, + "objective/train/tokens_used": 649605600, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087662988966901, + "loss": 1.452, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040875626880641926, + "loss": 1.5106, + "theoretical_loss": 3.8494530880891054, + "tokens_seen": 629211136 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040874623871614844, + "loss": 1.5596, + "theoretical_loss": 3.8494119092286887, + "tokens_seen": 629276672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087362086258776, + "loss": 1.5886, + "theoretical_loss": 3.849370735857277, + "tokens_seen": 629342208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040872617853560686, + "loss": 1.4917, + "theoretical_loss": 3.849329567973567, + "tokens_seen": 629407744 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408716148445336, + "loss": 1.6057, + "theoretical_loss": 3.8492884055762566, + "tokens_seen": 629473280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004087061183550652, + "loss": 1.4506, + "theoretical_loss": 3.849247248664043, + "tokens_seen": 629538816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040869608826479435, + "loss": 1.4092, + "theoretical_loss": 3.8492060972356255, + "tokens_seen": 629604352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086860581745236, + "loss": 1.4264, + "theoretical_loss": 3.8491649512897017, + "tokens_seen": 629669888 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040867602808425277, + "loss": 1.573, + "theoretical_loss": 3.849123810824971, + "tokens_seen": 629735424 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040866599799398195, + "loss": 1.5953, + "theoretical_loss": 3.8490826758401324, + "tokens_seen": 629800960 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040865596790371113, + "loss": 1.615, + "theoretical_loss": 3.849041546333887, + "tokens_seen": 629866496 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040864593781344036, + "loss": 1.4758, + "theoretical_loss": 3.8490004223049343, + "tokens_seen": 629932032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086359077231695, + "loss": 1.5158, + "theoretical_loss": 3.8489593037519754, + "tokens_seen": 629997568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086258776328987, + "loss": 1.4844, + "theoretical_loss": 3.8489181906737118, + "tokens_seen": 630063104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040861584754262785, + "loss": 1.5509, + "theoretical_loss": 3.848877083068845, + "tokens_seen": 630128640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004086058174523571, + "loss": 1.5991, + "theoretical_loss": 3.8488359809360766, + "tokens_seen": 630194176 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040859578736208627, + "loss": 1.5696, + "theoretical_loss": 3.8487948842741107, + "tokens_seen": 630259712 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040858575727181545, + "loss": 1.5235, + "theoretical_loss": 3.8487537930816487, + "tokens_seen": 630325248 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040857572718154463, + "loss": 1.6065, + "theoretical_loss": 3.8487127073573957, + "tokens_seen": 630390784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085656970912738, + "loss": 1.6098, + "theoretical_loss": 3.8486716271000545, + "tokens_seen": 630456320 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408555667001003, + "loss": 1.513, + "theoretical_loss": 3.8486305523083297, + "tokens_seen": 630521856 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040854563691073223, + "loss": 1.6026, + "theoretical_loss": 3.848589482980927, + "tokens_seen": 630587392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040853560682046136, + "loss": 1.5343, + "theoretical_loss": 3.8485484191165504, + "tokens_seen": 630652928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085255767301906, + "loss": 1.5459, + "theoretical_loss": 3.8485073607139064, + "tokens_seen": 630718464 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004085155466399197, + "loss": 1.5197, + "theoretical_loss": 3.8484663077717007, + "tokens_seen": 630784000 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040850551654964895, + "loss": 1.4473, + "theoretical_loss": 3.84842526028864, + "tokens_seen": 630849536 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040849548645937814, + "loss": 1.6753, + "theoretical_loss": 3.8483842182634325, + "tokens_seen": 630915072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084854563691073, + "loss": 1.5231, + "theoretical_loss": 3.848343181694784, + "tokens_seen": 630980608 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084754262788365, + "loss": 1.4233, + "theoretical_loss": 3.8483021505814032, + "tokens_seen": 631046144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040846539618856573, + "loss": 1.5378, + "theoretical_loss": 3.8482611249219985, + "tokens_seen": 631111680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040845536609829486, + "loss": 1.5958, + "theoretical_loss": 3.8482201047152786, + "tokens_seen": 631177216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084453360080241, + "loss": 1.5318, + "theoretical_loss": 3.8481790899599524, + "tokens_seen": 631242752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084353059177532, + "loss": 1.6704, + "theoretical_loss": 3.84813808065473, + "tokens_seen": 631308288 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040842527582748246, + "loss": 1.5496, + "theoretical_loss": 3.8480970767983216, + "tokens_seen": 631373824 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040841524573721164, + "loss": 1.6253, + "theoretical_loss": 3.8480560783894373, + "tokens_seen": 631439360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004084052156469408, + "loss": 1.6176, + "theoretical_loss": 3.8480150854267885, + "tokens_seen": 631504896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040839518555667005, + "loss": 1.5124, + "theoretical_loss": 3.847974097909087, + "tokens_seen": 631570432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083851554663992, + "loss": 1.4343, + "theoretical_loss": 3.8479331158350436, + "tokens_seen": 631635968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083751253761284, + "loss": 1.4744, + "theoretical_loss": 3.8478921392033714, + "tokens_seen": 631701504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083650952858576, + "loss": 1.468, + "theoretical_loss": 3.8478511680127827, + "tokens_seen": 631767040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083550651955868, + "loss": 1.5735, + "theoretical_loss": 3.847810202261991, + "tokens_seen": 631832576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040834503510531596, + "loss": 1.7198, + "theoretical_loss": 3.8477692419497105, + "tokens_seen": 631898112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083350050150452, + "loss": 1.54, + "theoretical_loss": 3.847728287074654, + "tokens_seen": 631963648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083249749247743, + "loss": 1.5837, + "theoretical_loss": 3.847687337635537, + "tokens_seen": 632029184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040831494483450356, + "loss": 1.5567, + "theoretical_loss": 3.8476463936310736, + "tokens_seen": 632094720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004083049147442327, + "loss": 1.5941, + "theoretical_loss": 3.8476054550599796, + "tokens_seen": 632160256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082948846539619, + "loss": 1.685, + "theoretical_loss": 3.8475645219209715, + "tokens_seen": 632225792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082848545636911, + "loss": 1.6897, + "theoretical_loss": 3.847523594212765, + "tokens_seen": 632291328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082748244734203, + "loss": 1.5321, + "theoretical_loss": 3.8474826719340762, + "tokens_seen": 632356864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 235545, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8557286262512207, + "objective/train/theoretical_loss": 3.847441755083623, + "objective/train/tokens_used": 652882400, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040826479438314946, + "loss": 1.7271, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040825476429287864, + "loss": 1.5969, + "theoretical_loss": 3.8474008436601226, + "tokens_seen": 632487936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082447342026078, + "loss": 1.6062, + "theoretical_loss": 3.8473599376622927, + "tokens_seen": 632553472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040823470411233706, + "loss": 1.4892, + "theoretical_loss": 3.8473190370888526, + "tokens_seen": 632619008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082246740220662, + "loss": 1.5431, + "theoretical_loss": 3.8472781419385207, + "tokens_seen": 632684544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004082146439317954, + "loss": 1.5526, + "theoretical_loss": 3.8472372522100162, + "tokens_seen": 632750080 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040820461384152455, + "loss": 1.4687, + "theoretical_loss": 3.8471963679020593, + "tokens_seen": 632815616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081945837512538, + "loss": 1.5975, + "theoretical_loss": 3.8471554890133692, + "tokens_seen": 632881152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040818455366098297, + "loss": 1.5126, + "theoretical_loss": 3.847114615542667, + "tokens_seen": 632946688 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040817452357071215, + "loss": 1.5849, + "theoretical_loss": 3.8470737474886745, + "tokens_seen": 633012224 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040816449348044133, + "loss": 1.5163, + "theoretical_loss": 3.847032884850112, + "tokens_seen": 633077760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040815446339017056, + "loss": 1.5905, + "theoretical_loss": 3.846992027625702, + "tokens_seen": 633143296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081444332998997, + "loss": 1.5329, + "theoretical_loss": 3.8469511758141666, + "tokens_seen": 633208832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081344032096289, + "loss": 1.4854, + "theoretical_loss": 3.8469103294142286, + "tokens_seen": 633274368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040812437311935805, + "loss": 1.5379, + "theoretical_loss": 3.8468694884246117, + "tokens_seen": 633339904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004081143430290873, + "loss": 1.5361, + "theoretical_loss": 3.8468286528440396, + "tokens_seen": 633405440 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040810431293881647, + "loss": 1.5075, + "theoretical_loss": 3.846787822671235, + "tokens_seen": 633470976 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040809428284854565, + "loss": 1.5241, + "theoretical_loss": 3.8467469979049245, + "tokens_seen": 633536512 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040808425275827483, + "loss": 1.5234, + "theoretical_loss": 3.8467061785438315, + "tokens_seen": 633602048 + }, + { + "epoch": 0.19, + "learning_rate": 0.000408074222668004, + "loss": 1.4625, + "theoretical_loss": 3.8466653645866815, + "tokens_seen": 633667584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080641925777332, + "loss": 1.593, + "theoretical_loss": 3.8466245560322005, + "tokens_seen": 633733120 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040805416248746243, + "loss": 1.6065, + "theoretical_loss": 3.8465837528791154, + "tokens_seen": 633798656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040804413239719156, + "loss": 1.5561, + "theoretical_loss": 3.846542955126152, + "tokens_seen": 633864192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080341023069208, + "loss": 1.566, + "theoretical_loss": 3.8465021627720377, + "tokens_seen": 633929728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004080240722166499, + "loss": 1.6512, + "theoretical_loss": 3.8464613758155, + "tokens_seen": 633995264 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040801404212637915, + "loss": 1.6203, + "theoretical_loss": 3.8464205942552665, + "tokens_seen": 634060800 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040800401203610834, + "loss": 1.6446, + "theoretical_loss": 3.8463798180900666, + "tokens_seen": 634126336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079939819458375, + "loss": 1.5449, + "theoretical_loss": 3.8463390473186285, + "tokens_seen": 634191872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079839518555667, + "loss": 1.5319, + "theoretical_loss": 3.8462982819396814, + "tokens_seen": 634257408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040797392176529593, + "loss": 1.4887, + "theoretical_loss": 3.846257521951955, + "tokens_seen": 634322944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040796389167502506, + "loss": 1.5316, + "theoretical_loss": 3.84621676735418, + "tokens_seen": 634388480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079538615847543, + "loss": 1.6153, + "theoretical_loss": 3.8461760181450857, + "tokens_seen": 634454016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079438314944834, + "loss": 1.4907, + "theoretical_loss": 3.8461352743234043, + "tokens_seen": 634519552 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040793380140421266, + "loss": 1.5497, + "theoretical_loss": 3.8460945358878664, + "tokens_seen": 634585088 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040792377131394184, + "loss": 1.6061, + "theoretical_loss": 3.846053802837204, + "tokens_seen": 634650624 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407913741223671, + "loss": 1.5631, + "theoretical_loss": 3.8460130751701502, + "tokens_seen": 634716160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004079037111334002, + "loss": 1.511, + "theoretical_loss": 3.845972352885437, + "tokens_seen": 634781696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078936810431294, + "loss": 1.5487, + "theoretical_loss": 3.845931635981797, + "tokens_seen": 634847232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040788365095285856, + "loss": 1.5311, + "theoretical_loss": 3.845890924457965, + "tokens_seen": 634912768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078736208625878, + "loss": 1.4891, + "theoretical_loss": 3.845850218312674, + "tokens_seen": 634978304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078635907723169, + "loss": 1.6366, + "theoretical_loss": 3.8458095175446587, + "tokens_seen": 635043840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040785356068204616, + "loss": 1.5425, + "theoretical_loss": 3.8457688221526545, + "tokens_seen": 635109376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078435305917753, + "loss": 1.5193, + "theoretical_loss": 3.8457281321353953, + "tokens_seen": 635174912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078335005015045, + "loss": 1.5147, + "theoretical_loss": 3.8456874474916183, + "tokens_seen": 635240448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078234704112337, + "loss": 1.5124, + "theoretical_loss": 3.8456467682200586, + "tokens_seen": 635305984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004078134403209629, + "loss": 1.5616, + "theoretical_loss": 3.8456060943194537, + "tokens_seen": 635371520 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040780341023069207, + "loss": 1.499, + "theoretical_loss": 3.845565425788539, + "tokens_seen": 635437056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077933801404213, + "loss": 1.5549, + "theoretical_loss": 3.8455247626260536, + "tokens_seen": 635502592 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040778335005015043, + "loss": 1.6149, + "theoretical_loss": 3.845484104830734, + "tokens_seen": 635568128 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040777331995987966, + "loss": 1.5163, + "theoretical_loss": 3.84544345240132, + "tokens_seen": 635633664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 236526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5637685060501099, + "objective/train/theoretical_loss": 3.8454028053365494, + "objective/train/tokens_used": 656159200, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077632898696088, + "loss": 1.5403, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407753259779338, + "loss": 1.5685, + "theoretical_loss": 3.8453621636351607, + "tokens_seen": 635764736 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077432296890672, + "loss": 1.5175, + "theoretical_loss": 3.8453215272958943, + "tokens_seen": 635830272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004077331995987964, + "loss": 1.5683, + "theoretical_loss": 3.84528089631749, + "tokens_seen": 635895808 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040772316950852557, + "loss": 1.4538, + "theoretical_loss": 3.845240270698688, + "tokens_seen": 635961344 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040771313941825475, + "loss": 1.5264, + "theoretical_loss": 3.845199650438229, + "tokens_seen": 636026880 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040770310932798393, + "loss": 1.5559, + "theoretical_loss": 3.845159035534855, + "tokens_seen": 636092416 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040769307923771317, + "loss": 1.5929, + "theoretical_loss": 3.845118425987307, + "tokens_seen": 636157952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076830491474423, + "loss": 1.6021, + "theoretical_loss": 3.845077821794327, + "tokens_seen": 636223488 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040767301905717153, + "loss": 1.6037, + "theoretical_loss": 3.845037222954658, + "tokens_seen": 636289024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076629889669007, + "loss": 1.5918, + "theoretical_loss": 3.8449966294670426, + "tokens_seen": 636354560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076529588766299, + "loss": 1.5621, + "theoretical_loss": 3.8449560413302244, + "tokens_seen": 636420096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076429287863591, + "loss": 1.6512, + "theoretical_loss": 3.844915458542947, + "tokens_seen": 636485632 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040763289869608825, + "loss": 1.5133, + "theoretical_loss": 3.844874881103955, + "tokens_seen": 636551168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004076228686058175, + "loss": 1.4839, + "theoretical_loss": 3.844834309011992, + "tokens_seen": 636616704 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040761283851554667, + "loss": 1.4863, + "theoretical_loss": 3.8447937422658045, + "tokens_seen": 636682240 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040760280842527585, + "loss": 1.5184, + "theoretical_loss": 3.8447531808641373, + "tokens_seen": 636747776 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040759277833500503, + "loss": 1.5222, + "theoretical_loss": 3.844712624805736, + "tokens_seen": 636813312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075827482447342, + "loss": 1.7884, + "theoretical_loss": 3.844672074089347, + "tokens_seen": 636878848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075727181544634, + "loss": 1.4945, + "theoretical_loss": 3.844631528713718, + "tokens_seen": 636944384 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040756268806419263, + "loss": 1.5775, + "theoretical_loss": 3.844590988677595, + "tokens_seen": 637009920 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040755265797392176, + "loss": 1.472, + "theoretical_loss": 3.844550453979726, + "tokens_seen": 637075456 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407542627883651, + "loss": 1.5602, + "theoretical_loss": 3.8445099246188597, + "tokens_seen": 637140992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075325977933801, + "loss": 1.4594, + "theoretical_loss": 3.8444694005937436, + "tokens_seen": 637206528 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040752256770310935, + "loss": 1.6233, + "theoretical_loss": 3.844428881903127, + "tokens_seen": 637272064 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040751253761283854, + "loss": 1.5225, + "theoretical_loss": 3.844388368545759, + "tokens_seen": 637337600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004075025075225677, + "loss": 1.5078, + "theoretical_loss": 3.84434786052039, + "tokens_seen": 637403136 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074924774322969, + "loss": 1.5161, + "theoretical_loss": 3.8443073578257687, + "tokens_seen": 637468672 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040748244734202613, + "loss": 1.4539, + "theoretical_loss": 3.8442668604606474, + "tokens_seen": 637534208 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040747241725175526, + "loss": 1.5416, + "theoretical_loss": 3.844226368423776, + "tokens_seen": 637599744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074623871614845, + "loss": 1.3917, + "theoretical_loss": 3.844185881713906, + "tokens_seen": 637665280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074523570712136, + "loss": 1.5222, + "theoretical_loss": 3.844145400329789, + "tokens_seen": 637730816 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040744232698094286, + "loss": 1.5393, + "theoretical_loss": 3.8441049242701784, + "tokens_seen": 637796352 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040743229689067204, + "loss": 1.5483, + "theoretical_loss": 3.8440644535338255, + "tokens_seen": 637861888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074222668004012, + "loss": 1.5269, + "theoretical_loss": 3.844023988119484, + "tokens_seen": 637927424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074122367101304, + "loss": 1.5005, + "theoretical_loss": 3.8439835280259076, + "tokens_seen": 637992960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004074022066198596, + "loss": 1.5344, + "theoretical_loss": 3.84394307325185, + "tokens_seen": 638058496 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040739217652958876, + "loss": 1.6283, + "theoretical_loss": 3.8439026237960654, + "tokens_seen": 638124032 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407382146439318, + "loss": 1.5138, + "theoretical_loss": 3.8438621796573087, + "tokens_seen": 638189568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073721163490471, + "loss": 1.4514, + "theoretical_loss": 3.843821740834335, + "tokens_seen": 638255104 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040736208625877636, + "loss": 1.4313, + "theoretical_loss": 3.8437813073258997, + "tokens_seen": 638320640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073520561685055, + "loss": 1.5265, + "theoretical_loss": 3.8437408791307597, + "tokens_seen": 638386176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073420260782347, + "loss": 1.5691, + "theoretical_loss": 3.843700456247671, + "tokens_seen": 638451712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073319959879639, + "loss": 1.5339, + "theoretical_loss": 3.84366003867539, + "tokens_seen": 638517248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073219658976931, + "loss": 1.6089, + "theoretical_loss": 3.8436196264126736, + "tokens_seen": 638582784 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040731193580742227, + "loss": 1.5577, + "theoretical_loss": 3.843579219458281, + "tokens_seen": 638648320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004073019057171515, + "loss": 1.6523, + "theoretical_loss": 3.843538817810969, + "tokens_seen": 638713856 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040729187562688063, + "loss": 1.5776, + "theoretical_loss": 3.843498421469497, + "tokens_seen": 638779392 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040728184553660986, + "loss": 1.4831, + "theoretical_loss": 3.843458030432623, + "tokens_seen": 638844928 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407271815446339, + "loss": 1.5338, + "theoretical_loss": 3.8434176446991075, + "tokens_seen": 638910464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 237858, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3501896858215332, + "objective/train/theoretical_loss": 3.8433772642677093, + "objective/train/tokens_used": 659436000, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072617853560682, + "loss": 1.3135, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072517552657974, + "loss": 1.5432, + "theoretical_loss": 3.8433368891371895, + "tokens_seen": 639041536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004072417251755266, + "loss": 1.5863, + "theoretical_loss": 3.843296519306308, + "tokens_seen": 639107072 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040723169508525577, + "loss": 1.5192, + "theoretical_loss": 3.843256154773826, + "tokens_seen": 639172608 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040722166499498495, + "loss": 1.5203, + "theoretical_loss": 3.843215795538505, + "tokens_seen": 639238144 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040721163490471413, + "loss": 1.5799, + "theoretical_loss": 3.843175441599107, + "tokens_seen": 639303680 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040720160481444337, + "loss": 1.6912, + "theoretical_loss": 3.8431350929543946, + "tokens_seen": 639369216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071915747241725, + "loss": 1.5126, + "theoretical_loss": 3.8430947496031296, + "tokens_seen": 639434752 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040718154463390173, + "loss": 1.4154, + "theoretical_loss": 3.8430544115440752, + "tokens_seen": 639500288 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071715145436309, + "loss": 1.5388, + "theoretical_loss": 3.8430140787759957, + "tokens_seen": 639565824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004071614844533601, + "loss": 1.5556, + "theoretical_loss": 3.842973751297655, + "tokens_seen": 639631360 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040715145436308927, + "loss": 1.5245, + "theoretical_loss": 3.8429334291078168, + "tokens_seen": 639696896 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040714142427281845, + "loss": 1.5104, + "theoretical_loss": 3.842893112205246, + "tokens_seen": 639762432 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040713139418254763, + "loss": 1.5316, + "theoretical_loss": 3.842852800588709, + "tokens_seen": 639827968 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040712136409227687, + "loss": 1.5107, + "theoretical_loss": 3.8428124942569695, + "tokens_seen": 639893504 + }, + { + "epoch": 0.19, + "learning_rate": 0.000407111334002006, + "loss": 1.4888, + "theoretical_loss": 3.842772193208795, + "tokens_seen": 639959040 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040710130391173523, + "loss": 1.4733, + "theoretical_loss": 3.8427318974429507, + "tokens_seen": 640024576 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040709127382146436, + "loss": 1.467, + "theoretical_loss": 3.8426916069582044, + "tokens_seen": 640090112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070812437311936, + "loss": 1.5504, + "theoretical_loss": 3.842651321753323, + "tokens_seen": 640155648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070712136409228, + "loss": 1.6126, + "theoretical_loss": 3.8426110418270745, + "tokens_seen": 640221184 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040706118355065196, + "loss": 1.6394, + "theoretical_loss": 3.842570767178227, + "tokens_seen": 640286720 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040705115346038114, + "loss": 1.4629, + "theoretical_loss": 3.8425304978055483, + "tokens_seen": 640352256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070411233701103, + "loss": 1.5523, + "theoretical_loss": 3.8424902337078084, + "tokens_seen": 640417792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070310932798395, + "loss": 1.6199, + "theoretical_loss": 3.8424499748837753, + "tokens_seen": 640483328 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040702106318956874, + "loss": 1.4774, + "theoretical_loss": 3.8424097213322197, + "tokens_seen": 640548864 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040701103309929786, + "loss": 1.4376, + "theoretical_loss": 3.8423694730519116, + "tokens_seen": 640614400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004070010030090271, + "loss": 1.5801, + "theoretical_loss": 3.8423292300416216, + "tokens_seen": 640679936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069909729187563, + "loss": 1.5227, + "theoretical_loss": 3.8422889923001207, + "tokens_seen": 640745472 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040698094282848546, + "loss": 1.5053, + "theoretical_loss": 3.84224875982618, + "tokens_seen": 640811008 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040697091273821464, + "loss": 1.3897, + "theoretical_loss": 3.8422085326185718, + "tokens_seen": 640876544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069608826479438, + "loss": 1.6027, + "theoretical_loss": 3.842168310676067, + "tokens_seen": 640942080 + }, + { + "epoch": 0.19, + "learning_rate": 0.000406950852557673, + "loss": 1.5084, + "theoretical_loss": 3.8421280939974407, + "tokens_seen": 641007616 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040694082246740224, + "loss": 1.5864, + "theoretical_loss": 3.8420878825814637, + "tokens_seen": 641073152 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040693079237713137, + "loss": 1.575, + "theoretical_loss": 3.84204767642691, + "tokens_seen": 641138688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004069207622868606, + "loss": 1.5235, + "theoretical_loss": 3.8420074755325544, + "tokens_seen": 641204224 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040691073219658973, + "loss": 1.5468, + "theoretical_loss": 3.8419672798971707, + "tokens_seen": 641269760 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040690070210631896, + "loss": 1.5618, + "theoretical_loss": 3.8419270895195323, + "tokens_seen": 641335296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068906720160482, + "loss": 1.5377, + "theoretical_loss": 3.8418869043984163, + "tokens_seen": 641400832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068806419257773, + "loss": 1.5648, + "theoretical_loss": 3.841846724532597, + "tokens_seen": 641466368 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040687061183550656, + "loss": 1.5282, + "theoretical_loss": 3.8418065499208502, + "tokens_seen": 641531904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068605817452357, + "loss": 1.5086, + "theoretical_loss": 3.8417663805619533, + "tokens_seen": 641597440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068505516549649, + "loss": 1.4734, + "theoretical_loss": 3.841726216454682, + "tokens_seen": 641662976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068405215646941, + "loss": 1.6179, + "theoretical_loss": 3.8416860575978142, + "tokens_seen": 641728512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068304914744233, + "loss": 1.4551, + "theoretical_loss": 3.841645903990127, + "tokens_seen": 641794048 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040682046138415247, + "loss": 1.4732, + "theoretical_loss": 3.841605755630398, + "tokens_seen": 641859584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004068104312938817, + "loss": 1.6127, + "theoretical_loss": 3.8415656125174067, + "tokens_seen": 641925120 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040680040120361083, + "loss": 1.4876, + "theoretical_loss": 3.8415254746499308, + "tokens_seen": 641990656 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040679037111334006, + "loss": 1.4941, + "theoretical_loss": 3.8414853420267496, + "tokens_seen": 642056192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067803410230692, + "loss": 1.4102, + "theoretical_loss": 3.841445214646644, + "tokens_seen": 642121728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067703109327984, + "loss": 1.5399, + "theoretical_loss": 3.8414050925083925, + "tokens_seen": 642187264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 239059, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5365581512451172, + "objective/train/theoretical_loss": 3.8413649756107757, + "objective/train/tokens_used": 662712800, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067602808425276, + "loss": 1.6487, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067502507522568, + "loss": 1.5221, + "theoretical_loss": 3.8413248639525754, + "tokens_seen": 642318336 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040674022066198597, + "loss": 1.5418, + "theoretical_loss": 3.8412847575325717, + "tokens_seen": 642383872 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040673019057171515, + "loss": 1.6577, + "theoretical_loss": 3.8412446563495477, + "tokens_seen": 642449408 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040672016048144433, + "loss": 1.6609, + "theoretical_loss": 3.8412045604022835, + "tokens_seen": 642514944 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040671013039117357, + "loss": 1.4784, + "theoretical_loss": 3.8411644696895633, + "tokens_seen": 642580480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004067001003009027, + "loss": 1.4818, + "theoretical_loss": 3.8411243842101688, + "tokens_seen": 642646016 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040669007021063193, + "loss": 1.5629, + "theoretical_loss": 3.841084303962884, + "tokens_seen": 642711552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066800401203611, + "loss": 1.5254, + "theoretical_loss": 3.841044228946492, + "tokens_seen": 642777088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066700100300903, + "loss": 1.467, + "theoretical_loss": 3.8410041591597777, + "tokens_seen": 642842624 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040665997993981947, + "loss": 1.5073, + "theoretical_loss": 3.840964094601525, + "tokens_seen": 642908160 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040664994984954865, + "loss": 1.557, + "theoretical_loss": 3.8409240352705187, + "tokens_seen": 642973696 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040663991975927783, + "loss": 1.5135, + "theoretical_loss": 3.840883981165544, + "tokens_seen": 643039232 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040662988966900707, + "loss": 1.6123, + "theoretical_loss": 3.8408439322853876, + "tokens_seen": 643104768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004066198595787362, + "loss": 1.4489, + "theoretical_loss": 3.8408038886288347, + "tokens_seen": 643170304 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040660982948846543, + "loss": 1.5614, + "theoretical_loss": 3.840763850194672, + "tokens_seen": 643235840 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040659979939819456, + "loss": 1.5785, + "theoretical_loss": 3.8407238169816864, + "tokens_seen": 643301376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0004065897693079238, + "loss": 1.5334, + "theoretical_loss": 3.840683788988666, + "tokens_seen": 643366912 + }, + { + "epoch": 0.19, + "learning_rate": 0.000406579739217653, + "loss": 1.4942, + "theoretical_loss": 3.840643766214397, + "tokens_seen": 643432448 + }, + { + "epoch": 0.19, + "learning_rate": 0.00040656970912738216, + "loss": 1.4938, + "theoretical_loss": 3.840603748657669, + "tokens_seen": 643497984 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040655967903711134, + "loss": 1.4285, + "theoretical_loss": 3.8405637363172693, + "tokens_seen": 643563520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065496489468405, + "loss": 1.4125, + "theoretical_loss": 3.8405237291919883, + "tokens_seen": 643629056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065396188565697, + "loss": 1.5202, + "theoretical_loss": 3.8404837272806143, + "tokens_seen": 643694592 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040652958876629894, + "loss": 1.5266, + "theoretical_loss": 3.840443730581937, + "tokens_seen": 643760128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040651955867602806, + "loss": 1.5428, + "theoretical_loss": 3.8404037390947474, + "tokens_seen": 643825664 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004065095285857573, + "loss": 1.4779, + "theoretical_loss": 3.8403637528178356, + "tokens_seen": 643891200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064994984954865, + "loss": 1.5054, + "theoretical_loss": 3.8403237717499925, + "tokens_seen": 643956736 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040648946840521566, + "loss": 1.5338, + "theoretical_loss": 3.8402837958900093, + "tokens_seen": 644022272 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040647943831494484, + "loss": 1.5052, + "theoretical_loss": 3.840243825236678, + "tokens_seen": 644087808 + }, + { + "epoch": 0.2, + "learning_rate": 0.000406469408224674, + "loss": 1.4812, + "theoretical_loss": 3.840203859788791, + "tokens_seen": 644153344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064593781344032, + "loss": 1.4694, + "theoretical_loss": 3.840163899545141, + "tokens_seen": 644218880 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040644934804413244, + "loss": 1.5376, + "theoretical_loss": 3.8401239445045205, + "tokens_seen": 644284416 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040643931795386157, + "loss": 1.5073, + "theoretical_loss": 3.840083994665723, + "tokens_seen": 644349952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004064292878635908, + "loss": 1.4699, + "theoretical_loss": 3.840044050027543, + "tokens_seen": 644415488 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040641925777331993, + "loss": 1.4745, + "theoretical_loss": 3.8400041105887732, + "tokens_seen": 644481024 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040640922768304916, + "loss": 1.6088, + "theoretical_loss": 3.8399641763482095, + "tokens_seen": 644546560 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040639919759277834, + "loss": 1.5633, + "theoretical_loss": 3.8399242473046464, + "tokens_seen": 644612096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063891675025075, + "loss": 1.5584, + "theoretical_loss": 3.8398843234568796, + "tokens_seen": 644677632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063791374122367, + "loss": 1.5323, + "theoretical_loss": 3.8398444048037046, + "tokens_seen": 644743168 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063691073219659, + "loss": 1.5219, + "theoretical_loss": 3.8398044913439175, + "tokens_seen": 644808704 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040635907723169507, + "loss": 1.4713, + "theoretical_loss": 3.839764583076316, + "tokens_seen": 644874240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004063490471414243, + "loss": 1.5955, + "theoretical_loss": 3.839724679999695, + "tokens_seen": 644939776 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040633901705115343, + "loss": 1.6057, + "theoretical_loss": 3.8396847821128537, + "tokens_seen": 645005312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040632898696088267, + "loss": 1.6401, + "theoretical_loss": 3.83964488941459, + "tokens_seen": 645070848 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040631895687061185, + "loss": 1.5346, + "theoretical_loss": 3.8396050019037005, + "tokens_seen": 645136384 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040630892678034103, + "loss": 1.5136, + "theoretical_loss": 3.8395651195789853, + "tokens_seen": 645201920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062988966900702, + "loss": 1.5413, + "theoretical_loss": 3.839525242439243, + "tokens_seen": 645267456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062888665997994, + "loss": 1.5223, + "theoretical_loss": 3.8394853704832723, + "tokens_seen": 645332992 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040627883650952857, + "loss": 1.5748, + "theoretical_loss": 3.8394455037098743, + "tokens_seen": 645398528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062688064192578, + "loss": 1.5471, + "theoretical_loss": 3.8394056421178484, + "tokens_seen": 645464064 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 240316, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5447896718978882, + "objective/train/theoretical_loss": 3.8393657857059953, + "objective/train/tokens_used": 665989600, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040625877632898693, + "loss": 1.4992, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040624874623871617, + "loss": 1.6208, + "theoretical_loss": 3.839325934473116, + "tokens_seen": 645595136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062387161484453, + "loss": 1.7084, + "theoretical_loss": 3.839286088418012, + "tokens_seen": 645660672 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040622868605817453, + "loss": 1.5142, + "theoretical_loss": 3.839246247539485, + "tokens_seen": 645726208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062186559679037, + "loss": 1.5852, + "theoretical_loss": 3.8392064118363374, + "tokens_seen": 645791744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004062086258776329, + "loss": 1.5626, + "theoretical_loss": 3.839166581307372, + "tokens_seen": 645857280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061985957873621, + "loss": 1.536, + "theoretical_loss": 3.8391267559513915, + "tokens_seen": 645922816 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040618856569709126, + "loss": 1.591, + "theoretical_loss": 3.839086935767199, + "tokens_seen": 645988352 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040617853560682044, + "loss": 1.5866, + "theoretical_loss": 3.839047120753599, + "tokens_seen": 646053888 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040616850551654967, + "loss": 1.4159, + "theoretical_loss": 3.8390073109093947, + "tokens_seen": 646119424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061584754262788, + "loss": 1.5251, + "theoretical_loss": 3.8389675062333923, + "tokens_seen": 646184960 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040614844533600804, + "loss": 1.5122, + "theoretical_loss": 3.838927706724395, + "tokens_seen": 646250496 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040613841524573727, + "loss": 1.4471, + "theoretical_loss": 3.8388879123812094, + "tokens_seen": 646316032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004061283851554664, + "loss": 1.5016, + "theoretical_loss": 3.8388481232026406, + "tokens_seen": 646381568 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040611835506519563, + "loss": 1.464, + "theoretical_loss": 3.838808339187495, + "tokens_seen": 646447104 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040610832497492476, + "loss": 1.5134, + "theoretical_loss": 3.8387685603345805, + "tokens_seen": 646512640 + }, + { + "epoch": 0.2, + "learning_rate": 0.000406098294884654, + "loss": 1.5466, + "theoretical_loss": 3.838728786642702, + "tokens_seen": 646578176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060882647943832, + "loss": 1.62, + "theoretical_loss": 3.838689018110667, + "tokens_seen": 646643712 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040607823470411236, + "loss": 1.4736, + "theoretical_loss": 3.838649254737285, + "tokens_seen": 646709248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040606820461384154, + "loss": 1.6105, + "theoretical_loss": 3.838609496521363, + "tokens_seen": 646774784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060581745235707, + "loss": 1.6878, + "theoretical_loss": 3.8385697434617096, + "tokens_seen": 646840320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060481444332999, + "loss": 1.5331, + "theoretical_loss": 3.838529995557134, + "tokens_seen": 646905856 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040603811434302914, + "loss": 1.4136, + "theoretical_loss": 3.8384902528064457, + "tokens_seen": 646971392 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040602808425275826, + "loss": 1.5273, + "theoretical_loss": 3.8384505152084545, + "tokens_seen": 647036928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060180541624875, + "loss": 1.5923, + "theoretical_loss": 3.8384107827619696, + "tokens_seen": 647102464 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004060080240722167, + "loss": 1.5648, + "theoretical_loss": 3.8383710554658026, + "tokens_seen": 647168000 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040599799398194586, + "loss": 1.5041, + "theoretical_loss": 3.838331333318764, + "tokens_seen": 647233536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040598796389167504, + "loss": 1.4872, + "theoretical_loss": 3.8382916163196654, + "tokens_seen": 647299072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004059779338014042, + "loss": 1.5001, + "theoretical_loss": 3.8382519044673176, + "tokens_seen": 647364608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004059679037111334, + "loss": 1.5208, + "theoretical_loss": 3.8382121977605346, + "tokens_seen": 647430144 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040595787362086264, + "loss": 1.5819, + "theoretical_loss": 3.8381724961981267, + "tokens_seen": 647495680 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040594784353059177, + "loss": 1.5743, + "theoretical_loss": 3.8381327997789088, + "tokens_seen": 647561216 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405937813440321, + "loss": 1.6321, + "theoretical_loss": 3.8380931085016927, + "tokens_seen": 647626752 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040592778335005013, + "loss": 1.5802, + "theoretical_loss": 3.8380534223652933, + "tokens_seen": 647692288 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040591775325977936, + "loss": 1.6076, + "theoretical_loss": 3.838013741368523, + "tokens_seen": 647757824 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040590772316950854, + "loss": 1.5583, + "theoretical_loss": 3.837974065510198, + "tokens_seen": 647823360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058976930792377, + "loss": 1.512, + "theoretical_loss": 3.837934394789133, + "tokens_seen": 647888896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058876629889669, + "loss": 1.4956, + "theoretical_loss": 3.8378947292041428, + "tokens_seen": 647954432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058776328986961, + "loss": 1.4939, + "theoretical_loss": 3.8378550687540427, + "tokens_seen": 648019968 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040586760280842527, + "loss": 1.5365, + "theoretical_loss": 3.8378154134376494, + "tokens_seen": 648085504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058575727181545, + "loss": 1.6332, + "theoretical_loss": 3.837775763253779, + "tokens_seen": 648151040 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040584754262788363, + "loss": 1.6153, + "theoretical_loss": 3.8377361182012484, + "tokens_seen": 648216576 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040583751253761287, + "loss": 1.6038, + "theoretical_loss": 3.8376964782788754, + "tokens_seen": 648282112 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040582748244734205, + "loss": 1.5432, + "theoretical_loss": 3.8376568434854765, + "tokens_seen": 648347648 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040581745235707123, + "loss": 1.5573, + "theoretical_loss": 3.837617213819871, + "tokens_seen": 648413184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004058074222668004, + "loss": 1.5003, + "theoretical_loss": 3.837577589280876, + "tokens_seen": 648478720 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057973921765296, + "loss": 1.5499, + "theoretical_loss": 3.8375379698673115, + "tokens_seen": 648544256 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040578736208625877, + "loss": 1.4866, + "theoretical_loss": 3.837498355577996, + "tokens_seen": 648609792 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405777331995988, + "loss": 1.5908, + "theoretical_loss": 3.837458746411749, + "tokens_seen": 648675328 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040576730190571713, + "loss": 1.4954, + "theoretical_loss": 3.8374191423673913, + "tokens_seen": 648740864 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 241544, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.447023630142212, + "objective/train/theoretical_loss": 3.8373795434437423, + "objective/train/tokens_used": 669266400, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040575727181544637, + "loss": 1.5337, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057472417251755, + "loss": 1.489, + "theoretical_loss": 3.8373399496396234, + "tokens_seen": 648871936 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040573721163490473, + "loss": 1.562, + "theoretical_loss": 3.8373003609538556, + "tokens_seen": 648937472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057271815446339, + "loss": 1.5749, + "theoretical_loss": 3.8372607773852603, + "tokens_seen": 649003008 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057171514543631, + "loss": 1.6223, + "theoretical_loss": 3.8372211989326592, + "tokens_seen": 649068544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004057071213640923, + "loss": 1.5731, + "theoretical_loss": 3.8371816255948756, + "tokens_seen": 649134080 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040569709127382146, + "loss": 1.4758, + "theoretical_loss": 3.837142057370731, + "tokens_seen": 649199616 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040568706118355064, + "loss": 1.5744, + "theoretical_loss": 3.837102494259049, + "tokens_seen": 649265152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056770310932799, + "loss": 1.5116, + "theoretical_loss": 3.8370629362586532, + "tokens_seen": 649330688 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405667001003009, + "loss": 1.6106, + "theoretical_loss": 3.8370233833683676, + "tokens_seen": 649396224 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040565697091273824, + "loss": 1.7007, + "theoretical_loss": 3.8369838355870165, + "tokens_seen": 649461760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056469408224674, + "loss": 1.5249, + "theoretical_loss": 3.836944292913424, + "tokens_seen": 649527296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056369107321966, + "loss": 1.4675, + "theoretical_loss": 3.8369047553464153, + "tokens_seen": 649592832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004056268806419258, + "loss": 1.4931, + "theoretical_loss": 3.836865222884817, + "tokens_seen": 649658368 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040561685055165496, + "loss": 1.6166, + "theoretical_loss": 3.8368256955274527, + "tokens_seen": 649723904 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040560682046138414, + "loss": 1.5307, + "theoretical_loss": 3.8367861732731505, + "tokens_seen": 649789440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055967903711134, + "loss": 1.4475, + "theoretical_loss": 3.836746656120737, + "tokens_seen": 649854976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055867602808425, + "loss": 1.5561, + "theoretical_loss": 3.8367071440690372, + "tokens_seen": 649920512 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040557673019057174, + "loss": 1.5282, + "theoretical_loss": 3.836667637116881, + "tokens_seen": 649986048 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040556670010030087, + "loss": 1.4795, + "theoretical_loss": 3.8366281352630947, + "tokens_seen": 650051584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055566700100301, + "loss": 1.4678, + "theoretical_loss": 3.8365886385065067, + "tokens_seen": 650117120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055466399197593, + "loss": 1.4638, + "theoretical_loss": 3.836549146845946, + "tokens_seen": 650182656 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040553660982948846, + "loss": 1.4787, + "theoretical_loss": 3.8365096602802407, + "tokens_seen": 650248192 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040552657973921764, + "loss": 1.6573, + "theoretical_loss": 3.83647017880822, + "tokens_seen": 650313728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004055165496489469, + "loss": 1.4977, + "theoretical_loss": 3.836430702428715, + "tokens_seen": 650379264 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405506519558676, + "loss": 1.5475, + "theoretical_loss": 3.8363912311405546, + "tokens_seen": 650444800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040549648946840524, + "loss": 1.5315, + "theoretical_loss": 3.8363517649425702, + "tokens_seen": 650510336 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040548645937813437, + "loss": 1.5494, + "theoretical_loss": 3.8363123038335916, + "tokens_seen": 650575872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054764292878636, + "loss": 1.5108, + "theoretical_loss": 3.8362728478124506, + "tokens_seen": 650641408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054663991975928, + "loss": 1.5751, + "theoretical_loss": 3.8362333968779794, + "tokens_seen": 650706944 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040545636910732197, + "loss": 1.5332, + "theoretical_loss": 3.836193951029008, + "tokens_seen": 650772480 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040544633901705115, + "loss": 1.6001, + "theoretical_loss": 3.8361545102643717, + "tokens_seen": 650838016 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040543630892678033, + "loss": 1.5348, + "theoretical_loss": 3.836115074582901, + "tokens_seen": 650903552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004054262788365095, + "loss": 1.5005, + "theoretical_loss": 3.8360756439834303, + "tokens_seen": 650969088 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040541624874623874, + "loss": 1.4385, + "theoretical_loss": 3.836036218464792, + "tokens_seen": 651034624 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040540621865596787, + "loss": 1.4244, + "theoretical_loss": 3.8359967980258216, + "tokens_seen": 651100160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053961885656971, + "loss": 1.4695, + "theoretical_loss": 3.8359573826653524, + "tokens_seen": 651165696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053861584754263, + "loss": 1.5805, + "theoretical_loss": 3.8359179723822194, + "tokens_seen": 651231232 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040537612838515547, + "loss": 1.6391, + "theoretical_loss": 3.835878567175257, + "tokens_seen": 651296768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053660982948847, + "loss": 1.5481, + "theoretical_loss": 3.8358391670433023, + "tokens_seen": 651362304 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040535606820461383, + "loss": 1.5623, + "theoretical_loss": 3.8357997719851893, + "tokens_seen": 651427840 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040534603811434307, + "loss": 1.5634, + "theoretical_loss": 3.835760381999756, + "tokens_seen": 651493376 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040533600802407225, + "loss": 1.6343, + "theoretical_loss": 3.835720997085838, + "tokens_seen": 651558912 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040532597793380143, + "loss": 1.5858, + "theoretical_loss": 3.835681617242272, + "tokens_seen": 651624448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053159478435306, + "loss": 1.5317, + "theoretical_loss": 3.8356422424678964, + "tokens_seen": 651689984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004053059177532598, + "loss": 1.5877, + "theoretical_loss": 3.835602872761549, + "tokens_seen": 651755520 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040529588766298897, + "loss": 1.5689, + "theoretical_loss": 3.8355635081220667, + "tokens_seen": 651821056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052858575727182, + "loss": 1.5687, + "theoretical_loss": 3.8355241485482887, + "tokens_seen": 651886592 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040527582748244733, + "loss": 1.5169, + "theoretical_loss": 3.835484794039055, + "tokens_seen": 651952128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040526579739217657, + "loss": 1.551, + "theoretical_loss": 3.835445444593203, + "tokens_seen": 652017664 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 242832, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.557466983795166, + "objective/train/theoretical_loss": 3.8354061002095743, + "objective/train/tokens_used": 672543200, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052557673019057, + "loss": 1.5053, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040524573721163493, + "loss": 1.5862, + "theoretical_loss": 3.835366760887008, + "tokens_seen": 652148736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052357071213641, + "loss": 1.4799, + "theoretical_loss": 3.8353274266243442, + "tokens_seen": 652214272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052256770310933, + "loss": 1.5282, + "theoretical_loss": 3.835288097420425, + "tokens_seen": 652279808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004052156469408225, + "loss": 1.6301, + "theoretical_loss": 3.83524877327409, + "tokens_seen": 652345344 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040520561685055166, + "loss": 1.5876, + "theoretical_loss": 3.8352094541841817, + "tokens_seen": 652410880 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040519558676028084, + "loss": 1.4886, + "theoretical_loss": 3.8351701401495424, + "tokens_seen": 652476416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051855566700101, + "loss": 1.4838, + "theoretical_loss": 3.8351308311690144, + "tokens_seen": 652541952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051755265797392, + "loss": 1.5666, + "theoretical_loss": 3.83509152724144, + "tokens_seen": 652607488 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040516549648946844, + "loss": 1.6496, + "theoretical_loss": 3.8350522283656625, + "tokens_seen": 652673024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051554663991976, + "loss": 1.5623, + "theoretical_loss": 3.8350129345405257, + "tokens_seen": 652738560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051454363089268, + "loss": 1.5296, + "theoretical_loss": 3.8349736457648733, + "tokens_seen": 652804096 + }, + { + "epoch": 0.2, + "learning_rate": 0.000405135406218656, + "loss": 1.4913, + "theoretical_loss": 3.83493436203755, + "tokens_seen": 652869632 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040512537612838516, + "loss": 1.5501, + "theoretical_loss": 3.834895083357399, + "tokens_seen": 652935168 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040511534603811434, + "loss": 1.544, + "theoretical_loss": 3.8348558097232672, + "tokens_seen": 653000704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004051053159478436, + "loss": 1.601, + "theoretical_loss": 3.834816541133999, + "tokens_seen": 653066240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050952858575727, + "loss": 1.4543, + "theoretical_loss": 3.83477727758844, + "tokens_seen": 653131776 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040508525576730194, + "loss": 1.691, + "theoretical_loss": 3.8347380190854374, + "tokens_seen": 653197312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040507522567703107, + "loss": 1.5359, + "theoretical_loss": 3.8346987656238367, + "tokens_seen": 653262848 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050651955867603, + "loss": 1.506, + "theoretical_loss": 3.8346595172024855, + "tokens_seen": 653328384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050551654964895, + "loss": 1.4623, + "theoretical_loss": 3.834620273820231, + "tokens_seen": 653393920 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040504513540621866, + "loss": 1.6161, + "theoretical_loss": 3.834581035475921, + "tokens_seen": 653459456 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040503510531594784, + "loss": 1.5334, + "theoretical_loss": 3.8345418021684035, + "tokens_seen": 653524992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050250752256771, + "loss": 1.552, + "theoretical_loss": 3.8345025738965264, + "tokens_seen": 653590528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004050150451354062, + "loss": 1.5299, + "theoretical_loss": 3.8344633506591395, + "tokens_seen": 653656064 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040500501504513544, + "loss": 1.589, + "theoretical_loss": 3.8344241324550907, + "tokens_seen": 653721600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040499498495486457, + "loss": 1.5792, + "theoretical_loss": 3.8343849192832313, + "tokens_seen": 653787136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049849548645938, + "loss": 1.5865, + "theoretical_loss": 3.83434571114241, + "tokens_seen": 653852672 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404974924774323, + "loss": 1.4532, + "theoretical_loss": 3.834306508031478, + "tokens_seen": 653918208 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040496489468405217, + "loss": 1.6311, + "theoretical_loss": 3.8342673099492854, + "tokens_seen": 653983744 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040495486459378135, + "loss": 1.5773, + "theoretical_loss": 3.8342281168946832, + "tokens_seen": 654049280 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040494483450351053, + "loss": 1.5835, + "theoretical_loss": 3.8341889288665234, + "tokens_seen": 654114816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049348044132397, + "loss": 1.5053, + "theoretical_loss": 3.8341497458636575, + "tokens_seen": 654180352 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040492477432296894, + "loss": 1.5216, + "theoretical_loss": 3.834110567884938, + "tokens_seen": 654245888 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040491474423269807, + "loss": 1.6225, + "theoretical_loss": 3.8340713949292176, + "tokens_seen": 654311424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004049047141424273, + "loss": 1.5133, + "theoretical_loss": 3.834032226995349, + "tokens_seen": 654376960 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040489468405215643, + "loss": 1.5333, + "theoretical_loss": 3.8339930640821853, + "tokens_seen": 654442496 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040488465396188567, + "loss": 1.4656, + "theoretical_loss": 3.8339539061885812, + "tokens_seen": 654508032 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040487462387161485, + "loss": 1.4766, + "theoretical_loss": 3.83391475331339, + "tokens_seen": 654573568 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040486459378134403, + "loss": 1.5238, + "theoretical_loss": 3.833875605455466, + "tokens_seen": 654639104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048545636910732, + "loss": 1.538, + "theoretical_loss": 3.8338364626136654, + "tokens_seen": 654704640 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040484453360080245, + "loss": 1.4265, + "theoretical_loss": 3.833797324786842, + "tokens_seen": 654770176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048345035105316, + "loss": 1.5611, + "theoretical_loss": 3.8337581919738524, + "tokens_seen": 654835712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004048244734202608, + "loss": 1.625, + "theoretical_loss": 3.8337190641735512, + "tokens_seen": 654901248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040481444332998994, + "loss": 1.5734, + "theoretical_loss": 3.833679941384797, + "tokens_seen": 654966784 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040480441323971917, + "loss": 1.3988, + "theoretical_loss": 3.833640823606445, + "tokens_seen": 655032320 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040479438314944835, + "loss": 1.5191, + "theoretical_loss": 3.8336017108373523, + "tokens_seen": 655097856 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040478435305917753, + "loss": 1.6306, + "theoretical_loss": 3.833562603076377, + "tokens_seen": 655163392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047743229689067, + "loss": 1.6017, + "theoretical_loss": 3.833523500322377, + "tokens_seen": 655228928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047642928786359, + "loss": 1.4836, + "theoretical_loss": 3.83348440257421, + "tokens_seen": 655294464 + }, + { + "debugging/distinct-1-grams": 0.7560193640250044, + "debugging/entropy-1-grams": 5.2534808493234895, + "debugging/length": 422.2307692307692, + "debugging/num_segments": 13, + "debugging/score": 0.01638830111902192, + "debugging/score_std": 0.019821935735086043, + "epoch": 0.2, + "objective/train/docs_used": 243976, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4795438051223755, + "objective/train/theoretical_loss": 3.833445309830735, + "objective/train/tokens_used": 675820000, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047542627883651, + "loss": 1.5896, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047442326980943, + "loss": 1.5666, + "theoretical_loss": 3.833406222090811, + "tokens_seen": 655425536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040473420260782344, + "loss": 1.4626, + "theoretical_loss": 3.8333671393532973, + "tokens_seen": 655491072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047241725175527, + "loss": 1.558, + "theoretical_loss": 3.8333280616170535, + "tokens_seen": 655556608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004047141424272818, + "loss": 1.529, + "theoretical_loss": 3.8332889888809403, + "tokens_seen": 655622144 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040470411233701104, + "loss": 1.5126, + "theoretical_loss": 3.8332499211438176, + "tokens_seen": 655687680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046940822467402, + "loss": 1.5555, + "theoretical_loss": 3.8332108584045463, + "tokens_seen": 655753216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046840521564694, + "loss": 1.6047, + "theoretical_loss": 3.8331718006619884, + "tokens_seen": 655818752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046740220661986, + "loss": 1.5241, + "theoretical_loss": 3.833132747915004, + "tokens_seen": 655884288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046639919759278, + "loss": 1.4768, + "theoretical_loss": 3.8330937001624568, + "tokens_seen": 655949824 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404653961885657, + "loss": 1.3988, + "theoretical_loss": 3.833054657403208, + "tokens_seen": 656015360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046439317953862, + "loss": 1.5991, + "theoretical_loss": 3.8330156196361207, + "tokens_seen": 656080896 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040463390170511536, + "loss": 1.6194, + "theoretical_loss": 3.8329765868600583, + "tokens_seen": 656146432 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040462387161484454, + "loss": 1.5677, + "theoretical_loss": 3.832937559073884, + "tokens_seen": 656211968 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046138415245738, + "loss": 1.5595, + "theoretical_loss": 3.8328985362764616, + "tokens_seen": 656277504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004046038114343029, + "loss": 1.7165, + "theoretical_loss": 3.8328595184666554, + "tokens_seen": 656343040 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040459378134403214, + "loss": 1.5619, + "theoretical_loss": 3.8328205056433298, + "tokens_seen": 656408576 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040458375125376127, + "loss": 1.4766, + "theoretical_loss": 3.8327814978053505, + "tokens_seen": 656474112 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004045737211634905, + "loss": 1.5975, + "theoretical_loss": 3.8327424949515816, + "tokens_seen": 656539648 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004045636910732197, + "loss": 1.6513, + "theoretical_loss": 3.83270349708089, + "tokens_seen": 656605184 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040455366098294886, + "loss": 1.5492, + "theoretical_loss": 3.832664504192141, + "tokens_seen": 656670720 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040454363089267804, + "loss": 1.5404, + "theoretical_loss": 3.8326255162842022, + "tokens_seen": 656736256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004045336008024073, + "loss": 1.4725, + "theoretical_loss": 3.832586533355939, + "tokens_seen": 656801792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004045235707121364, + "loss": 1.5181, + "theoretical_loss": 3.832547555406219, + "tokens_seen": 656867328 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040451354062186564, + "loss": 1.4764, + "theoretical_loss": 3.8325085824339102, + "tokens_seen": 656932864 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040450351053159477, + "loss": 1.5165, + "theoretical_loss": 3.8324696144378803, + "tokens_seen": 656998400 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404493480441324, + "loss": 1.4161, + "theoretical_loss": 3.8324306514169977, + "tokens_seen": 657063936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004044834503510532, + "loss": 1.5267, + "theoretical_loss": 3.832391693370131, + "tokens_seen": 657129472 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040447342026078237, + "loss": 1.6134, + "theoretical_loss": 3.832352740296149, + "tokens_seen": 657195008 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040446339017051155, + "loss": 1.5679, + "theoretical_loss": 3.8323137921939217, + "tokens_seen": 657260544 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040445336008024073, + "loss": 1.428, + "theoretical_loss": 3.832274849062318, + "tokens_seen": 657326080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004044433299899699, + "loss": 1.5044, + "theoretical_loss": 3.8322359109002093, + "tokens_seen": 657391616 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040443329989969915, + "loss": 1.4919, + "theoretical_loss": 3.8321969777064653, + "tokens_seen": 657457152 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040442326980942827, + "loss": 1.5071, + "theoretical_loss": 3.8321580494799568, + "tokens_seen": 657522688 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004044132397191575, + "loss": 1.6447, + "theoretical_loss": 3.8321191262195553, + "tokens_seen": 657588224 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040440320962888663, + "loss": 1.5762, + "theoretical_loss": 3.832080207924133, + "tokens_seen": 657653760 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040439317953861587, + "loss": 1.5101, + "theoretical_loss": 3.832041294592561, + "tokens_seen": 657719296 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040438314944834505, + "loss": 1.5038, + "theoretical_loss": 3.832002386223712, + "tokens_seen": 657784832 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040437311935807423, + "loss": 1.4594, + "theoretical_loss": 3.831963482816459, + "tokens_seen": 657850368 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004043630892678034, + "loss": 1.547, + "theoretical_loss": 3.8319245843696743, + "tokens_seen": 657915904 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040435305917753265, + "loss": 1.587, + "theoretical_loss": 3.8318856908822325, + "tokens_seen": 657981440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004043430290872618, + "loss": 1.507, + "theoretical_loss": 3.831846802353007, + "tokens_seen": 658046976 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404332998996991, + "loss": 1.5955, + "theoretical_loss": 3.8318079187808713, + "tokens_seen": 658112512 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040432296890672014, + "loss": 1.5781, + "theoretical_loss": 3.831769040164701, + "tokens_seen": 658178048 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040431293881644937, + "loss": 1.4909, + "theoretical_loss": 3.8317301665033705, + "tokens_seen": 658243584 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040430290872617855, + "loss": 1.5836, + "theoretical_loss": 3.8316912977957553, + "tokens_seen": 658309120 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040429287863590773, + "loss": 1.5198, + "theoretical_loss": 3.8316524340407314, + "tokens_seen": 658374656 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042828485456369, + "loss": 1.5251, + "theoretical_loss": 3.8316135752371743, + "tokens_seen": 658440192 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042728184553661, + "loss": 1.5287, + "theoretical_loss": 3.8315747213839604, + "tokens_seen": 658505728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042627883650953, + "loss": 1.4784, + "theoretical_loss": 3.831535872479967, + "tokens_seen": 658571264 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 245288, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4219818115234375, + "objective/train/theoretical_loss": 3.831497028524071, + "objective/train/tokens_used": 679096800, + "theoretical_loss": 3.831497028524071, + "tokens_seen": 658636800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042527582748245, + "loss": 1.5046, + "theoretical_loss": 3.831497028524071, + "tokens_seen": 658636800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040424272818455364, + "loss": 1.5212, + "theoretical_loss": 3.83145818951515, + "tokens_seen": 658702336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042326980942829, + "loss": 1.4871, + "theoretical_loss": 3.8314193554520806, + "tokens_seen": 658767872 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404222668004012, + "loss": 1.5401, + "theoretical_loss": 3.831380526333743, + "tokens_seen": 658833408 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040421263791374124, + "loss": 1.5385, + "theoretical_loss": 3.8313417021590155, + "tokens_seen": 658898944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004042026078234704, + "loss": 1.5719, + "theoretical_loss": 3.831302882926776, + "tokens_seen": 658964480 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041925777331996, + "loss": 1.4927, + "theoretical_loss": 3.831264068635905, + "tokens_seen": 659030016 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041825476429288, + "loss": 1.5312, + "theoretical_loss": 3.831225259285281, + "tokens_seen": 659095552 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404172517552658, + "loss": 1.5406, + "theoretical_loss": 3.8311864548737855, + "tokens_seen": 659161088 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040416248746238714, + "loss": 1.5721, + "theoretical_loss": 3.831147655400298, + "tokens_seen": 659226624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041524573721164, + "loss": 1.5872, + "theoretical_loss": 3.831108860863699, + "tokens_seen": 659292160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041424272818455, + "loss": 1.5869, + "theoretical_loss": 3.8310700712628707, + "tokens_seen": 659357696 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040413239719157474, + "loss": 1.5855, + "theoretical_loss": 3.8310312865966942, + "tokens_seen": 659423232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041223671013039, + "loss": 1.5551, + "theoretical_loss": 3.830992506864051, + "tokens_seen": 659488768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041123370110331, + "loss": 1.5861, + "theoretical_loss": 3.830953732063824, + "tokens_seen": 659554304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004041023069207623, + "loss": 1.5982, + "theoretical_loss": 3.830914962194895, + "tokens_seen": 659619840 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040409227683049147, + "loss": 1.6327, + "theoretical_loss": 3.830876197256148, + "tokens_seen": 659685376 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040408224674022065, + "loss": 1.5356, + "theoretical_loss": 3.8308374372464655, + "tokens_seen": 659750912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004040722166499499, + "loss": 1.5454, + "theoretical_loss": 3.830798682164732, + "tokens_seen": 659816448 + }, + { + "epoch": 0.2, + "learning_rate": 0.000404062186559679, + "loss": 1.5117, + "theoretical_loss": 3.8307599320098307, + "tokens_seen": 659881984 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040405215646940824, + "loss": 1.6013, + "theoretical_loss": 3.8307211867806465, + "tokens_seen": 659947520 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040404212637913737, + "loss": 1.4368, + "theoretical_loss": 3.8306824464760645, + "tokens_seen": 660013056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004040320962888666, + "loss": 1.5076, + "theoretical_loss": 3.83064371109497, + "tokens_seen": 660078592 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004040220661985958, + "loss": 1.5676, + "theoretical_loss": 3.830604980636248, + "tokens_seen": 660144128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040401203610832497, + "loss": 1.5794, + "theoretical_loss": 3.8305662550987845, + "tokens_seen": 660209664 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040400200601805415, + "loss": 1.5568, + "theoretical_loss": 3.8305275344814658, + "tokens_seen": 660275200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004039919759277834, + "loss": 1.4453, + "theoretical_loss": 3.8304888187831785, + "tokens_seen": 660340736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004039819458375125, + "loss": 1.5279, + "theoretical_loss": 3.8304501080028097, + "tokens_seen": 660406272 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040397191574724175, + "loss": 1.5774, + "theoretical_loss": 3.830411402139247, + "tokens_seen": 660471808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004039618856569709, + "loss": 1.5346, + "theoretical_loss": 3.830372701191377, + "tokens_seen": 660537344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004039518555667001, + "loss": 1.5166, + "theoretical_loss": 3.8303340051580896, + "tokens_seen": 660602880 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004039418254764293, + "loss": 1.5708, + "theoretical_loss": 3.8302953140382714, + "tokens_seen": 660668416 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040393179538615847, + "loss": 1.4552, + "theoretical_loss": 3.8302566278308126, + "tokens_seen": 660733952 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040392176529588765, + "loss": 1.462, + "theoretical_loss": 3.830217946534602, + "tokens_seen": 660799488 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040391173520561683, + "loss": 1.5004, + "theoretical_loss": 3.830179270148528, + "tokens_seen": 660865024 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040390170511534607, + "loss": 1.682, + "theoretical_loss": 3.8301405986714823, + "tokens_seen": 660930560 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040389167502507525, + "loss": 1.5385, + "theoretical_loss": 3.8301019321023544, + "tokens_seen": 660996096 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040388164493480443, + "loss": 1.5329, + "theoretical_loss": 3.830063270440034, + "tokens_seen": 661061632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004038716148445336, + "loss": 1.5072, + "theoretical_loss": 3.8300246136834133, + "tokens_seen": 661127168 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040386158475426285, + "loss": 1.419, + "theoretical_loss": 3.829985961831383, + "tokens_seen": 661192704 + }, + { + "epoch": 0.2, + "learning_rate": 0.000403851554663992, + "loss": 1.5099, + "theoretical_loss": 3.829947314882835, + "tokens_seen": 661258240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004038415245737212, + "loss": 1.4448, + "theoretical_loss": 3.829908672836661, + "tokens_seen": 661323776 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040383149448345034, + "loss": 1.5646, + "theoretical_loss": 3.829870035691754, + "tokens_seen": 661389312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040382146439317957, + "loss": 1.5881, + "theoretical_loss": 3.829831403447007, + "tokens_seen": 661454848 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040381143430290875, + "loss": 1.4609, + "theoretical_loss": 3.8297927761013115, + "tokens_seen": 661520384 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040380140421263794, + "loss": 1.5772, + "theoretical_loss": 3.829754153653563, + "tokens_seen": 661585920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037913741223671, + "loss": 1.5546, + "theoretical_loss": 3.829715536102654, + "tokens_seen": 661651456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037813440320963, + "loss": 1.5373, + "theoretical_loss": 3.8296769234474795, + "tokens_seen": 661716992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037713139418255, + "loss": 1.4539, + "theoretical_loss": 3.829638315686933, + "tokens_seen": 661782528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037612838515547, + "loss": 1.6024, + "theoretical_loss": 3.8295997128199106, + "tokens_seen": 661848064 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 246530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5785856246948242, + "objective/train/theoretical_loss": 3.8295611148453075, + "objective/train/tokens_used": 682373600, + "theoretical_loss": 3.8295611148453075, + "tokens_seen": 661913600 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040375125376128384, + "loss": 1.6456, + "theoretical_loss": 3.8295611148453075, + "tokens_seen": 661913600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037412236710131, + "loss": 1.4744, + "theoretical_loss": 3.8295225217620183, + "tokens_seen": 661979136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037311935807422, + "loss": 1.5288, + "theoretical_loss": 3.82948393356894, + "tokens_seen": 662044672 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040372116349047144, + "loss": 1.5823, + "theoretical_loss": 3.829445350264969, + "tokens_seen": 662110208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037111334002006, + "loss": 1.6247, + "theoretical_loss": 3.8294067718490012, + "tokens_seen": 662175744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004037011033099298, + "loss": 1.6375, + "theoretical_loss": 3.829368198319934, + "tokens_seen": 662241280 + }, + { + "epoch": 0.2, + "learning_rate": 0.000403691073219659, + "loss": 1.5443, + "theoretical_loss": 3.829329629676665, + "tokens_seen": 662306816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036810431293882, + "loss": 1.5419, + "theoretical_loss": 3.829291065918092, + "tokens_seen": 662372352 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040367101303911734, + "loss": 1.4521, + "theoretical_loss": 3.8292525070431136, + "tokens_seen": 662437888 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036609829488466, + "loss": 1.4937, + "theoretical_loss": 3.829213953050627, + "tokens_seen": 662503424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036509528585757, + "loss": 1.4584, + "theoretical_loss": 3.8291754039395323, + "tokens_seen": 662568960 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040364092276830494, + "loss": 1.4021, + "theoretical_loss": 3.8291368597087283, + "tokens_seen": 662634496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036308926780341, + "loss": 1.5078, + "theoretical_loss": 3.8290983203571147, + "tokens_seen": 662700032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036208625877633, + "loss": 1.62, + "theoretical_loss": 3.829059785883591, + "tokens_seen": 662765568 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004036108324974925, + "loss": 1.4497, + "theoretical_loss": 3.8290212562870583, + "tokens_seen": 662831104 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040360080240722167, + "loss": 1.3835, + "theoretical_loss": 3.8289827315664158, + "tokens_seen": 662896640 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040359077231695085, + "loss": 1.5276, + "theoretical_loss": 3.8289442117205663, + "tokens_seen": 662962176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004035807422266801, + "loss": 1.5756, + "theoretical_loss": 3.82890569674841, + "tokens_seen": 663027712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004035707121364092, + "loss": 1.5339, + "theoretical_loss": 3.828867186648849, + "tokens_seen": 663093248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040356068204613844, + "loss": 1.689, + "theoretical_loss": 3.8288286814207853, + "tokens_seen": 663158784 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040355065195586757, + "loss": 1.5466, + "theoretical_loss": 3.828790181063121, + "tokens_seen": 663224320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004035406218655968, + "loss": 1.4979, + "theoretical_loss": 3.828751685574759, + "tokens_seen": 663289856 + }, + { + "epoch": 0.2, + "learning_rate": 0.000403530591775326, + "loss": 1.5115, + "theoretical_loss": 3.828713194954604, + "tokens_seen": 663355392 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040352056168505517, + "loss": 1.4988, + "theoretical_loss": 3.8286747092015574, + "tokens_seen": 663420928 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040351053159478435, + "loss": 1.5689, + "theoretical_loss": 3.828636228314523, + "tokens_seen": 663486464 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004035005015045136, + "loss": 1.5901, + "theoretical_loss": 3.8285977522924073, + "tokens_seen": 663552000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004034904714142427, + "loss": 1.632, + "theoretical_loss": 3.8285592811341127, + "tokens_seen": 663617536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040348044132397195, + "loss": 1.492, + "theoretical_loss": 3.8285208148385443, + "tokens_seen": 663683072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004034704112337011, + "loss": 1.4763, + "theoretical_loss": 3.8284823534046084, + "tokens_seen": 663748608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004034603811434303, + "loss": 1.5144, + "theoretical_loss": 3.8284438968312102, + "tokens_seen": 663814144 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004034503510531595, + "loss": 1.6428, + "theoretical_loss": 3.828405445117255, + "tokens_seen": 663879680 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040344032096288867, + "loss": 1.5269, + "theoretical_loss": 3.82836699826165, + "tokens_seen": 663945216 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040343029087261785, + "loss": 1.5254, + "theoretical_loss": 3.828328556263302, + "tokens_seen": 664010752 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040342026078234703, + "loss": 1.4888, + "theoretical_loss": 3.8282901191211174, + "tokens_seen": 664076288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004034102306920762, + "loss": 1.5102, + "theoretical_loss": 3.8282516868340037, + "tokens_seen": 664141824 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040340020060180545, + "loss": 1.5899, + "theoretical_loss": 3.8282132594008687, + "tokens_seen": 664207360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004033901705115346, + "loss": 1.5049, + "theoretical_loss": 3.828174836820621, + "tokens_seen": 664272896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004033801404212638, + "loss": 1.6341, + "theoretical_loss": 3.8281364190921683, + "tokens_seen": 664338432 + }, + { + "epoch": 0.2, + "learning_rate": 0.000403370110330993, + "loss": 1.5355, + "theoretical_loss": 3.82809800621442, + "tokens_seen": 664403968 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004033600802407222, + "loss": 1.3927, + "theoretical_loss": 3.8280595981862846, + "tokens_seen": 664469504 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040335005015045136, + "loss": 1.5433, + "theoretical_loss": 3.8280211950066723, + "tokens_seen": 664535040 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040334002006018054, + "loss": 1.5823, + "theoretical_loss": 3.827982796674493, + "tokens_seen": 664600576 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004033299899699097, + "loss": 1.5009, + "theoretical_loss": 3.8279444031886563, + "tokens_seen": 664666112 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040331995987963895, + "loss": 1.5647, + "theoretical_loss": 3.827906014548073, + "tokens_seen": 664731648 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004033099297893681, + "loss": 1.4502, + "theoretical_loss": 3.827867630751654, + "tokens_seen": 664797184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032998996990973, + "loss": 1.5538, + "theoretical_loss": 3.827829251798311, + "tokens_seen": 664862720 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040328986960882644, + "loss": 1.4688, + "theoretical_loss": 3.8277908776869554, + "tokens_seen": 664928256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032798395185557, + "loss": 1.566, + "theoretical_loss": 3.8277525084164994, + "tokens_seen": 664993792 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040326980942828486, + "loss": 1.5431, + "theoretical_loss": 3.8277141439858546, + "tokens_seen": 665059328 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040325977933801404, + "loss": 1.678, + "theoretical_loss": 3.827675784393934, + "tokens_seen": 665124864 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 247583, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6462876796722412, + "objective/train/theoretical_loss": 3.827637429639651, + "objective/train/tokens_used": 685650400, + "theoretical_loss": 3.827637429639651, + "tokens_seen": 665190400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032497492477432, + "loss": 1.569, + "theoretical_loss": 3.827637429639651, + "tokens_seen": 665190400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032397191574724, + "loss": 1.4483, + "theoretical_loss": 3.827599079721918, + "tokens_seen": 665255936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032296890672016, + "loss": 1.5527, + "theoretical_loss": 3.8275607346396505, + "tokens_seen": 665321472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004032196589769308, + "loss": 1.6215, + "theoretical_loss": 3.827522394391761, + "tokens_seen": 665387008 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040320962888665995, + "loss": 1.4969, + "theoretical_loss": 3.8274840589771655, + "tokens_seen": 665452544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031995987963892, + "loss": 1.46, + "theoretical_loss": 3.8274457283947765, + "tokens_seen": 665518080 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040318956870611836, + "loss": 1.5099, + "theoretical_loss": 3.8274074026435114, + "tokens_seen": 665583616 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040317953861584754, + "loss": 1.5029, + "theoretical_loss": 3.8273690817222845, + "tokens_seen": 665649152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031695085255767, + "loss": 1.3984, + "theoretical_loss": 3.8273307656300117, + "tokens_seen": 665714688 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031594784353059, + "loss": 1.463, + "theoretical_loss": 3.8272924543656095, + "tokens_seen": 665780224 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040314944834503514, + "loss": 1.5329, + "theoretical_loss": 3.8272541479279942, + "tokens_seen": 665845760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031394182547643, + "loss": 1.5295, + "theoretical_loss": 3.827215846316083, + "tokens_seen": 665911296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031293881644935, + "loss": 1.4999, + "theoretical_loss": 3.827177549528793, + "tokens_seen": 665976832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004031193580742227, + "loss": 1.5258, + "theoretical_loss": 3.8271392575650416, + "tokens_seen": 666042368 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040310932798395187, + "loss": 1.4935, + "theoretical_loss": 3.827100970423747, + "tokens_seen": 666107904 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040309929789368105, + "loss": 1.5639, + "theoretical_loss": 3.827062688103827, + "tokens_seen": 666173440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004030892678034103, + "loss": 1.5406, + "theoretical_loss": 3.827024410604201, + "tokens_seen": 666238976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004030792377131394, + "loss": 1.5394, + "theoretical_loss": 3.8269861379237877, + "tokens_seen": 666304512 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040306920762286864, + "loss": 1.4906, + "theoretical_loss": 3.826947870061506, + "tokens_seen": 666370048 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040305917753259777, + "loss": 1.5227, + "theoretical_loss": 3.826909607016276, + "tokens_seen": 666435584 + }, + { + "epoch": 0.2, + "learning_rate": 0.000403049147442327, + "loss": 1.5493, + "theoretical_loss": 3.8268713487870176, + "tokens_seen": 666501120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004030391173520562, + "loss": 1.4297, + "theoretical_loss": 3.8268330953726513, + "tokens_seen": 666566656 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040302908726178537, + "loss": 1.5484, + "theoretical_loss": 3.826794846772098, + "tokens_seen": 666632192 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040301905717151455, + "loss": 1.6141, + "theoretical_loss": 3.826756602984278, + "tokens_seen": 666697728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004030090270812438, + "loss": 1.4851, + "theoretical_loss": 3.826718364008114, + "tokens_seen": 666763264 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004029989969909729, + "loss": 1.5915, + "theoretical_loss": 3.826680129842526, + "tokens_seen": 666828800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040298896690070215, + "loss": 1.567, + "theoretical_loss": 3.826641900486438, + "tokens_seen": 666894336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004029789368104313, + "loss": 1.4926, + "theoretical_loss": 3.8266036759387716, + "tokens_seen": 666959872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004029689067201605, + "loss": 1.4501, + "theoretical_loss": 3.8265654561984492, + "tokens_seen": 667025408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004029588766298897, + "loss": 1.6439, + "theoretical_loss": 3.8265272412643947, + "tokens_seen": 667090944 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040294884653961887, + "loss": 1.4141, + "theoretical_loss": 3.8264890311355315, + "tokens_seen": 667156480 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040293881644934805, + "loss": 1.4502, + "theoretical_loss": 3.8264508258107828, + "tokens_seen": 667222016 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040292878635907723, + "loss": 1.5498, + "theoretical_loss": 3.8264126252890738, + "tokens_seen": 667287552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004029187562688064, + "loss": 1.5026, + "theoretical_loss": 3.826374429569328, + "tokens_seen": 667353088 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040290872617853565, + "loss": 1.5128, + "theoretical_loss": 3.8263362386504713, + "tokens_seen": 667418624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028986960882648, + "loss": 1.5689, + "theoretical_loss": 3.826298052531428, + "tokens_seen": 667484160 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402888665997994, + "loss": 1.5639, + "theoretical_loss": 3.8262598712111244, + "tokens_seen": 667549696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028786359077232, + "loss": 1.5081, + "theoretical_loss": 3.8262216946884866, + "tokens_seen": 667615232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028686058174524, + "loss": 1.5236, + "theoretical_loss": 3.8261835229624395, + "tokens_seen": 667680768 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040285857572718156, + "loss": 1.4695, + "theoretical_loss": 3.8261453560319114, + "tokens_seen": 667746304 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040284854563691074, + "loss": 1.4256, + "theoretical_loss": 3.8261071938958278, + "tokens_seen": 667811840 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028385155466399, + "loss": 1.5827, + "theoretical_loss": 3.826069036553118, + "tokens_seen": 667877376 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040282848545636915, + "loss": 1.505, + "theoretical_loss": 3.8260308840027073, + "tokens_seen": 667942912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028184553660983, + "loss": 1.5311, + "theoretical_loss": 3.825992736243525, + "tokens_seen": 668008448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004028084252758275, + "loss": 1.4591, + "theoretical_loss": 3.8259545932745, + "tokens_seen": 668073984 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040279839518555664, + "loss": 1.5048, + "theoretical_loss": 3.8259164550945597, + "tokens_seen": 668139520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004027883650952859, + "loss": 1.6054, + "theoretical_loss": 3.8258783217026338, + "tokens_seen": 668205056 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040277833500501506, + "loss": 1.5105, + "theoretical_loss": 3.825840193097652, + "tokens_seen": 668270592 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040276830491474424, + "loss": 1.5329, + "theoretical_loss": 3.8258020692785433, + "tokens_seen": 668336128 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004027582748244734, + "loss": 1.5099, + "theoretical_loss": 3.825763950244238, + "tokens_seen": 668401664 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 248744, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.56922447681427, + "objective/train/theoretical_loss": 3.8257258359936666, + "objective/train/tokens_used": 688927200, + "theoretical_loss": 3.8257258359936666, + "tokens_seen": 668467200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004027482447342026, + "loss": 1.6198, + "theoretical_loss": 3.8257258359936666, + "tokens_seen": 668467200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004027382146439318, + "loss": 1.4714, + "theoretical_loss": 3.82568772652576, + "tokens_seen": 668532736 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402728184553661, + "loss": 1.5795, + "theoretical_loss": 3.8256496218394496, + "tokens_seen": 668598272 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040271815446339015, + "loss": 1.4817, + "theoretical_loss": 3.825611521933666, + "tokens_seen": 668663808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004027081243731194, + "loss": 1.4393, + "theoretical_loss": 3.8255734268073414, + "tokens_seen": 668729344 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040269809428284856, + "loss": 1.5832, + "theoretical_loss": 3.8255353364594082, + "tokens_seen": 668794880 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040268806419257774, + "loss": 1.4423, + "theoretical_loss": 3.825497250888799, + "tokens_seen": 668860416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004026780341023069, + "loss": 1.5498, + "theoretical_loss": 3.8254591700944456, + "tokens_seen": 668925952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004026680040120361, + "loss": 1.4648, + "theoretical_loss": 3.825421094075282, + "tokens_seen": 668991488 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004026579739217653, + "loss": 1.4737, + "theoretical_loss": 3.8253830228302417, + "tokens_seen": 669057024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004026479438314945, + "loss": 1.519, + "theoretical_loss": 3.825344956358258, + "tokens_seen": 669122560 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040263791374122365, + "loss": 1.566, + "theoretical_loss": 3.825306894658266, + "tokens_seen": 669188096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004026278836509529, + "loss": 1.4855, + "theoretical_loss": 3.8252688377292, + "tokens_seen": 669253632 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402617853560682, + "loss": 1.5506, + "theoretical_loss": 3.8252307855699943, + "tokens_seen": 669319168 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040260782347041125, + "loss": 1.5087, + "theoretical_loss": 3.8251927381795845, + "tokens_seen": 669384704 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040259779338014043, + "loss": 1.5074, + "theoretical_loss": 3.825154695556906, + "tokens_seen": 669450240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004025877632898696, + "loss": 1.503, + "theoretical_loss": 3.825116657700894, + "tokens_seen": 669515776 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004025777331995988, + "loss": 1.6027, + "theoretical_loss": 3.8250786246104864, + "tokens_seen": 669581312 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040256770310932797, + "loss": 1.5051, + "theoretical_loss": 3.825040596284619, + "tokens_seen": 669646848 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040255767301905715, + "loss": 1.5115, + "theoretical_loss": 3.8250025727222283, + "tokens_seen": 669712384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004025476429287864, + "loss": 1.585, + "theoretical_loss": 3.8249645539222525, + "tokens_seen": 669777920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004025376128385155, + "loss": 1.4876, + "theoretical_loss": 3.824926539883628, + "tokens_seen": 669843456 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040252758274824475, + "loss": 1.4795, + "theoretical_loss": 3.8248885306052935, + "tokens_seen": 669908992 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040251755265797393, + "loss": 1.5956, + "theoretical_loss": 3.8248505260861876, + "tokens_seen": 669974528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004025075225677031, + "loss": 1.4887, + "theoretical_loss": 3.824812526325248, + "tokens_seen": 670040064 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024974924774323, + "loss": 1.5121, + "theoretical_loss": 3.824774531321414, + "tokens_seen": 670105600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024874623871615, + "loss": 1.4603, + "theoretical_loss": 3.8247365410736256, + "tokens_seen": 670171136 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040247743229689066, + "loss": 1.6015, + "theoretical_loss": 3.824698555580822, + "tokens_seen": 670236672 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024674022066199, + "loss": 1.5463, + "theoretical_loss": 3.8246605748419427, + "tokens_seen": 670302208 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402457372116349, + "loss": 1.4786, + "theoretical_loss": 3.824622598855929, + "tokens_seen": 670367744 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040244734202607825, + "loss": 1.3182, + "theoretical_loss": 3.824584627621721, + "tokens_seen": 670433280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024373119358074, + "loss": 1.4229, + "theoretical_loss": 3.8245466611382595, + "tokens_seen": 670498816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024272818455366, + "loss": 1.4925, + "theoretical_loss": 3.824508699404486, + "tokens_seen": 670564352 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004024172517552658, + "loss": 1.5761, + "theoretical_loss": 3.8244707424193427, + "tokens_seen": 670629888 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402407221664995, + "loss": 1.5157, + "theoretical_loss": 3.8244327901817705, + "tokens_seen": 670695424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023971915747242, + "loss": 1.5636, + "theoretical_loss": 3.8243948426907135, + "tokens_seen": 670760960 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023871614844534, + "loss": 1.5437, + "theoretical_loss": 3.824356899945113, + "tokens_seen": 670826496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023771313941826, + "loss": 1.5781, + "theoretical_loss": 3.824318961943912, + "tokens_seen": 670892032 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040236710130391176, + "loss": 1.4606, + "theoretical_loss": 3.8242810286860545, + "tokens_seen": 670957568 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040235707121364094, + "loss": 1.4934, + "theoretical_loss": 3.8242431001704844, + "tokens_seen": 671023104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023470411233701, + "loss": 1.5921, + "theoretical_loss": 3.8242051763961458, + "tokens_seen": 671088640 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040233701103309935, + "loss": 1.5199, + "theoretical_loss": 3.824167257361982, + "tokens_seen": 671154176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023269809428285, + "loss": 1.4381, + "theoretical_loss": 3.8241293430669385, + "tokens_seen": 671219712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004023169508525577, + "loss": 1.4791, + "theoretical_loss": 3.824091433509961, + "tokens_seen": 671285248 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040230692076228684, + "loss": 1.4817, + "theoretical_loss": 3.824053528689994, + "tokens_seen": 671350784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004022968906720161, + "loss": 1.5504, + "theoretical_loss": 3.8240156286059834, + "tokens_seen": 671416320 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040228686058174526, + "loss": 1.5437, + "theoretical_loss": 3.823977733256875, + "tokens_seen": 671481856 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040227683049147444, + "loss": 1.5691, + "theoretical_loss": 3.8239398426416162, + "tokens_seen": 671547392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004022668004012036, + "loss": 1.5923, + "theoretical_loss": 3.823901956759153, + "tokens_seen": 671612928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004022567703109328, + "loss": 1.5684, + "theoretical_loss": 3.8238640756084328, + "tokens_seen": 671678464 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 249880, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4036436080932617, + "objective/train/theoretical_loss": 3.823826199188403, + "objective/train/tokens_used": 692204000, + "theoretical_loss": 3.823826199188403, + "tokens_seen": 671744000 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402246740220662, + "loss": 1.4452, + "theoretical_loss": 3.823826199188403, + "tokens_seen": 671744000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004022367101303912, + "loss": 1.396, + "theoretical_loss": 3.8237883274980113, + "tokens_seen": 671809536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040222668004012035, + "loss": 1.4578, + "theoretical_loss": 3.823750460536206, + "tokens_seen": 671875072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004022166499498496, + "loss": 1.4385, + "theoretical_loss": 3.8237125983019347, + "tokens_seen": 671940608 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040220661985957876, + "loss": 1.5652, + "theoretical_loss": 3.8236747407941474, + "tokens_seen": 672006144 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040219658976930794, + "loss": 1.4945, + "theoretical_loss": 3.8236368880117926, + "tokens_seen": 672071680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021865596790371, + "loss": 1.419, + "theoretical_loss": 3.82359903995382, + "tokens_seen": 672137216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021765295887663, + "loss": 1.4835, + "theoretical_loss": 3.823561196619179, + "tokens_seen": 672202752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021664994984955, + "loss": 1.5448, + "theoretical_loss": 3.82352335800682, + "tokens_seen": 672268288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021564694082247, + "loss": 1.5353, + "theoretical_loss": 3.8234855241156938, + "tokens_seen": 672333824 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040214643931795385, + "loss": 1.4578, + "theoretical_loss": 3.8234476949447505, + "tokens_seen": 672399360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021364092276831, + "loss": 1.5606, + "theoretical_loss": 3.8234098704929416, + "tokens_seen": 672464896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004021263791374122, + "loss": 1.4304, + "theoretical_loss": 3.823372050759218, + "tokens_seen": 672530432 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040211634904714145, + "loss": 1.56, + "theoretical_loss": 3.823334235742533, + "tokens_seen": 672595968 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040210631895687063, + "loss": 1.5339, + "theoretical_loss": 3.8232964254418373, + "tokens_seen": 672661504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004020962888665998, + "loss": 1.5456, + "theoretical_loss": 3.823258619856084, + "tokens_seen": 672727040 + }, + { + "epoch": 0.2, + "learning_rate": 0.000402086258776329, + "loss": 1.5878, + "theoretical_loss": 3.8232208189842254, + "tokens_seen": 672792576 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040207622868605817, + "loss": 1.4391, + "theoretical_loss": 3.8231830228252157, + "tokens_seen": 672858112 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040206619859578735, + "loss": 1.5797, + "theoretical_loss": 3.8231452313780068, + "tokens_seen": 672923648 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004020561685055166, + "loss": 1.4787, + "theoretical_loss": 3.823107444641554, + "tokens_seen": 672989184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004020461384152457, + "loss": 1.5559, + "theoretical_loss": 3.8230696626148113, + "tokens_seen": 673054720 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040203610832497495, + "loss": 1.467, + "theoretical_loss": 3.823031885296732, + "tokens_seen": 673120256 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040202607823470413, + "loss": 1.6441, + "theoretical_loss": 3.822994112686272, + "tokens_seen": 673185792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004020160481444333, + "loss": 1.5641, + "theoretical_loss": 3.8229563447823858, + "tokens_seen": 673251328 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004020060180541625, + "loss": 1.4195, + "theoretical_loss": 3.8229185815840294, + "tokens_seen": 673316864 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019959879638917, + "loss": 1.5097, + "theoretical_loss": 3.8228808230901583, + "tokens_seen": 673382400 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040198595787362086, + "loss": 1.5109, + "theoretical_loss": 3.822843069299729, + "tokens_seen": 673447936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019759277833501, + "loss": 1.551, + "theoretical_loss": 3.822805320211698, + "tokens_seen": 673513472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019658976930792, + "loss": 1.5592, + "theoretical_loss": 3.8227675758250212, + "tokens_seen": 673579008 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040195586760280845, + "loss": 1.5261, + "theoretical_loss": 3.822729836138657, + "tokens_seen": 673644544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019458375125376, + "loss": 1.4248, + "theoretical_loss": 3.8226921011515618, + "tokens_seen": 673710080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019358074222668, + "loss": 1.4446, + "theoretical_loss": 3.822654370862694, + "tokens_seen": 673775616 + }, + { + "epoch": 0.2, + "learning_rate": 0.000401925777331996, + "loss": 1.475, + "theoretical_loss": 3.8226166452710117, + "tokens_seen": 673841152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004019157472417252, + "loss": 1.5766, + "theoretical_loss": 3.822578924375473, + "tokens_seen": 673906688 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040190571715145436, + "loss": 1.538, + "theoretical_loss": 3.8225412081750374, + "tokens_seen": 673972224 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018956870611836, + "loss": 1.4886, + "theoretical_loss": 3.8225034966686637, + "tokens_seen": 674037760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018856569709127, + "loss": 1.5465, + "theoretical_loss": 3.822465789855311, + "tokens_seen": 674103296 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040187562688064196, + "loss": 1.5721, + "theoretical_loss": 3.82242808773394, + "tokens_seen": 674168832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018655967903711, + "loss": 1.5195, + "theoretical_loss": 3.82239039030351, + "tokens_seen": 674234368 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018555667001003, + "loss": 1.5136, + "theoretical_loss": 3.822352697562982, + "tokens_seen": 674299904 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018455366098295, + "loss": 1.3985, + "theoretical_loss": 3.8223150095113163, + "tokens_seen": 674365440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018355065195587, + "loss": 1.5686, + "theoretical_loss": 3.822277326147475, + "tokens_seen": 674430976 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040182547642928786, + "loss": 1.4601, + "theoretical_loss": 3.822239647470418, + "tokens_seen": 674496512 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040181544633901704, + "loss": 1.4675, + "theoretical_loss": 3.822201973479108, + "tokens_seen": 674562048 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004018054162487462, + "loss": 1.5063, + "theoretical_loss": 3.8221643041725075, + "tokens_seen": 674627584 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040179538615847546, + "loss": 1.529, + "theoretical_loss": 3.8221266395495785, + "tokens_seen": 674693120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004017853560682046, + "loss": 1.5737, + "theoretical_loss": 3.822088979609284, + "tokens_seen": 674758656 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004017753259779338, + "loss": 1.424, + "theoretical_loss": 3.8220513243505865, + "tokens_seen": 674824192 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040176529588766295, + "loss": 1.4374, + "theoretical_loss": 3.8220136737724504, + "tokens_seen": 674889728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004017552657973922, + "loss": 1.4119, + "theoretical_loss": 3.8219760278738386, + "tokens_seen": 674955264 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 251058, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.511181116104126, + "objective/train/theoretical_loss": 3.821938386653716, + "objective/train/tokens_used": 695480800, + "theoretical_loss": 3.821938386653716, + "tokens_seen": 675020800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040174523570712137, + "loss": 1.4834, + "theoretical_loss": 3.821938386653716, + "tokens_seen": 675020800 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040173520561685055, + "loss": 1.472, + "theoretical_loss": 3.8219007501110465, + "tokens_seen": 675086336 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040172517552657973, + "loss": 1.5175, + "theoretical_loss": 3.8218631182447944, + "tokens_seen": 675151872 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040171514543630896, + "loss": 1.5105, + "theoretical_loss": 3.821825491053926, + "tokens_seen": 675217408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004017051153460381, + "loss": 1.455, + "theoretical_loss": 3.821787868537406, + "tokens_seen": 675282944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004016950852557673, + "loss": 1.506, + "theoretical_loss": 3.8217502506942003, + "tokens_seen": 675348480 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040168505516549645, + "loss": 1.4501, + "theoretical_loss": 3.821712637523275, + "tokens_seen": 675414016 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004016750250752257, + "loss": 1.4552, + "theoretical_loss": 3.8216750290235963, + "tokens_seen": 675479552 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040166499498495487, + "loss": 1.4741, + "theoretical_loss": 3.8216374251941314, + "tokens_seen": 675545088 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040165496489468405, + "loss": 1.5158, + "theoretical_loss": 3.8215998260338466, + "tokens_seen": 675610624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004016449348044133, + "loss": 1.5776, + "theoretical_loss": 3.8215622315417104, + "tokens_seen": 675676160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004016349047141424, + "loss": 1.5212, + "theoretical_loss": 3.8215246417166893, + "tokens_seen": 675741696 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040162487462387165, + "loss": 1.4467, + "theoretical_loss": 3.821487056557752, + "tokens_seen": 675807232 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040161484453360083, + "loss": 1.4659, + "theoretical_loss": 3.8214494760638673, + "tokens_seen": 675872768 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040160481444333, + "loss": 1.4359, + "theoretical_loss": 3.8214119002340032, + "tokens_seen": 675938304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004015947843530592, + "loss": 1.5462, + "theoretical_loss": 3.821374329067129, + "tokens_seen": 676003840 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040158475426278837, + "loss": 1.463, + "theoretical_loss": 3.8213367625622143, + "tokens_seen": 676069376 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040157472417251755, + "loss": 1.4381, + "theoretical_loss": 3.8212992007182285, + "tokens_seen": 676134912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004015646940822468, + "loss": 1.5659, + "theoretical_loss": 3.8212616435341413, + "tokens_seen": 676200448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004015546639919759, + "loss": 1.4986, + "theoretical_loss": 3.821224091008924, + "tokens_seen": 676265984 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040154463390170515, + "loss": 1.5129, + "theoretical_loss": 3.821186543141547, + "tokens_seen": 676331520 + }, + { + "epoch": 0.2, + "learning_rate": 0.00040153460381143433, + "loss": 1.493, + "theoretical_loss": 3.8211489999309807, + "tokens_seen": 676397056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0004015245737211635, + "loss": 1.5298, + "theoretical_loss": 3.821111461376196, + "tokens_seen": 676462592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004015145436308927, + "loss": 1.5444, + "theoretical_loss": 3.8210739274761663, + "tokens_seen": 676528128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004015045135406219, + "loss": 1.4936, + "theoretical_loss": 3.8210363982298623, + "tokens_seen": 676593664 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040149448345035106, + "loss": 1.5532, + "theoretical_loss": 3.820998873636257, + "tokens_seen": 676659200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014844533600803, + "loss": 1.5299, + "theoretical_loss": 3.8209613536943223, + "tokens_seen": 676724736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014744232698094, + "loss": 1.4314, + "theoretical_loss": 3.8209238384030315, + "tokens_seen": 676790272 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040146439317953865, + "loss": 1.5374, + "theoretical_loss": 3.820886327761358, + "tokens_seen": 676855808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014543630892678, + "loss": 1.5441, + "theoretical_loss": 3.820848821768275, + "tokens_seen": 676921344 + }, + { + "epoch": 0.21, + "learning_rate": 0.000401444332998997, + "loss": 1.3666, + "theoretical_loss": 3.8208113204227576, + "tokens_seen": 676986880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014343029087262, + "loss": 1.4715, + "theoretical_loss": 3.8207738237237785, + "tokens_seen": 677052416 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014242728184554, + "loss": 1.509, + "theoretical_loss": 3.8207363316703136, + "tokens_seen": 677117952 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040141424272818456, + "loss": 1.3661, + "theoretical_loss": 3.820698844261337, + "tokens_seen": 677183488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004014042126379138, + "loss": 1.5683, + "theoretical_loss": 3.8206613614958242, + "tokens_seen": 677249024 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013941825476429, + "loss": 1.5244, + "theoretical_loss": 3.820623883372751, + "tokens_seen": 677314560 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040138415245737216, + "loss": 1.5109, + "theoretical_loss": 3.820586409891093, + "tokens_seen": 677380096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013741223671013, + "loss": 1.5234, + "theoretical_loss": 3.820548941049826, + "tokens_seen": 677445632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013640922768305, + "loss": 1.435, + "theoretical_loss": 3.820511476847928, + "tokens_seen": 677511168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013540621865597, + "loss": 1.5124, + "theoretical_loss": 3.8204740172843747, + "tokens_seen": 677576704 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013440320962889, + "loss": 1.514, + "theoretical_loss": 3.820436562358143, + "tokens_seen": 677642240 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040133400200601806, + "loss": 1.5706, + "theoretical_loss": 3.8203991120682117, + "tokens_seen": 677707776 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040132397191574724, + "loss": 1.4779, + "theoretical_loss": 3.8203616664135573, + "tokens_seen": 677773312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004013139418254764, + "loss": 1.5123, + "theoretical_loss": 3.8203242253931586, + "tokens_seen": 677838848 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040130391173520566, + "loss": 1.4098, + "theoretical_loss": 3.820286789005994, + "tokens_seen": 677904384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004012938816449348, + "loss": 1.4821, + "theoretical_loss": 3.820249357251043, + "tokens_seen": 677969920 + }, + { + "epoch": 0.21, + "learning_rate": 0.000401283851554664, + "loss": 1.5042, + "theoretical_loss": 3.8202119301272837, + "tokens_seen": 678035456 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040127382146439315, + "loss": 1.487, + "theoretical_loss": 3.8201745076336957, + "tokens_seen": 678100992 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004012637913741224, + "loss": 1.5653, + "theoretical_loss": 3.8201370897692595, + "tokens_seen": 678166528 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040125376128385157, + "loss": 1.5548, + "theoretical_loss": 3.820099676532955, + "tokens_seen": 678232064 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 252282, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.436499834060669, + "objective/train/theoretical_loss": 3.8200622679237615, + "objective/train/tokens_used": 698757600, + "theoretical_loss": 3.8200622679237615, + "tokens_seen": 678297600 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040124373119358075, + "loss": 1.4203, + "theoretical_loss": 3.8200622679237615, + "tokens_seen": 678297600 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040123370110330993, + "loss": 1.4956, + "theoretical_loss": 3.8200248639406613, + "tokens_seen": 678363136 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040122367101303916, + "loss": 1.3966, + "theoretical_loss": 3.819987464582635, + "tokens_seen": 678428672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004012136409227683, + "loss": 1.5722, + "theoretical_loss": 3.8199500698486633, + "tokens_seen": 678494208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004012036108324975, + "loss": 1.5058, + "theoretical_loss": 3.8199126797377287, + "tokens_seen": 678559744 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040119358074222665, + "loss": 1.6068, + "theoretical_loss": 3.8198752942488134, + "tokens_seen": 678625280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004011835506519559, + "loss": 1.5608, + "theoretical_loss": 3.8198379133808995, + "tokens_seen": 678690816 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040117352056168507, + "loss": 1.5524, + "theoretical_loss": 3.819800537132969, + "tokens_seen": 678756352 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040116349047141425, + "loss": 1.5926, + "theoretical_loss": 3.8197631655040056, + "tokens_seen": 678821888 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040115346038114343, + "loss": 1.5706, + "theoretical_loss": 3.8197257984929927, + "tokens_seen": 678887424 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004011434302908726, + "loss": 1.523, + "theoretical_loss": 3.8196884360989136, + "tokens_seen": 678952960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004011334002006018, + "loss": 1.5716, + "theoretical_loss": 3.819651078320753, + "tokens_seen": 679018496 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040112337011033103, + "loss": 1.5835, + "theoretical_loss": 3.819613725157494, + "tokens_seen": 679084032 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040111334002006016, + "loss": 1.5672, + "theoretical_loss": 3.819576376608122, + "tokens_seen": 679149568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004011033099297894, + "loss": 1.6537, + "theoretical_loss": 3.8195390326716225, + "tokens_seen": 679215104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004010932798395185, + "loss": 1.4022, + "theoretical_loss": 3.8195016933469796, + "tokens_seen": 679280640 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040108324974924775, + "loss": 1.4886, + "theoretical_loss": 3.819464358633179, + "tokens_seen": 679346176 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040107321965897693, + "loss": 1.4651, + "theoretical_loss": 3.819427028529208, + "tokens_seen": 679411712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004010631895687061, + "loss": 1.5, + "theoretical_loss": 3.8193897030340516, + "tokens_seen": 679477248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004010531594784353, + "loss": 1.6477, + "theoretical_loss": 3.819352382146696, + "tokens_seen": 679542784 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040104312938816453, + "loss": 1.4906, + "theoretical_loss": 3.819315065866129, + "tokens_seen": 679608320 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040103309929789366, + "loss": 1.5128, + "theoretical_loss": 3.8192777541913374, + "tokens_seen": 679673856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004010230692076229, + "loss": 1.463, + "theoretical_loss": 3.819240447121309, + "tokens_seen": 679739392 + }, + { + "epoch": 0.21, + "learning_rate": 0.000401013039117352, + "loss": 1.5395, + "theoretical_loss": 3.8192031446550314, + "tokens_seen": 679804928 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040100300902708126, + "loss": 1.4859, + "theoretical_loss": 3.819165846791493, + "tokens_seen": 679870464 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040099297893681044, + "loss": 1.5497, + "theoretical_loss": 3.8191285535296813, + "tokens_seen": 679936000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009829488465396, + "loss": 1.5202, + "theoretical_loss": 3.819091264868586, + "tokens_seen": 680001536 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009729187562688, + "loss": 1.5135, + "theoretical_loss": 3.8190539808071966, + "tokens_seen": 680067072 + }, + { + "epoch": 0.21, + "learning_rate": 0.000400962888665998, + "loss": 1.4345, + "theoretical_loss": 3.819016701344501, + "tokens_seen": 680132608 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040095285857572716, + "loss": 1.5048, + "theoretical_loss": 3.8189794264794905, + "tokens_seen": 680198144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009428284854564, + "loss": 1.5924, + "theoretical_loss": 3.8189421562111545, + "tokens_seen": 680263680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009327983951855, + "loss": 1.418, + "theoretical_loss": 3.8189048905384837, + "tokens_seen": 680329216 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040092276830491476, + "loss": 1.4858, + "theoretical_loss": 3.8188676294604686, + "tokens_seen": 680394752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009127382146439, + "loss": 1.5633, + "theoretical_loss": 3.8188303729761, + "tokens_seen": 680460288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004009027081243731, + "loss": 1.4377, + "theoretical_loss": 3.818793121084369, + "tokens_seen": 680525824 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040089267803410236, + "loss": 1.5357, + "theoretical_loss": 3.818755873784268, + "tokens_seen": 680591360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004008826479438315, + "loss": 1.5597, + "theoretical_loss": 3.8187186310747885, + "tokens_seen": 680656896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004008726178535607, + "loss": 1.5184, + "theoretical_loss": 3.8186813929549235, + "tokens_seen": 680722432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004008625877632899, + "loss": 1.4358, + "theoretical_loss": 3.818644159423665, + "tokens_seen": 680787968 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004008525576730191, + "loss": 1.5346, + "theoretical_loss": 3.818606930480006, + "tokens_seen": 680853504 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040084252758274826, + "loss": 1.5364, + "theoretical_loss": 3.8185697061229398, + "tokens_seen": 680919040 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040083249749247744, + "loss": 1.3985, + "theoretical_loss": 3.8185324863514594, + "tokens_seen": 680984576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004008224674022066, + "loss": 1.5988, + "theoretical_loss": 3.8184952711645597, + "tokens_seen": 681050112 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040081243731193586, + "loss": 1.5306, + "theoretical_loss": 3.8184580605612344, + "tokens_seen": 681115648 + }, + { + "epoch": 0.21, + "learning_rate": 0.000400802407221665, + "loss": 1.421, + "theoretical_loss": 3.818420854540478, + "tokens_seen": 681181184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004007923771313942, + "loss": 1.5469, + "theoretical_loss": 3.818383653101286, + "tokens_seen": 681246720 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040078234704112335, + "loss": 1.4926, + "theoretical_loss": 3.8183464562426526, + "tokens_seen": 681312256 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004007723169508526, + "loss": 1.5514, + "theoretical_loss": 3.8183092639635734, + "tokens_seen": 681377792 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040076228686058177, + "loss": 1.547, + "theoretical_loss": 3.8182720762630447, + "tokens_seen": 681443328 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040075225677031095, + "loss": 1.5437, + "theoretical_loss": 3.8182348931400627, + "tokens_seen": 681508864 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 253498, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5209672451019287, + "objective/train/theoretical_loss": 3.818197714593623, + "objective/train/tokens_used": 702034400, + "theoretical_loss": 3.818197714593623, + "tokens_seen": 681574400 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040074222668004013, + "loss": 1.5448, + "theoretical_loss": 3.818197714593623, + "tokens_seen": 681574400 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040073219658976936, + "loss": 1.488, + "theoretical_loss": 3.818160540622723, + "tokens_seen": 681639936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004007221664994985, + "loss": 1.6257, + "theoretical_loss": 3.8181233712263594, + "tokens_seen": 681705472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004007121364092277, + "loss": 1.5336, + "theoretical_loss": 3.81808620640353, + "tokens_seen": 681771008 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040070210631895685, + "loss": 1.5059, + "theoretical_loss": 3.8180490461532317, + "tokens_seen": 681836544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004006920762286861, + "loss": 1.5771, + "theoretical_loss": 3.818011890474464, + "tokens_seen": 681902080 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040068204613841527, + "loss": 1.4953, + "theoretical_loss": 3.817974739366224, + "tokens_seen": 681967616 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040067201604814445, + "loss": 1.5374, + "theoretical_loss": 3.8179375928275103, + "tokens_seen": 682033152 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040066198595787363, + "loss": 1.5734, + "theoretical_loss": 3.817900450857322, + "tokens_seen": 682098688 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004006519558676028, + "loss": 1.4704, + "theoretical_loss": 3.817863313454659, + "tokens_seen": 682164224 + }, + { + "epoch": 0.21, + "learning_rate": 0.000400641925777332, + "loss": 1.5826, + "theoretical_loss": 3.8178261806185207, + "tokens_seen": 682229760 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040063189568706123, + "loss": 1.5781, + "theoretical_loss": 3.817789052347906, + "tokens_seen": 682295296 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040062186559679036, + "loss": 1.582, + "theoretical_loss": 3.8177519286418162, + "tokens_seen": 682360832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004006118355065196, + "loss": 1.528, + "theoretical_loss": 3.817714809499252, + "tokens_seen": 682426368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004006018054162487, + "loss": 1.5906, + "theoretical_loss": 3.8176776949192135, + "tokens_seen": 682491904 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040059177532597795, + "loss": 1.5361, + "theoretical_loss": 3.8176405849007016, + "tokens_seen": 682557440 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040058174523570713, + "loss": 1.4427, + "theoretical_loss": 3.8176034794427185, + "tokens_seen": 682622976 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004005717151454363, + "loss": 1.5512, + "theoretical_loss": 3.8175663785442655, + "tokens_seen": 682688512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004005616850551655, + "loss": 1.5927, + "theoretical_loss": 3.8175292822043456, + "tokens_seen": 682754048 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040055165496489473, + "loss": 1.6002, + "theoretical_loss": 3.81749219042196, + "tokens_seen": 682819584 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040054162487462386, + "loss": 1.4049, + "theoretical_loss": 3.8174551031961124, + "tokens_seen": 682885120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004005315947843531, + "loss": 1.5094, + "theoretical_loss": 3.817418020525805, + "tokens_seen": 682950656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004005215646940822, + "loss": 1.4765, + "theoretical_loss": 3.817380942410042, + "tokens_seen": 683016192 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040051153460381146, + "loss": 1.4575, + "theoretical_loss": 3.8173438688478267, + "tokens_seen": 683081728 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040050150451354064, + "loss": 1.4896, + "theoretical_loss": 3.817306799838163, + "tokens_seen": 683147264 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004914744232698, + "loss": 1.5391, + "theoretical_loss": 3.817269735380055, + "tokens_seen": 683212800 + }, + { + "epoch": 0.21, + "learning_rate": 0.000400481444332999, + "loss": 1.5109, + "theoretical_loss": 3.817232675472508, + "tokens_seen": 683278336 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004714142427282, + "loss": 1.553, + "theoretical_loss": 3.817195620114526, + "tokens_seen": 683343872 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040046138415245736, + "loss": 1.5811, + "theoretical_loss": 3.817158569305115, + "tokens_seen": 683409408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004513540621866, + "loss": 1.4715, + "theoretical_loss": 3.817121523043281, + "tokens_seen": 683474944 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004413239719157, + "loss": 1.6887, + "theoretical_loss": 3.8170844813280285, + "tokens_seen": 683540480 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040043129388164496, + "loss": 1.5431, + "theoretical_loss": 3.817047444158364, + "tokens_seen": 683606016 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004212637913741, + "loss": 1.4539, + "theoretical_loss": 3.8170104115332943, + "tokens_seen": 683671552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004112337011033, + "loss": 1.5176, + "theoretical_loss": 3.816973383451827, + "tokens_seen": 683737088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004004012036108325, + "loss": 1.5171, + "theoretical_loss": 3.816936359912968, + "tokens_seen": 683802624 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004003911735205617, + "loss": 1.5763, + "theoretical_loss": 3.8168993409157252, + "tokens_seen": 683868160 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040038114343029087, + "loss": 1.5093, + "theoretical_loss": 3.8168623264591064, + "tokens_seen": 683933696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004003711133400201, + "loss": 1.5215, + "theoretical_loss": 3.8168253165421198, + "tokens_seen": 683999232 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040036108324974923, + "loss": 1.6365, + "theoretical_loss": 3.816788311163773, + "tokens_seen": 684064768 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040035105315947846, + "loss": 1.5091, + "theoretical_loss": 3.8167513103230757, + "tokens_seen": 684130304 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004003410230692076, + "loss": 1.4707, + "theoretical_loss": 3.816714314019036, + "tokens_seen": 684195840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004003309929789368, + "loss": 1.6544, + "theoretical_loss": 3.816677322250664, + "tokens_seen": 684261376 + }, + { + "epoch": 0.21, + "learning_rate": 0.000400320962888666, + "loss": 1.3954, + "theoretical_loss": 3.8166403350169684, + "tokens_seen": 684326912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004003109327983952, + "loss": 1.4524, + "theoretical_loss": 3.81660335231696, + "tokens_seen": 684392448 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040030090270812437, + "loss": 1.5396, + "theoretical_loss": 3.816566374149648, + "tokens_seen": 684457984 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040029087261785355, + "loss": 1.476, + "theoretical_loss": 3.816529400514044, + "tokens_seen": 684523520 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040028084252758273, + "loss": 1.5107, + "theoretical_loss": 3.8164924314091584, + "tokens_seen": 684589056 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040027081243731197, + "loss": 1.4541, + "theoretical_loss": 3.816455466834003, + "tokens_seen": 684654592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004002607823470411, + "loss": 1.5159, + "theoretical_loss": 3.8164185067875875, + "tokens_seen": 684720128 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040025075225677033, + "loss": 1.5177, + "theoretical_loss": 3.8163815512689254, + "tokens_seen": 684785664 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 254674, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4948501586914062, + "objective/train/theoretical_loss": 3.816344600277029, + "objective/train/tokens_used": 705311200, + "theoretical_loss": 3.816344600277029, + "tokens_seen": 684851200 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040024072216649946, + "loss": 1.5019, + "theoretical_loss": 3.816344600277029, + "tokens_seen": 684851200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004002306920762287, + "loss": 1.5377, + "theoretical_loss": 3.8163076538109086, + "tokens_seen": 684916736 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040022066198595787, + "loss": 1.5526, + "theoretical_loss": 3.816270711869579, + "tokens_seen": 684982272 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040021063189568705, + "loss": 1.4517, + "theoretical_loss": 3.8162337744520523, + "tokens_seen": 685047808 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040020060180541623, + "loss": 1.4645, + "theoretical_loss": 3.816196841557342, + "tokens_seen": 685113344 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040019057171514547, + "loss": 1.4712, + "theoretical_loss": 3.816159913184462, + "tokens_seen": 685178880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004001805416248746, + "loss": 1.5301, + "theoretical_loss": 3.8161229893324258, + "tokens_seen": 685244416 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040017051153460383, + "loss": 1.5679, + "theoretical_loss": 3.8160860700002477, + "tokens_seen": 685309952 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040016048144433296, + "loss": 1.4847, + "theoretical_loss": 3.816049155186943, + "tokens_seen": 685375488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004001504513540622, + "loss": 1.5279, + "theoretical_loss": 3.8160122448915255, + "tokens_seen": 685441024 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040014042126379143, + "loss": 1.4925, + "theoretical_loss": 3.815975339113011, + "tokens_seen": 685506560 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040013039117352056, + "loss": 1.6203, + "theoretical_loss": 3.8159384378504146, + "tokens_seen": 685572096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004001203610832498, + "loss": 1.448, + "theoretical_loss": 3.8159015411027526, + "tokens_seen": 685637632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004001103309929789, + "loss": 1.515, + "theoretical_loss": 3.815864648869041, + "tokens_seen": 685703168 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040010030090270815, + "loss": 1.5486, + "theoretical_loss": 3.815827761148296, + "tokens_seen": 685768704 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040009027081243733, + "loss": 1.4747, + "theoretical_loss": 3.8157908779395346, + "tokens_seen": 685834240 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004000802407221665, + "loss": 1.5098, + "theoretical_loss": 3.8157539992417737, + "tokens_seen": 685899776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004000702106318957, + "loss": 1.6611, + "theoretical_loss": 3.8157171250540305, + "tokens_seen": 685965312 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040006018054162493, + "loss": 1.5749, + "theoretical_loss": 3.815680255375323, + "tokens_seen": 686030848 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040005015045135406, + "loss": 1.3995, + "theoretical_loss": 3.815643390204669, + "tokens_seen": 686096384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004000401203610833, + "loss": 1.5226, + "theoretical_loss": 3.8156065295410864, + "tokens_seen": 686161920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004000300902708124, + "loss": 1.5221, + "theoretical_loss": 3.8155696733835947, + "tokens_seen": 686227456 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040002006018054166, + "loss": 1.5862, + "theoretical_loss": 3.8155328217312117, + "tokens_seen": 686292992 + }, + { + "epoch": 0.21, + "learning_rate": 0.00040001003009027084, + "loss": 1.563, + "theoretical_loss": 3.815495974582957, + "tokens_seen": 686358528 + }, + { + "epoch": 0.21, + "learning_rate": 0.0004, + "loss": 1.4183, + "theoretical_loss": 3.8154591319378506, + "tokens_seen": 686424064 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999899699097292, + "loss": 1.5779, + "theoretical_loss": 3.815422293794912, + "tokens_seen": 686489600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999799398194584, + "loss": 1.4954, + "theoretical_loss": 3.8153854601531605, + "tokens_seen": 686555136 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039996990972918756, + "loss": 1.5253, + "theoretical_loss": 3.815348631011618, + "tokens_seen": 686620672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999598796389168, + "loss": 1.4535, + "theoretical_loss": 3.8153118063693037, + "tokens_seen": 686686208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999498495486459, + "loss": 1.463, + "theoretical_loss": 3.8152749862252398, + "tokens_seen": 686751744 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039993981945837516, + "loss": 1.4972, + "theoretical_loss": 3.8152381705784473, + "tokens_seen": 686817280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999297893681043, + "loss": 1.5033, + "theoretical_loss": 3.815201359427948, + "tokens_seen": 686882816 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999197592778335, + "loss": 1.5529, + "theoretical_loss": 3.815164552772763, + "tokens_seen": 686948352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003999097291875627, + "loss": 1.4666, + "theoretical_loss": 3.8151277506119152, + "tokens_seen": 687013888 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003998996990972919, + "loss": 1.4605, + "theoretical_loss": 3.8150909529444275, + "tokens_seen": 687079424 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039988966900702107, + "loss": 1.481, + "theoretical_loss": 3.815054159769322, + "tokens_seen": 687144960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003998796389167503, + "loss": 1.5993, + "theoretical_loss": 3.8150173710856228, + "tokens_seen": 687210496 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039986960882647943, + "loss": 1.4533, + "theoretical_loss": 3.8149805868923528, + "tokens_seen": 687276032 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039985957873620866, + "loss": 1.4445, + "theoretical_loss": 3.814943807188536, + "tokens_seen": 687341568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003998495486459378, + "loss": 1.5369, + "theoretical_loss": 3.8149070319731955, + "tokens_seen": 687407104 + }, + { + "epoch": 0.21, + "learning_rate": 0.000399839518555667, + "loss": 1.6655, + "theoretical_loss": 3.8148702612453573, + "tokens_seen": 687472640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003998294884653962, + "loss": 1.5854, + "theoretical_loss": 3.814833495004045, + "tokens_seen": 687538176 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003998194583751254, + "loss": 1.4639, + "theoretical_loss": 3.8147967332482846, + "tokens_seen": 687603712 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039980942828485457, + "loss": 1.6681, + "theoretical_loss": 3.8147599759771, + "tokens_seen": 687669248 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039979939819458375, + "loss": 1.5176, + "theoretical_loss": 3.814723223189518, + "tokens_seen": 687734784 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039978936810431293, + "loss": 1.4811, + "theoretical_loss": 3.8146864748845637, + "tokens_seen": 687800320 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039977933801404217, + "loss": 1.3833, + "theoretical_loss": 3.814649731061264, + "tokens_seen": 687865856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003997693079237713, + "loss": 1.5049, + "theoretical_loss": 3.8146129917186453, + "tokens_seen": 687931392 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039975927783350053, + "loss": 1.4418, + "theoretical_loss": 3.8145762568557338, + "tokens_seen": 687996928 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039974924774322966, + "loss": 1.5678, + "theoretical_loss": 3.814539526471558, + "tokens_seen": 688062464 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 255827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4779702425003052, + "objective/train/theoretical_loss": 3.8145028005651436, + "objective/train/tokens_used": 708588000, + "theoretical_loss": 3.8145028005651436, + "tokens_seen": 688128000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003997392176529589, + "loss": 1.5128, + "theoretical_loss": 3.8145028005651436, + "tokens_seen": 688128000 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039972918756268807, + "loss": 1.4961, + "theoretical_loss": 3.8144660791355194, + "tokens_seen": 688193536 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039971915747241725, + "loss": 1.5457, + "theoretical_loss": 3.8144293621817136, + "tokens_seen": 688259072 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039970912738214643, + "loss": 1.5303, + "theoretical_loss": 3.8143926497027536, + "tokens_seen": 688324608 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039969909729187567, + "loss": 1.5146, + "theoretical_loss": 3.814355941697669, + "tokens_seen": 688390144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003996890672016048, + "loss": 1.4937, + "theoretical_loss": 3.814319238165489, + "tokens_seen": 688455680 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039967903711133403, + "loss": 1.5792, + "theoretical_loss": 3.8142825391052417, + "tokens_seen": 688521216 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039966900702106316, + "loss": 1.5117, + "theoretical_loss": 3.8142458445159573, + "tokens_seen": 688586752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003996589769307924, + "loss": 1.5101, + "theoretical_loss": 3.8142091543966656, + "tokens_seen": 688652288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003996489468405216, + "loss": 1.4189, + "theoretical_loss": 3.814172468746397, + "tokens_seen": 688717824 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039963891675025076, + "loss": 1.5676, + "theoretical_loss": 3.8141357875641813, + "tokens_seen": 688783360 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039962888665997994, + "loss": 1.4743, + "theoretical_loss": 3.81409911084905, + "tokens_seen": 688848896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003996188565697091, + "loss": 1.5417, + "theoretical_loss": 3.8140624386000344, + "tokens_seen": 688914432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003996088264794383, + "loss": 1.4841, + "theoretical_loss": 3.8140257708161647, + "tokens_seen": 688979968 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039959879638916754, + "loss": 1.5613, + "theoretical_loss": 3.8139891074964734, + "tokens_seen": 689045504 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039958876629889666, + "loss": 1.5156, + "theoretical_loss": 3.8139524486399927, + "tokens_seen": 689111040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003995787362086259, + "loss": 1.522, + "theoretical_loss": 3.813915794245754, + "tokens_seen": 689176576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003995687061183551, + "loss": 1.404, + "theoretical_loss": 3.813879144312791, + "tokens_seen": 689242112 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039955867602808426, + "loss": 1.5346, + "theoretical_loss": 3.8138424988401365, + "tokens_seen": 689307648 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039954864593781344, + "loss": 1.6499, + "theoretical_loss": 3.8138058578268224, + "tokens_seen": 689373184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003995386158475426, + "loss": 1.4484, + "theoretical_loss": 3.8137692212718832, + "tokens_seen": 689438720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003995285857572718, + "loss": 1.5331, + "theoretical_loss": 3.8137325891743528, + "tokens_seen": 689504256 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039951855566700104, + "loss": 1.5277, + "theoretical_loss": 3.8136959615332646, + "tokens_seen": 689569792 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039950852557673017, + "loss": 1.484, + "theoretical_loss": 3.813659338347654, + "tokens_seen": 689635328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003994984954864594, + "loss": 1.5666, + "theoretical_loss": 3.8136227196165544, + "tokens_seen": 689700864 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039948846539618853, + "loss": 1.5283, + "theoretical_loss": 3.813586105339002, + "tokens_seen": 689766400 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039947843530591776, + "loss": 1.5482, + "theoretical_loss": 3.813549495514031, + "tokens_seen": 689831936 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039946840521564694, + "loss": 1.499, + "theoretical_loss": 3.8135128901406783, + "tokens_seen": 689897472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003994583751253761, + "loss": 1.5282, + "theoretical_loss": 3.813476289217979, + "tokens_seen": 689963008 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003994483450351053, + "loss": 1.4263, + "theoretical_loss": 3.813439692744969, + "tokens_seen": 690028544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003994383149448345, + "loss": 1.6184, + "theoretical_loss": 3.813403100720686, + "tokens_seen": 690094080 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039942828485456367, + "loss": 1.497, + "theoretical_loss": 3.813366513144165, + "tokens_seen": 690159616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003994182547642929, + "loss": 1.4725, + "theoretical_loss": 3.8133299300144445, + "tokens_seen": 690225152 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039940822467402203, + "loss": 1.514, + "theoretical_loss": 3.813293351330562, + "tokens_seen": 690290688 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039939819458375127, + "loss": 1.5666, + "theoretical_loss": 3.8132567770915546, + "tokens_seen": 690356224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003993881644934805, + "loss": 1.6007, + "theoretical_loss": 3.8132202072964603, + "tokens_seen": 690421760 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039937813440320963, + "loss": 1.4877, + "theoretical_loss": 3.8131836419443172, + "tokens_seen": 690487296 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039936810431293886, + "loss": 1.4643, + "theoretical_loss": 3.8131470810341646, + "tokens_seen": 690552832 + }, + { + "epoch": 0.21, + "learning_rate": 0.000399358074222668, + "loss": 1.5409, + "theoretical_loss": 3.8131105245650403, + "tokens_seen": 690618368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003993480441323972, + "loss": 1.558, + "theoretical_loss": 3.8130739725359852, + "tokens_seen": 690683904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003993380140421264, + "loss": 1.5977, + "theoretical_loss": 3.8130374249460375, + "tokens_seen": 690749440 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003993279839518556, + "loss": 1.5617, + "theoretical_loss": 3.8130008817942374, + "tokens_seen": 690814976 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039931795386158477, + "loss": 1.4561, + "theoretical_loss": 3.8129643430796243, + "tokens_seen": 690880512 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039930792377131395, + "loss": 1.4087, + "theoretical_loss": 3.8129278088012395, + "tokens_seen": 690946048 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039929789368104313, + "loss": 1.4569, + "theoretical_loss": 3.8128912789581237, + "tokens_seen": 691011584 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039928786359077237, + "loss": 1.6216, + "theoretical_loss": 3.8128547535493174, + "tokens_seen": 691077120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003992778335005015, + "loss": 1.5934, + "theoretical_loss": 3.8128182325738624, + "tokens_seen": 691142656 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039926780341023073, + "loss": 1.64, + "theoretical_loss": 3.8127817160307993, + "tokens_seen": 691208192 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039925777331995986, + "loss": 1.4208, + "theoretical_loss": 3.812745203919171, + "tokens_seen": 691273728 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003992477432296891, + "loss": 1.5114, + "theoretical_loss": 3.8127086962380194, + "tokens_seen": 691339264 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 257040, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4978456497192383, + "objective/train/theoretical_loss": 3.812672192986387, + "objective/train/tokens_used": 711864800, + "theoretical_loss": 3.812672192986387, + "tokens_seen": 691404800 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039923771313941827, + "loss": 1.5176, + "theoretical_loss": 3.812672192986387, + "tokens_seen": 691404800 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039922768304914745, + "loss": 1.3901, + "theoretical_loss": 3.8126356941633164, + "tokens_seen": 691470336 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039921765295887663, + "loss": 1.5115, + "theoretical_loss": 3.8125991997678503, + "tokens_seen": 691535872 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039920762286860587, + "loss": 1.4319, + "theoretical_loss": 3.8125627097990327, + "tokens_seen": 691601408 + }, + { + "epoch": 0.21, + "learning_rate": 0.000399197592778335, + "loss": 1.4812, + "theoretical_loss": 3.8125262242559073, + "tokens_seen": 691666944 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039918756268806423, + "loss": 1.5277, + "theoretical_loss": 3.812489743137518, + "tokens_seen": 691732480 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039917753259779336, + "loss": 1.4304, + "theoretical_loss": 3.8124532664429087, + "tokens_seen": 691798016 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003991675025075226, + "loss": 1.4167, + "theoretical_loss": 3.8124167941711242, + "tokens_seen": 691863552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003991574724172518, + "loss": 1.5579, + "theoretical_loss": 3.8123803263212097, + "tokens_seen": 691929088 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039914744232698096, + "loss": 1.5706, + "theoretical_loss": 3.8123438628922095, + "tokens_seen": 691994624 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039913741223671014, + "loss": 1.4638, + "theoretical_loss": 3.8123074038831697, + "tokens_seen": 692060160 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003991273821464393, + "loss": 1.5588, + "theoretical_loss": 3.8122709492931364, + "tokens_seen": 692125696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003991173520561685, + "loss": 1.4619, + "theoretical_loss": 3.8122344991211543, + "tokens_seen": 692191232 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039910732196589774, + "loss": 1.5347, + "theoretical_loss": 3.8121980533662714, + "tokens_seen": 692256768 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039909729187562686, + "loss": 1.4016, + "theoretical_loss": 3.8121616120275332, + "tokens_seen": 692322304 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003990872617853561, + "loss": 1.5698, + "theoretical_loss": 3.812125175103987, + "tokens_seen": 692387840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003990772316950853, + "loss": 1.485, + "theoretical_loss": 3.81208874259468, + "tokens_seen": 692453376 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039906720160481446, + "loss": 1.5056, + "theoretical_loss": 3.8120523144986596, + "tokens_seen": 692518912 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039905717151454364, + "loss": 1.5218, + "theoretical_loss": 3.8120158908149735, + "tokens_seen": 692584448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003990471414242728, + "loss": 1.6273, + "theoretical_loss": 3.8119794715426707, + "tokens_seen": 692649984 + }, + { + "epoch": 0.21, + "learning_rate": 0.000399037111334002, + "loss": 1.5304, + "theoretical_loss": 3.8119430566807986, + "tokens_seen": 692715520 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039902708124373124, + "loss": 1.5508, + "theoretical_loss": 3.8119066462284064, + "tokens_seen": 692781056 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039901705115346037, + "loss": 1.4934, + "theoretical_loss": 3.8118702401845423, + "tokens_seen": 692846592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003990070210631896, + "loss": 1.4982, + "theoretical_loss": 3.8118338385482575, + "tokens_seen": 692912128 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039899699097291873, + "loss": 1.5581, + "theoretical_loss": 3.8117974413185993, + "tokens_seen": 692977664 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039898696088264796, + "loss": 1.5096, + "theoretical_loss": 3.8117610484946187, + "tokens_seen": 693043200 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039897693079237714, + "loss": 1.6033, + "theoretical_loss": 3.8117246600753667, + "tokens_seen": 693108736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003989669007021063, + "loss": 1.6413, + "theoretical_loss": 3.8116882760598925, + "tokens_seen": 693174272 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003989568706118355, + "loss": 1.4982, + "theoretical_loss": 3.811651896447247, + "tokens_seen": 693239808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003989468405215647, + "loss": 1.5294, + "theoretical_loss": 3.8116155212364817, + "tokens_seen": 693305344 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039893681043129387, + "loss": 1.4717, + "theoretical_loss": 3.811579150426648, + "tokens_seen": 693370880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003989267803410231, + "loss": 1.5633, + "theoretical_loss": 3.811542784016797, + "tokens_seen": 693436416 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039891675025075223, + "loss": 1.5733, + "theoretical_loss": 3.811506422005981, + "tokens_seen": 693501952 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039890672016048147, + "loss": 1.5284, + "theoretical_loss": 3.8114700643932524, + "tokens_seen": 693567488 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039889669007021065, + "loss": 1.4712, + "theoretical_loss": 3.8114337111776635, + "tokens_seen": 693633024 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039888665997993983, + "loss": 1.5599, + "theoretical_loss": 3.8113973623582673, + "tokens_seen": 693698560 + }, + { + "epoch": 0.21, + "learning_rate": 0.000398876629889669, + "loss": 1.5175, + "theoretical_loss": 3.8113610179341166, + "tokens_seen": 693764096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003988665997993982, + "loss": 1.5712, + "theoretical_loss": 3.811324677904265, + "tokens_seen": 693829632 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039885656970912737, + "loss": 1.5257, + "theoretical_loss": 3.811288342267767, + "tokens_seen": 693895168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003988465396188566, + "loss": 1.5584, + "theoretical_loss": 3.811252011023675, + "tokens_seen": 693960704 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039883650952858573, + "loss": 1.4308, + "theoretical_loss": 3.811215684171044, + "tokens_seen": 694026240 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039882647943831497, + "loss": 1.5976, + "theoretical_loss": 3.8111793617089296, + "tokens_seen": 694091776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003988164493480441, + "loss": 1.544, + "theoretical_loss": 3.8111430436363856, + "tokens_seen": 694157312 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039880641925777333, + "loss": 1.6917, + "theoretical_loss": 3.8111067299524666, + "tokens_seen": 694222848 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003987963891675025, + "loss": 1.4979, + "theoretical_loss": 3.81107042065623, + "tokens_seen": 694288384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003987863590772317, + "loss": 1.4415, + "theoretical_loss": 3.811034115746729, + "tokens_seen": 694353920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003987763289869609, + "loss": 1.513, + "theoretical_loss": 3.810997815223022, + "tokens_seen": 694419456 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039876629889669006, + "loss": 1.6239, + "theoretical_loss": 3.810961519084164, + "tokens_seen": 694484992 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039875626880641924, + "loss": 1.4457, + "theoretical_loss": 3.8109252273292125, + "tokens_seen": 694550528 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039874623871614847, + "loss": 1.4472, + "theoretical_loss": 3.8108889399572234, + "tokens_seen": 694616064 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 258341, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5276737213134766, + "objective/train/theoretical_loss": 3.810852656967255, + "objective/train/tokens_used": 715141600, + "theoretical_loss": 3.810852656967255, + "tokens_seen": 694681600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003987362086258776, + "loss": 1.6103, + "theoretical_loss": 3.810852656967255, + "tokens_seen": 694681600 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039872617853560683, + "loss": 1.5194, + "theoretical_loss": 3.8108163783583637, + "tokens_seen": 694747136 + }, + { + "epoch": 0.21, + "learning_rate": 0.000398716148445336, + "loss": 1.5559, + "theoretical_loss": 3.8107801041296083, + "tokens_seen": 694812672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003987061183550652, + "loss": 1.596, + "theoretical_loss": 3.8107438342800464, + "tokens_seen": 694878208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003986960882647944, + "loss": 1.5276, + "theoretical_loss": 3.810707568808736, + "tokens_seen": 694943744 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039868605817452356, + "loss": 1.5042, + "theoretical_loss": 3.8106713077147365, + "tokens_seen": 695009280 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039867602808425274, + "loss": 1.575, + "theoretical_loss": 3.810635050997107, + "tokens_seen": 695074816 + }, + { + "epoch": 0.21, + "learning_rate": 0.000398665997993982, + "loss": 1.4326, + "theoretical_loss": 3.810598798654906, + "tokens_seen": 695140352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003986559679037111, + "loss": 1.5767, + "theoretical_loss": 3.8105625506871936, + "tokens_seen": 695205888 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039864593781344034, + "loss": 1.5137, + "theoretical_loss": 3.810526307093029, + "tokens_seen": 695271424 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003986359077231695, + "loss": 1.4278, + "theoretical_loss": 3.810490067871473, + "tokens_seen": 695336960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003986258776328987, + "loss": 1.5855, + "theoretical_loss": 3.810453833021586, + "tokens_seen": 695402496 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039861584754262794, + "loss": 1.4665, + "theoretical_loss": 3.810417602542428, + "tokens_seen": 695468032 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039860581745235706, + "loss": 1.4899, + "theoretical_loss": 3.8103813764330607, + "tokens_seen": 695533568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003985957873620863, + "loss": 1.4206, + "theoretical_loss": 3.8103451546925453, + "tokens_seen": 695599104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003985857572718155, + "loss": 1.6455, + "theoretical_loss": 3.8103089373199435, + "tokens_seen": 695664640 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039857572718154466, + "loss": 1.4074, + "theoretical_loss": 3.8102727243143164, + "tokens_seen": 695730176 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039856569709127384, + "loss": 1.5079, + "theoretical_loss": 3.8102365156747267, + "tokens_seen": 695795712 + }, + { + "epoch": 0.21, + "learning_rate": 0.000398555667001003, + "loss": 1.5476, + "theoretical_loss": 3.810200311400237, + "tokens_seen": 695861248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003985456369107322, + "loss": 1.474, + "theoretical_loss": 3.8101641114899105, + "tokens_seen": 695926784 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039853560682046144, + "loss": 1.5257, + "theoretical_loss": 3.810127915942809, + "tokens_seen": 695992320 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039852557673019057, + "loss": 1.4146, + "theoretical_loss": 3.8100917247579966, + "tokens_seen": 696057856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003985155466399198, + "loss": 1.4588, + "theoretical_loss": 3.8100555379345367, + "tokens_seen": 696123392 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039850551654964893, + "loss": 1.5627, + "theoretical_loss": 3.810019355471493, + "tokens_seen": 696188928 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039849548645937816, + "loss": 1.517, + "theoretical_loss": 3.80998317736793, + "tokens_seen": 696254464 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039848545636910734, + "loss": 1.5342, + "theoretical_loss": 3.809947003622913, + "tokens_seen": 696320000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003984754262788365, + "loss": 1.5749, + "theoretical_loss": 3.8099108342355046, + "tokens_seen": 696385536 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003984653961885657, + "loss": 1.4796, + "theoretical_loss": 3.809874669204772, + "tokens_seen": 696451072 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003984553660982949, + "loss": 1.525, + "theoretical_loss": 3.8098385085297792, + "tokens_seen": 696516608 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039844533600802407, + "loss": 1.479, + "theoretical_loss": 3.809802352209592, + "tokens_seen": 696582144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003984353059177533, + "loss": 1.4353, + "theoretical_loss": 3.809766200243277, + "tokens_seen": 696647680 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039842527582748243, + "loss": 1.5815, + "theoretical_loss": 3.8097300526299, + "tokens_seen": 696713216 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039841524573721167, + "loss": 1.6454, + "theoretical_loss": 3.8096939093685274, + "tokens_seen": 696778752 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039840521564694085, + "loss": 1.4267, + "theoretical_loss": 3.8096577704582266, + "tokens_seen": 696844288 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039839518555667003, + "loss": 1.543, + "theoretical_loss": 3.8096216358980635, + "tokens_seen": 696909824 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003983851554663992, + "loss": 1.6136, + "theoretical_loss": 3.809585505687106, + "tokens_seen": 696975360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003983751253761284, + "loss": 1.5592, + "theoretical_loss": 3.809549379824422, + "tokens_seen": 697040896 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039836509528585757, + "loss": 1.4933, + "theoretical_loss": 3.80951325830908, + "tokens_seen": 697106432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003983550651955868, + "loss": 1.4325, + "theoretical_loss": 3.809477141140147, + "tokens_seen": 697171968 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039834503510531593, + "loss": 1.4407, + "theoretical_loss": 3.8094410283166917, + "tokens_seen": 697237504 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039833500501504517, + "loss": 1.4879, + "theoretical_loss": 3.809404919837784, + "tokens_seen": 697303040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003983249749247743, + "loss": 1.6092, + "theoretical_loss": 3.8093688157024914, + "tokens_seen": 697368576 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039831494483450353, + "loss": 1.5302, + "theoretical_loss": 3.8093327159098846, + "tokens_seen": 697434112 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003983049147442327, + "loss": 1.4791, + "theoretical_loss": 3.8092966204590324, + "tokens_seen": 697499648 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982948846539619, + "loss": 1.6757, + "theoretical_loss": 3.8092605293490056, + "tokens_seen": 697565184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982848545636911, + "loss": 1.5543, + "theoretical_loss": 3.809224442578873, + "tokens_seen": 697630720 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039827482447342026, + "loss": 1.5663, + "theoretical_loss": 3.809188360147707, + "tokens_seen": 697696256 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039826479438314944, + "loss": 1.452, + "theoretical_loss": 3.8091522820545762, + "tokens_seen": 697761792 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039825476429287867, + "loss": 1.6326, + "theoretical_loss": 3.809116208298554, + "tokens_seen": 697827328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982447342026078, + "loss": 1.5913, + "theoretical_loss": 3.80908013887871, + "tokens_seen": 697892864 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 259086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5467989444732666, + "objective/train/theoretical_loss": 3.8090440737941176, + "objective/train/tokens_used": 718418400, + "theoretical_loss": 3.8090440737941176, + "tokens_seen": 697958400 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039823470411233703, + "loss": 1.5015, + "theoretical_loss": 3.8090440737941176, + "tokens_seen": 697958400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982246740220662, + "loss": 1.4806, + "theoretical_loss": 3.8090080130438464, + "tokens_seen": 698023936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982146439317954, + "loss": 1.5581, + "theoretical_loss": 3.80897195662697, + "tokens_seen": 698089472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003982046138415246, + "loss": 1.4053, + "theoretical_loss": 3.8089359045425617, + "tokens_seen": 698155008 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039819458375125376, + "loss": 1.4268, + "theoretical_loss": 3.808899856789693, + "tokens_seen": 698220544 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039818455366098294, + "loss": 1.6057, + "theoretical_loss": 3.808863813367437, + "tokens_seen": 698286080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003981745235707122, + "loss": 1.5356, + "theoretical_loss": 3.808827774274868, + "tokens_seen": 698351616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003981644934804413, + "loss": 1.4864, + "theoretical_loss": 3.808791739511059, + "tokens_seen": 698417152 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039815446339017054, + "loss": 1.5622, + "theoretical_loss": 3.808755709075084, + "tokens_seen": 698482688 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039814443329989966, + "loss": 1.5111, + "theoretical_loss": 3.8087196829660175, + "tokens_seen": 698548224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003981344032096289, + "loss": 1.5985, + "theoretical_loss": 3.8086836611829336, + "tokens_seen": 698613760 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003981243731193581, + "loss": 1.3808, + "theoretical_loss": 3.8086476437249077, + "tokens_seen": 698679296 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039811434302908726, + "loss": 1.5503, + "theoretical_loss": 3.808611630591014, + "tokens_seen": 698744832 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039810431293881644, + "loss": 1.5325, + "theoretical_loss": 3.808575621780329, + "tokens_seen": 698810368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003980942828485457, + "loss": 1.5715, + "theoretical_loss": 3.8085396172919275, + "tokens_seen": 698875904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003980842527582748, + "loss": 1.581, + "theoretical_loss": 3.8085036171248854, + "tokens_seen": 698941440 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039807422266800404, + "loss": 1.5867, + "theoretical_loss": 3.8084676212782793, + "tokens_seen": 699006976 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039806419257773317, + "loss": 1.6166, + "theoretical_loss": 3.808431629751186, + "tokens_seen": 699072512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003980541624874624, + "loss": 1.4823, + "theoretical_loss": 3.8083956425426813, + "tokens_seen": 699138048 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003980441323971916, + "loss": 1.5125, + "theoretical_loss": 3.808359659651843, + "tokens_seen": 699203584 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039803410230692077, + "loss": 1.4878, + "theoretical_loss": 3.8083236810777485, + "tokens_seen": 699269120 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039802407221664995, + "loss": 1.468, + "theoretical_loss": 3.808287706819475, + "tokens_seen": 699334656 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039801404212637913, + "loss": 1.5157, + "theoretical_loss": 3.808251736876101, + "tokens_seen": 699400192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003980040120361083, + "loss": 1.5546, + "theoretical_loss": 3.8082157712467044, + "tokens_seen": 699465728 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039799398194583754, + "loss": 1.5601, + "theoretical_loss": 3.8081798099303636, + "tokens_seen": 699531264 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039798395185556667, + "loss": 1.5361, + "theoretical_loss": 3.808143852926157, + "tokens_seen": 699596800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003979739217652959, + "loss": 1.5354, + "theoretical_loss": 3.8081079002331646, + "tokens_seen": 699662336 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039796389167502503, + "loss": 1.4908, + "theoretical_loss": 3.808071951850465, + "tokens_seen": 699727872 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039795386158475427, + "loss": 1.5848, + "theoretical_loss": 3.8080360077771385, + "tokens_seen": 699793408 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039794383149448345, + "loss": 1.4422, + "theoretical_loss": 3.808000068012264, + "tokens_seen": 699858944 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039793380140421263, + "loss": 1.4972, + "theoretical_loss": 3.8079641325549223, + "tokens_seen": 699924480 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003979237713139418, + "loss": 1.5865, + "theoretical_loss": 3.807928201404194, + "tokens_seen": 699990016 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039791374122367105, + "loss": 1.4072, + "theoretical_loss": 3.8078922745591597, + "tokens_seen": 700055552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003979037111334002, + "loss": 1.2482, + "theoretical_loss": 3.8078563520189004, + "tokens_seen": 700121088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003978936810431294, + "loss": 1.5335, + "theoretical_loss": 3.8078204337824975, + "tokens_seen": 700186624 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003978836509528586, + "loss": 1.5516, + "theoretical_loss": 3.807784519849032, + "tokens_seen": 700252160 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039787362086258777, + "loss": 1.5557, + "theoretical_loss": 3.807748610217587, + "tokens_seen": 700317696 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397863590772317, + "loss": 1.5552, + "theoretical_loss": 3.8077127048872437, + "tokens_seen": 700383232 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039785356068204613, + "loss": 1.5311, + "theoretical_loss": 3.8076768038570847, + "tokens_seen": 700448768 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039784353059177537, + "loss": 1.5562, + "theoretical_loss": 3.807640907126193, + "tokens_seen": 700514304 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003978335005015045, + "loss": 1.4725, + "theoretical_loss": 3.8076050146936513, + "tokens_seen": 700579840 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039782347041123373, + "loss": 1.5865, + "theoretical_loss": 3.807569126558543, + "tokens_seen": 700645376 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003978134403209629, + "loss": 1.4425, + "theoretical_loss": 3.807533242719952, + "tokens_seen": 700710912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003978034102306921, + "loss": 1.5642, + "theoretical_loss": 3.807497363176962, + "tokens_seen": 700776448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003977933801404213, + "loss": 1.4955, + "theoretical_loss": 3.8074614879286566, + "tokens_seen": 700841984 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039778335005015046, + "loss": 1.5072, + "theoretical_loss": 3.8074256169741205, + "tokens_seen": 700907520 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039777331995987964, + "loss": 1.5352, + "theoretical_loss": 3.8073897503124385, + "tokens_seen": 700973056 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039776328986960887, + "loss": 1.4979, + "theoretical_loss": 3.807353887942696, + "tokens_seen": 701038592 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397753259779338, + "loss": 1.4424, + "theoretical_loss": 3.8073180298639775, + "tokens_seen": 701104128 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039774322968906723, + "loss": 1.4904, + "theoretical_loss": 3.8072821760753683, + "tokens_seen": 701169664 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 260260, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5290884971618652, + "objective/train/theoretical_loss": 3.8072463265759557, + "objective/train/tokens_used": 721695200, + "theoretical_loss": 3.8072463265759557, + "tokens_seen": 701235200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003977331995987964, + "loss": 1.526, + "theoretical_loss": 3.8072463265759557, + "tokens_seen": 701235200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003977231695085256, + "loss": 1.5318, + "theoretical_loss": 3.8072104813648244, + "tokens_seen": 701300736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003977131394182548, + "loss": 1.4839, + "theoretical_loss": 3.8071746404410614, + "tokens_seen": 701366272 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039770310932798396, + "loss": 1.4875, + "theoretical_loss": 3.807138803803753, + "tokens_seen": 701431808 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039769307923771314, + "loss": 1.5254, + "theoretical_loss": 3.807102971451986, + "tokens_seen": 701497344 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003976830491474424, + "loss": 1.5299, + "theoretical_loss": 3.807067143384849, + "tokens_seen": 701562880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003976730190571715, + "loss": 1.5053, + "theoretical_loss": 3.8070313196014274, + "tokens_seen": 701628416 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039766298896690074, + "loss": 1.5602, + "theoretical_loss": 3.8069955001008102, + "tokens_seen": 701693952 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039765295887662987, + "loss": 1.5597, + "theoretical_loss": 3.8069596848820852, + "tokens_seen": 701759488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003976429287863591, + "loss": 1.5286, + "theoretical_loss": 3.806923873944341, + "tokens_seen": 701825024 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003976328986960883, + "loss": 1.5245, + "theoretical_loss": 3.8068880672866667, + "tokens_seen": 701890560 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039762286860581746, + "loss": 1.6187, + "theoretical_loss": 3.8068522649081498, + "tokens_seen": 701956096 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039761283851554664, + "loss": 1.5542, + "theoretical_loss": 3.80681646680788, + "tokens_seen": 702021632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003976028084252759, + "loss": 1.4778, + "theoretical_loss": 3.8067806729849467, + "tokens_seen": 702087168 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397592778335005, + "loss": 1.4767, + "theoretical_loss": 3.8067448834384408, + "tokens_seen": 702152704 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039758274824473424, + "loss": 1.3103, + "theoretical_loss": 3.8067090981674507, + "tokens_seen": 702218240 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039757271815446337, + "loss": 1.5437, + "theoretical_loss": 3.806673317171067, + "tokens_seen": 702283776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003975626880641926, + "loss": 1.5101, + "theoretical_loss": 3.8066375404483814, + "tokens_seen": 702349312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003975526579739218, + "loss": 1.439, + "theoretical_loss": 3.806601767998483, + "tokens_seen": 702414848 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039754262788365097, + "loss": 1.513, + "theoretical_loss": 3.806565999820465, + "tokens_seen": 702480384 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039753259779338015, + "loss": 1.4899, + "theoretical_loss": 3.806530235913417, + "tokens_seen": 702545920 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039752256770310933, + "loss": 1.4992, + "theoretical_loss": 3.8064944762764314, + "tokens_seen": 702611456 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003975125376128385, + "loss": 1.6493, + "theoretical_loss": 3.8064587209086005, + "tokens_seen": 702676992 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039750250752256774, + "loss": 1.5624, + "theoretical_loss": 3.8064229698090157, + "tokens_seen": 702742528 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039749247743229687, + "loss": 1.5295, + "theoretical_loss": 3.80638722297677, + "tokens_seen": 702808064 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003974824473420261, + "loss": 1.5116, + "theoretical_loss": 3.8063514804109557, + "tokens_seen": 702873600 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039747241725175523, + "loss": 1.5818, + "theoretical_loss": 3.806315742110667, + "tokens_seen": 702939136 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039746238716148447, + "loss": 1.5403, + "theoretical_loss": 3.806280008074996, + "tokens_seen": 703004672 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039745235707121365, + "loss": 1.5523, + "theoretical_loss": 3.806244278303037, + "tokens_seen": 703070208 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039744232698094283, + "loss": 1.5398, + "theoretical_loss": 3.806208552793884, + "tokens_seen": 703135744 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397432296890672, + "loss": 1.5678, + "theoretical_loss": 3.8061728315466308, + "tokens_seen": 703201280 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039742226680040125, + "loss": 1.455, + "theoretical_loss": 3.8061371145603715, + "tokens_seen": 703266816 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003974122367101304, + "loss": 1.4988, + "theoretical_loss": 3.806101401834202, + "tokens_seen": 703332352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003974022066198596, + "loss": 1.4753, + "theoretical_loss": 3.8060656933672163, + "tokens_seen": 703397888 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039739217652958874, + "loss": 1.5643, + "theoretical_loss": 3.8060299891585094, + "tokens_seen": 703463424 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039738214643931797, + "loss": 1.4532, + "theoretical_loss": 3.8059942892071783, + "tokens_seen": 703528960 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039737211634904715, + "loss": 1.5807, + "theoretical_loss": 3.8059585935123175, + "tokens_seen": 703594496 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039736208625877633, + "loss": 1.4072, + "theoretical_loss": 3.8059229020730236, + "tokens_seen": 703660032 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003973520561685055, + "loss": 1.4933, + "theoretical_loss": 3.805887214888393, + "tokens_seen": 703725568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003973420260782347, + "loss": 1.6198, + "theoretical_loss": 3.8058515319575226, + "tokens_seen": 703791104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003973319959879639, + "loss": 1.527, + "theoretical_loss": 3.8058158532795088, + "tokens_seen": 703856640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003973219658976931, + "loss": 1.6384, + "theoretical_loss": 3.805780178853449, + "tokens_seen": 703922176 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039731193580742224, + "loss": 1.4684, + "theoretical_loss": 3.8057445086784405, + "tokens_seen": 703987712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003973019057171515, + "loss": 1.4803, + "theoretical_loss": 3.8057088427535817, + "tokens_seen": 704053248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003972918756268806, + "loss": 1.5173, + "theoretical_loss": 3.80567318107797, + "tokens_seen": 704118784 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039728184553660984, + "loss": 1.5576, + "theoretical_loss": 3.8056375236507045, + "tokens_seen": 704184320 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397271815446339, + "loss": 1.5115, + "theoretical_loss": 3.805601870470883, + "tokens_seen": 704249856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003972617853560682, + "loss": 1.5015, + "theoretical_loss": 3.805566221537604, + "tokens_seen": 704315392 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003972517552657974, + "loss": 1.4417, + "theoretical_loss": 3.805530576849968, + "tokens_seen": 704380928 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003972417251755266, + "loss": 1.5388, + "theoretical_loss": 3.8054949364070736, + "tokens_seen": 704446464 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 261464, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.369011402130127, + "objective/train/theoretical_loss": 3.8054593002080206, + "objective/train/tokens_used": 724972000, + "theoretical_loss": 3.8054593002080206, + "tokens_seen": 704512000 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039723169508525574, + "loss": 1.4361, + "theoretical_loss": 3.8054593002080206, + "tokens_seen": 704512000 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397221664994985, + "loss": 1.4422, + "theoretical_loss": 3.805423668251909, + "tokens_seen": 704577536 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003972116349047141, + "loss": 1.5738, + "theoretical_loss": 3.8053880405378386, + "tokens_seen": 704643072 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039720160481444334, + "loss": 1.562, + "theoretical_loss": 3.8053524170649107, + "tokens_seen": 704708608 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971915747241725, + "loss": 1.5342, + "theoretical_loss": 3.805316797832225, + "tokens_seen": 704774144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971815446339017, + "loss": 1.4668, + "theoretical_loss": 3.805281182838884, + "tokens_seen": 704839680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971715145436309, + "loss": 1.6376, + "theoretical_loss": 3.8052455720839875, + "tokens_seen": 704905216 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039716148445336007, + "loss": 1.5279, + "theoretical_loss": 3.8052099655666387, + "tokens_seen": 704970752 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039715145436308925, + "loss": 1.5702, + "theoretical_loss": 3.8051743632859383, + "tokens_seen": 705036288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971414242728185, + "loss": 1.5041, + "theoretical_loss": 3.8051387652409887, + "tokens_seen": 705101824 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039713139418254766, + "loss": 1.5527, + "theoretical_loss": 3.8051031714308925, + "tokens_seen": 705167360 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039712136409227684, + "loss": 1.4269, + "theoretical_loss": 3.8050675818547526, + "tokens_seen": 705232896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971113340020061, + "loss": 1.5101, + "theoretical_loss": 3.8050319965116715, + "tokens_seen": 705298432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003971013039117352, + "loss": 1.4187, + "theoretical_loss": 3.804996415400753, + "tokens_seen": 705363968 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039709127382146444, + "loss": 1.408, + "theoretical_loss": 3.8049608385211, + "tokens_seen": 705429504 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039708124373119357, + "loss": 1.4394, + "theoretical_loss": 3.804925265871817, + "tokens_seen": 705495040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003970712136409228, + "loss": 1.5694, + "theoretical_loss": 3.804889697452007, + "tokens_seen": 705560576 + }, + { + "epoch": 0.21, + "learning_rate": 0.000397061183550652, + "loss": 1.5237, + "theoretical_loss": 3.8048541332607755, + "tokens_seen": 705626112 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039705115346038117, + "loss": 1.5361, + "theoretical_loss": 3.8048185732972266, + "tokens_seen": 705691648 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039704112337011035, + "loss": 1.4332, + "theoretical_loss": 3.8047830175604656, + "tokens_seen": 705757184 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039703109327983953, + "loss": 1.5323, + "theoretical_loss": 3.804747466049597, + "tokens_seen": 705822720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003970210631895687, + "loss": 1.574, + "theoretical_loss": 3.804711918763726, + "tokens_seen": 705888256 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039701103309929794, + "loss": 1.4812, + "theoretical_loss": 3.8046763757019595, + "tokens_seen": 705953792 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039700100300902707, + "loss": 1.4571, + "theoretical_loss": 3.8046408368634026, + "tokens_seen": 706019328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003969909729187563, + "loss": 1.5143, + "theoretical_loss": 3.804605302247162, + "tokens_seen": 706084864 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039698094282848543, + "loss": 1.5576, + "theoretical_loss": 3.8045697718523437, + "tokens_seen": 706150400 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039697091273821467, + "loss": 1.5847, + "theoretical_loss": 3.804534245678055, + "tokens_seen": 706215936 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039696088264794385, + "loss": 1.5641, + "theoretical_loss": 3.8044987237234027, + "tokens_seen": 706281472 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039695085255767303, + "loss": 1.4607, + "theoretical_loss": 3.804463205987494, + "tokens_seen": 706347008 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003969408224674022, + "loss": 1.5147, + "theoretical_loss": 3.804427692469437, + "tokens_seen": 706412544 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039693079237713145, + "loss": 1.4567, + "theoretical_loss": 3.8043921831683396, + "tokens_seen": 706478080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003969207622868606, + "loss": 1.5347, + "theoretical_loss": 3.804356678083309, + "tokens_seen": 706543616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003969107321965898, + "loss": 1.5732, + "theoretical_loss": 3.8043211772134544, + "tokens_seen": 706609152 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039690070210631894, + "loss": 1.5018, + "theoretical_loss": 3.804285680557885, + "tokens_seen": 706674688 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039689067201604817, + "loss": 1.4783, + "theoretical_loss": 3.8042501881157085, + "tokens_seen": 706740224 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039688064192577735, + "loss": 1.4979, + "theoretical_loss": 3.804214699886035, + "tokens_seen": 706805760 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039687061183550653, + "loss": 1.6108, + "theoretical_loss": 3.8041792158679737, + "tokens_seen": 706871296 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968605817452357, + "loss": 1.5698, + "theoretical_loss": 3.804143736060635, + "tokens_seen": 706936832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968505516549649, + "loss": 1.5022, + "theoretical_loss": 3.8041082604631278, + "tokens_seen": 707002368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968405215646941, + "loss": 1.4716, + "theoretical_loss": 3.804072789074563, + "tokens_seen": 707067904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968304914744233, + "loss": 1.5774, + "theoretical_loss": 3.8040373218940515, + "tokens_seen": 707133440 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039682046138415244, + "loss": 1.5747, + "theoretical_loss": 3.804001858920704, + "tokens_seen": 707198976 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968104312938817, + "loss": 1.4312, + "theoretical_loss": 3.8039664001536315, + "tokens_seen": 707264512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003968004012036108, + "loss": 1.4878, + "theoretical_loss": 3.8039309455919454, + "tokens_seen": 707330048 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039679037111334004, + "loss": 1.4193, + "theoretical_loss": 3.803895495234757, + "tokens_seen": 707395584 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967803410230692, + "loss": 1.5303, + "theoretical_loss": 3.8038600490811794, + "tokens_seen": 707461120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967703109327984, + "loss": 1.5505, + "theoretical_loss": 3.8038246071303234, + "tokens_seen": 707526656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967602808425276, + "loss": 1.4888, + "theoretical_loss": 3.803789169381303, + "tokens_seen": 707592192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967502507522568, + "loss": 1.495, + "theoretical_loss": 3.8037537358332294, + "tokens_seen": 707657728 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039674022066198594, + "loss": 1.5216, + "theoretical_loss": 3.8037183064852162, + "tokens_seen": 707723264 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 262675, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7008589506149292, + "objective/train/theoretical_loss": 3.8036828813363774, + "objective/train/tokens_used": 728248800, + "theoretical_loss": 3.8036828813363774, + "tokens_seen": 707788800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967301905717152, + "loss": 1.6322, + "theoretical_loss": 3.8036828813363774, + "tokens_seen": 707788800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967201604814443, + "loss": 1.4275, + "theoretical_loss": 3.8036474603858252, + "tokens_seen": 707854336 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039671013039117354, + "loss": 1.6217, + "theoretical_loss": 3.803612043632675, + "tokens_seen": 707919872 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003967001003009027, + "loss": 1.5083, + "theoretical_loss": 3.80357663107604, + "tokens_seen": 707985408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003966900702106319, + "loss": 1.4068, + "theoretical_loss": 3.803541222715034, + "tokens_seen": 708050944 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003966800401203611, + "loss": 1.5727, + "theoretical_loss": 3.803505818548773, + "tokens_seen": 708116480 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039667001003009027, + "loss": 1.6033, + "theoretical_loss": 3.803470418576371, + "tokens_seen": 708182016 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039665997993981945, + "loss": 1.53, + "theoretical_loss": 3.8034350227969433, + "tokens_seen": 708247552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003966499498495487, + "loss": 1.5454, + "theoretical_loss": 3.8033996312096052, + "tokens_seen": 708313088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003966399197592778, + "loss": 1.5364, + "theoretical_loss": 3.803364243813473, + "tokens_seen": 708378624 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039662988966900704, + "loss": 1.5708, + "theoretical_loss": 3.803328860607662, + "tokens_seen": 708444160 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039661985957873617, + "loss": 1.5607, + "theoretical_loss": 3.8032934815912895, + "tokens_seen": 708509696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003966098294884654, + "loss": 1.4935, + "theoretical_loss": 3.8032581067634705, + "tokens_seen": 708575232 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965997993981946, + "loss": 1.5317, + "theoretical_loss": 3.803222736123323, + "tokens_seen": 708640768 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039658976930792377, + "loss": 1.4193, + "theoretical_loss": 3.8031873696699634, + "tokens_seen": 708706304 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039657973921765295, + "loss": 1.6053, + "theoretical_loss": 3.8031520074025096, + "tokens_seen": 708771840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965697091273822, + "loss": 1.3589, + "theoretical_loss": 3.8031166493200788, + "tokens_seen": 708837376 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965596790371113, + "loss": 1.5222, + "theoretical_loss": 3.803081295421788, + "tokens_seen": 708902912 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039654964894684055, + "loss": 1.4692, + "theoretical_loss": 3.8030459457067574, + "tokens_seen": 708968448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965396188565697, + "loss": 1.6404, + "theoretical_loss": 3.803010600174103, + "tokens_seen": 709033984 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965295887662989, + "loss": 1.5465, + "theoretical_loss": 3.8029752588229453, + "tokens_seen": 709099520 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003965195586760281, + "loss": 1.4924, + "theoretical_loss": 3.8029399216524027, + "tokens_seen": 709165056 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039650952858575727, + "loss": 1.4944, + "theoretical_loss": 3.802904588661594, + "tokens_seen": 709230592 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039649949849548645, + "loss": 1.6032, + "theoretical_loss": 3.802869259849639, + "tokens_seen": 709296128 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039648946840521563, + "loss": 1.5484, + "theoretical_loss": 3.8028339352156575, + "tokens_seen": 709361664 + }, + { + "epoch": 0.21, + "learning_rate": 0.0003964794383149448, + "loss": 1.4097, + "theoretical_loss": 3.802798614758769, + "tokens_seen": 709427200 + }, + { + "epoch": 0.21, + "learning_rate": 0.00039646940822467405, + "loss": 1.6244, + "theoretical_loss": 3.8027632984780944, + "tokens_seen": 709492736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003964593781344032, + "loss": 1.5407, + "theoretical_loss": 3.8027279863727537, + "tokens_seen": 709558272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003964493480441324, + "loss": 1.506, + "theoretical_loss": 3.8026926784418675, + "tokens_seen": 709623808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003964393179538616, + "loss": 1.602, + "theoretical_loss": 3.802657374684558, + "tokens_seen": 709689344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003964292878635908, + "loss": 1.5465, + "theoretical_loss": 3.8026220750999453, + "tokens_seen": 709754880 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039641925777331996, + "loss": 1.4926, + "theoretical_loss": 3.802586779687152, + "tokens_seen": 709820416 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039640922768304914, + "loss": 1.6115, + "theoretical_loss": 3.8025514884452987, + "tokens_seen": 709885952 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039640922768304914, + "loss": 1.4656, + "theoretical_loss": 3.8025162013735088, + "tokens_seen": 709951488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963991975927783, + "loss": 1.5871, + "theoretical_loss": 3.802480918470904, + "tokens_seen": 710017024 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039638916750250755, + "loss": 1.4661, + "theoretical_loss": 3.8024456397366073, + "tokens_seen": 710082560 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039637913741223673, + "loss": 1.5116, + "theoretical_loss": 3.802410365169741, + "tokens_seen": 710148096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963691073219659, + "loss": 1.6321, + "theoretical_loss": 3.802375094769429, + "tokens_seen": 710213632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963590772316951, + "loss": 1.6381, + "theoretical_loss": 3.8023398285347945, + "tokens_seen": 710279168 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963490471414243, + "loss": 1.5634, + "theoretical_loss": 3.8023045664649606, + "tokens_seen": 710344704 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963390170511535, + "loss": 1.5452, + "theoretical_loss": 3.802269308559052, + "tokens_seen": 710410240 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039632898696088264, + "loss": 1.4718, + "theoretical_loss": 3.8022340548161933, + "tokens_seen": 710475776 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003963189568706119, + "loss": 1.6534, + "theoretical_loss": 3.8021988052355082, + "tokens_seen": 710541312 + }, + { + "epoch": 0.22, + "learning_rate": 0.000396308926780341, + "loss": 1.5824, + "theoretical_loss": 3.8021635598161216, + "tokens_seen": 710606848 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039629889669007024, + "loss": 1.5799, + "theoretical_loss": 3.802128318557159, + "tokens_seen": 710672384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962888665997994, + "loss": 1.6222, + "theoretical_loss": 3.802093081457745, + "tokens_seen": 710737920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962788365095286, + "loss": 1.5856, + "theoretical_loss": 3.8020578485170056, + "tokens_seen": 710803456 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962688064192578, + "loss": 1.4598, + "theoretical_loss": 3.8020226197340667, + "tokens_seen": 710868992 + }, + { + "epoch": 0.22, + "learning_rate": 0.000396258776328987, + "loss": 1.5972, + "theoretical_loss": 3.801987395108055, + "tokens_seen": 710934528 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039624874623871614, + "loss": 1.5353, + "theoretical_loss": 3.8019521746380955, + "tokens_seen": 711000064 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 263840, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.587327003479004, + "objective/train/theoretical_loss": 3.801916958323315, + "objective/train/tokens_used": 731525600, + "theoretical_loss": 3.801916958323315, + "tokens_seen": 711065600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962387161484454, + "loss": 1.5463, + "theoretical_loss": 3.801916958323315, + "tokens_seen": 711065600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962286860581745, + "loss": 1.4196, + "theoretical_loss": 3.8018817461628416, + "tokens_seen": 711131136 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039621865596790374, + "loss": 1.5352, + "theoretical_loss": 3.801846538155802, + "tokens_seen": 711196672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003962086258776329, + "loss": 1.598, + "theoretical_loss": 3.801811334301323, + "tokens_seen": 711262208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003961985957873621, + "loss": 1.5131, + "theoretical_loss": 3.8017761345985326, + "tokens_seen": 711327744 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003961885656970913, + "loss": 1.5692, + "theoretical_loss": 3.8017409390465584, + "tokens_seen": 711393280 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039617853560682047, + "loss": 1.5841, + "theoretical_loss": 3.8017057476445295, + "tokens_seen": 711458816 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039616850551654965, + "loss": 1.6001, + "theoretical_loss": 3.8016705603915737, + "tokens_seen": 711524352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003961584754262789, + "loss": 1.5312, + "theoretical_loss": 3.8016353772868197, + "tokens_seen": 711589888 + }, + { + "epoch": 0.22, + "learning_rate": 0.000396148445336008, + "loss": 1.4537, + "theoretical_loss": 3.801600198329397, + "tokens_seen": 711655424 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039613841524573724, + "loss": 1.4867, + "theoretical_loss": 3.8015650235184344, + "tokens_seen": 711720960 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039612838515546637, + "loss": 1.5206, + "theoretical_loss": 3.8015298528530614, + "tokens_seen": 711786496 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003961183550651956, + "loss": 1.5366, + "theoretical_loss": 3.801494686332408, + "tokens_seen": 711852032 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003961083249749248, + "loss": 1.521, + "theoretical_loss": 3.8014595239556046, + "tokens_seen": 711917568 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039609829488465397, + "loss": 1.6431, + "theoretical_loss": 3.8014243657217808, + "tokens_seen": 711983104 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039608826479438315, + "loss": 1.5259, + "theoretical_loss": 3.8013892116300676, + "tokens_seen": 712048640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003960782347041124, + "loss": 1.4851, + "theoretical_loss": 3.8013540616795956, + "tokens_seen": 712114176 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003960682046138415, + "loss": 1.5317, + "theoretical_loss": 3.801318915869496, + "tokens_seen": 712179712 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039605817452357075, + "loss": 1.5109, + "theoretical_loss": 3.8012837741989003, + "tokens_seen": 712245248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003960481444332999, + "loss": 1.5438, + "theoretical_loss": 3.80124863666694, + "tokens_seen": 712310784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003960381143430291, + "loss": 1.431, + "theoretical_loss": 3.8012135032727468, + "tokens_seen": 712376320 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003960280842527583, + "loss": 1.4945, + "theoretical_loss": 3.8011783740154534, + "tokens_seen": 712441856 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039601805416248747, + "loss": 1.4953, + "theoretical_loss": 3.801143248894192, + "tokens_seen": 712507392 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039600802407221665, + "loss": 1.5193, + "theoretical_loss": 3.801108127908095, + "tokens_seen": 712572928 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039599799398194583, + "loss": 1.467, + "theoretical_loss": 3.801073011056295, + "tokens_seen": 712638464 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395987963891675, + "loss": 1.6195, + "theoretical_loss": 3.801037898337926, + "tokens_seen": 712704000 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039597793380140425, + "loss": 1.4252, + "theoretical_loss": 3.801002789752121, + "tokens_seen": 712769536 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003959679037111334, + "loss": 1.6182, + "theoretical_loss": 3.800967685298014, + "tokens_seen": 712835072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003959578736208626, + "loss": 1.4588, + "theoretical_loss": 3.8009325849747384, + "tokens_seen": 712900608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003959478435305918, + "loss": 1.5748, + "theoretical_loss": 3.8008974887814295, + "tokens_seen": 712966144 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395937813440321, + "loss": 1.645, + "theoretical_loss": 3.8008623967172204, + "tokens_seen": 713031680 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039592778335005016, + "loss": 1.4662, + "theoretical_loss": 3.800827308781247, + "tokens_seen": 713097216 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039591775325977934, + "loss": 1.4408, + "theoretical_loss": 3.8007922249726445, + "tokens_seen": 713162752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003959077231695085, + "loss": 1.446, + "theoretical_loss": 3.8007571452905466, + "tokens_seen": 713228288 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039589769307923775, + "loss": 1.5687, + "theoretical_loss": 3.800722069734091, + "tokens_seen": 713293824 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003958876629889669, + "loss": 1.6129, + "theoretical_loss": 3.800686998302411, + "tokens_seen": 713359360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003958776328986961, + "loss": 1.5142, + "theoretical_loss": 3.8006519309946447, + "tokens_seen": 713424896 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039586760280842524, + "loss": 1.541, + "theoretical_loss": 3.8006168678099277, + "tokens_seen": 713490432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003958575727181545, + "loss": 1.5464, + "theoretical_loss": 3.800581808747397, + "tokens_seen": 713555968 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039584754262788366, + "loss": 1.367, + "theoretical_loss": 3.800546753806189, + "tokens_seen": 713621504 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039583751253761284, + "loss": 1.5287, + "theoretical_loss": 3.8005117029854403, + "tokens_seen": 713687040 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395827482447342, + "loss": 1.5421, + "theoretical_loss": 3.8004766562842893, + "tokens_seen": 713752576 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003958174523570712, + "loss": 1.4528, + "theoretical_loss": 3.800441613701873, + "tokens_seen": 713818112 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003958074222668004, + "loss": 1.5682, + "theoretical_loss": 3.8004065752373295, + "tokens_seen": 713883648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003957973921765296, + "loss": 1.5685, + "theoretical_loss": 3.800371540889797, + "tokens_seen": 713949184 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039578736208625875, + "loss": 1.5016, + "theoretical_loss": 3.800336510658414, + "tokens_seen": 714014720 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395777331995988, + "loss": 1.5389, + "theoretical_loss": 3.8003014845423184, + "tokens_seen": 714080256 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039576730190571716, + "loss": 1.5647, + "theoretical_loss": 3.80026646254065, + "tokens_seen": 714145792 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039575727181544634, + "loss": 1.5179, + "theoretical_loss": 3.800231444652548, + "tokens_seen": 714211328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003957472417251755, + "loss": 1.5775, + "theoretical_loss": 3.800196430877151, + "tokens_seen": 714276864 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 265036, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4154844284057617, + "objective/train/theoretical_loss": 3.8001614212135997, + "objective/train/tokens_used": 734802400, + "theoretical_loss": 3.8001614212135997, + "tokens_seen": 714342400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003957372116349047, + "loss": 1.524, + "theoretical_loss": 3.8001614212135997, + "tokens_seen": 714342400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003957271815446339, + "loss": 1.5476, + "theoretical_loss": 3.8001264156610333, + "tokens_seen": 714407936 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003957171514543631, + "loss": 1.4517, + "theoretical_loss": 3.800091414218592, + "tokens_seen": 714473472 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039570712136409225, + "loss": 1.5646, + "theoretical_loss": 3.800056416885417, + "tokens_seen": 714539008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003956970912738215, + "loss": 1.4649, + "theoretical_loss": 3.8000214236606484, + "tokens_seen": 714604544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003956870611835506, + "loss": 1.489, + "theoretical_loss": 3.7999864345434275, + "tokens_seen": 714670080 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039567703109327985, + "loss": 1.5064, + "theoretical_loss": 3.7999514495328954, + "tokens_seen": 714735616 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039566700100300903, + "loss": 1.5083, + "theoretical_loss": 3.7999164686281937, + "tokens_seen": 714801152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003956569709127382, + "loss": 1.572, + "theoretical_loss": 3.7998814918284642, + "tokens_seen": 714866688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003956469408224674, + "loss": 1.5172, + "theoretical_loss": 3.7998465191328483, + "tokens_seen": 714932224 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039563691073219657, + "loss": 1.4482, + "theoretical_loss": 3.7998115505404892, + "tokens_seen": 714997760 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003956268806419258, + "loss": 1.6902, + "theoretical_loss": 3.7997765860505295, + "tokens_seen": 715063296 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395616850551655, + "loss": 1.5299, + "theoretical_loss": 3.7997416256621106, + "tokens_seen": 715128832 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039560682046138417, + "loss": 1.5746, + "theoretical_loss": 3.7997066693743773, + "tokens_seen": 715194368 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039559679037111335, + "loss": 1.5032, + "theoretical_loss": 3.799671717186472, + "tokens_seen": 715259904 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003955867602808426, + "loss": 1.5093, + "theoretical_loss": 3.799636769097538, + "tokens_seen": 715325440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003955767301905717, + "loss": 1.5796, + "theoretical_loss": 3.7996018251067203, + "tokens_seen": 715390976 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039556670010030095, + "loss": 1.5731, + "theoretical_loss": 3.799566885213162, + "tokens_seen": 715456512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003955566700100301, + "loss": 1.5468, + "theoretical_loss": 3.799531949416007, + "tokens_seen": 715522048 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003955466399197593, + "loss": 1.4565, + "theoretical_loss": 3.799497017714401, + "tokens_seen": 715587584 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003955366098294885, + "loss": 1.4533, + "theoretical_loss": 3.7994620901074887, + "tokens_seen": 715653120 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039552657973921767, + "loss": 1.5934, + "theoretical_loss": 3.7994271665944144, + "tokens_seen": 715718656 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039552657973921767, + "loss": 1.4863, + "theoretical_loss": 3.799392247174324, + "tokens_seen": 715784192 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039551654964894685, + "loss": 1.4617, + "theoretical_loss": 3.799357331846364, + "tokens_seen": 715849728 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039550651955867603, + "loss": 1.5551, + "theoretical_loss": 3.7993224206096787, + "tokens_seen": 715915264 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003954964894684052, + "loss": 1.5827, + "theoretical_loss": 3.7992875134634154, + "tokens_seen": 715980800 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039548645937813445, + "loss": 1.5278, + "theoretical_loss": 3.79925261040672, + "tokens_seen": 716046336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003954764292878636, + "loss": 1.5045, + "theoretical_loss": 3.7992177114387387, + "tokens_seen": 716111872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003954663991975928, + "loss": 1.5014, + "theoretical_loss": 3.799182816558619, + "tokens_seen": 716177408 + }, + { + "epoch": 0.22, + "learning_rate": 0.000395456369107322, + "loss": 1.4864, + "theoretical_loss": 3.7991479257655083, + "tokens_seen": 716242944 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003954463390170512, + "loss": 1.4708, + "theoretical_loss": 3.7991130390585535, + "tokens_seen": 716308480 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039543630892678036, + "loss": 1.3939, + "theoretical_loss": 3.7990781564369023, + "tokens_seen": 716374016 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039542627883650954, + "loss": 1.4912, + "theoretical_loss": 3.7990432778997034, + "tokens_seen": 716439552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003954162487462387, + "loss": 1.5526, + "theoretical_loss": 3.799008403446104, + "tokens_seen": 716505088 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039540621865596795, + "loss": 1.4891, + "theoretical_loss": 3.798973533075253, + "tokens_seen": 716570624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953961885656971, + "loss": 1.53, + "theoretical_loss": 3.798938666786299, + "tokens_seen": 716636160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953861584754263, + "loss": 1.67, + "theoretical_loss": 3.798903804578391, + "tokens_seen": 716701696 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039537612838515544, + "loss": 1.5215, + "theoretical_loss": 3.7988689464506784, + "tokens_seen": 716767232 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953660982948847, + "loss": 1.449, + "theoretical_loss": 3.798834092402311, + "tokens_seen": 716832768 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039535606820461386, + "loss": 1.4984, + "theoretical_loss": 3.7987992424324366, + "tokens_seen": 716898304 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039534603811434304, + "loss": 1.5756, + "theoretical_loss": 3.7987643965402076, + "tokens_seen": 716963840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953360080240722, + "loss": 1.5454, + "theoretical_loss": 3.7987295547247735, + "tokens_seen": 717029376 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953259779338014, + "loss": 1.587, + "theoretical_loss": 3.7986947169852834, + "tokens_seen": 717094912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953159478435306, + "loss": 1.6002, + "theoretical_loss": 3.79865988332089, + "tokens_seen": 717160448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003953059177532598, + "loss": 1.5798, + "theoretical_loss": 3.798625053730743, + "tokens_seen": 717225984 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039529588766298895, + "loss": 1.5469, + "theoretical_loss": 3.7985902282139943, + "tokens_seen": 717291520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952858575727182, + "loss": 1.5554, + "theoretical_loss": 3.7985554067697946, + "tokens_seen": 717357056 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039527582748244736, + "loss": 1.5265, + "theoretical_loss": 3.798520589397297, + "tokens_seen": 717422592 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039526579739217654, + "loss": 1.6892, + "theoretical_loss": 3.798485776095652, + "tokens_seen": 717488128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952557673019057, + "loss": 1.4746, + "theoretical_loss": 3.7984509668640136, + "tokens_seen": 717553664 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 266305, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5381361246109009, + "objective/train/theoretical_loss": 3.7984161617015326, + "objective/train/tokens_used": 738079200, + "theoretical_loss": 3.7984161617015326, + "tokens_seen": 717619200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952457372116349, + "loss": 1.4772, + "theoretical_loss": 3.7984161617015326, + "tokens_seen": 717619200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952357071213641, + "loss": 1.5554, + "theoretical_loss": 3.7983813606073626, + "tokens_seen": 717684736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952256770310933, + "loss": 1.5352, + "theoretical_loss": 3.7983465635806564, + "tokens_seen": 717750272 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039521564694082245, + "loss": 1.6446, + "theoretical_loss": 3.7983117706205674, + "tokens_seen": 717815808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003952056168505517, + "loss": 1.6348, + "theoretical_loss": 3.798276981726249, + "tokens_seen": 717881344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003951955867602808, + "loss": 1.4971, + "theoretical_loss": 3.798242196896856, + "tokens_seen": 717946880 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039518555667001005, + "loss": 1.5586, + "theoretical_loss": 3.798207416131541, + "tokens_seen": 718012416 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039517552657973923, + "loss": 1.5208, + "theoretical_loss": 3.798172639429459, + "tokens_seen": 718077952 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003951654964894684, + "loss": 1.4894, + "theoretical_loss": 3.798137866789764, + "tokens_seen": 718143488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003951554663991976, + "loss": 1.5253, + "theoretical_loss": 3.798103098211611, + "tokens_seen": 718209024 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039514543630892677, + "loss": 1.512, + "theoretical_loss": 3.798068333694156, + "tokens_seen": 718274560 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039513540621865595, + "loss": 1.5468, + "theoretical_loss": 3.798033573236553, + "tokens_seen": 718340096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003951253761283852, + "loss": 1.4569, + "theoretical_loss": 3.797998816837959, + "tokens_seen": 718405632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003951153460381143, + "loss": 1.5414, + "theoretical_loss": 3.7979640644975277, + "tokens_seen": 718471168 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039510531594784355, + "loss": 1.4631, + "theoretical_loss": 3.797929316214417, + "tokens_seen": 718536704 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039509528585757273, + "loss": 1.4122, + "theoretical_loss": 3.797894571987783, + "tokens_seen": 718602240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950852557673019, + "loss": 1.4833, + "theoretical_loss": 3.7978598318167816, + "tokens_seen": 718667776 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950752256770311, + "loss": 1.5119, + "theoretical_loss": 3.7978250957005697, + "tokens_seen": 718733312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950651955867603, + "loss": 1.5265, + "theoretical_loss": 3.7977903636383044, + "tokens_seen": 718798848 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039505516549648946, + "loss": 1.4401, + "theoretical_loss": 3.7977556356291435, + "tokens_seen": 718864384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950451354062187, + "loss": 1.5696, + "theoretical_loss": 3.7977209116722443, + "tokens_seen": 718929920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950351053159478, + "loss": 1.4927, + "theoretical_loss": 3.797686191766765, + "tokens_seen": 718995456 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039502507522567705, + "loss": 1.5037, + "theoretical_loss": 3.797651475911863, + "tokens_seen": 719060992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950150451354062, + "loss": 1.5261, + "theoretical_loss": 3.7976167641066962, + "tokens_seen": 719126528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003950050150451354, + "loss": 1.449, + "theoretical_loss": 3.7975820563504246, + "tokens_seen": 719192064 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949949849548646, + "loss": 1.6264, + "theoretical_loss": 3.797547352642206, + "tokens_seen": 719257600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949849548645938, + "loss": 1.4921, + "theoretical_loss": 3.7975126529812, + "tokens_seen": 719323136 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039497492477432296, + "loss": 1.4697, + "theoretical_loss": 3.797477957366566, + "tokens_seen": 719388672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949648946840522, + "loss": 1.5167, + "theoretical_loss": 3.797443265797463, + "tokens_seen": 719454208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949548645937813, + "loss": 1.6177, + "theoretical_loss": 3.7974085782730516, + "tokens_seen": 719519744 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039494483450351056, + "loss": 1.4148, + "theoretical_loss": 3.797373894792491, + "tokens_seen": 719585280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949348044132397, + "loss": 1.4987, + "theoretical_loss": 3.797339215354942, + "tokens_seen": 719650816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949247743229689, + "loss": 1.575, + "theoretical_loss": 3.7973045399595655, + "tokens_seen": 719716352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949147442326981, + "loss": 1.5159, + "theoretical_loss": 3.797269868605522, + "tokens_seen": 719781888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003949047141424273, + "loss": 1.4943, + "theoretical_loss": 3.797235201291973, + "tokens_seen": 719847424 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003948946840521565, + "loss": 1.5376, + "theoretical_loss": 3.797200538018079, + "tokens_seen": 719912960 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039488465396188564, + "loss": 1.5101, + "theoretical_loss": 3.7971658787830025, + "tokens_seen": 719978496 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003948746238716149, + "loss": 1.5393, + "theoretical_loss": 3.7971312235859047, + "tokens_seen": 720044032 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039486459378134406, + "loss": 1.5957, + "theoretical_loss": 3.797096572425948, + "tokens_seen": 720109568 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039485456369107324, + "loss": 1.5611, + "theoretical_loss": 3.7970619253022946, + "tokens_seen": 720175104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003948445336008024, + "loss": 1.4338, + "theoretical_loss": 3.797027282214107, + "tokens_seen": 720240640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003948345035105316, + "loss": 1.4413, + "theoretical_loss": 3.7969926431605483, + "tokens_seen": 720306176 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003948244734202608, + "loss": 1.4324, + "theoretical_loss": 3.796958008140782, + "tokens_seen": 720371712 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039481444332999, + "loss": 1.5103, + "theoretical_loss": 3.7969233771539708, + "tokens_seen": 720437248 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039480441323971915, + "loss": 1.5447, + "theoretical_loss": 3.796888750199278, + "tokens_seen": 720502784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947943831494484, + "loss": 1.5031, + "theoretical_loss": 3.796854127275868, + "tokens_seen": 720568320 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039478435305917756, + "loss": 1.5544, + "theoretical_loss": 3.796819508382905, + "tokens_seen": 720633856 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039477432296890674, + "loss": 1.4908, + "theoretical_loss": 3.796784893519553, + "tokens_seen": 720699392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947642928786359, + "loss": 1.6107, + "theoretical_loss": 3.796750282684977, + "tokens_seen": 720764928 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947542627883651, + "loss": 1.3859, + "theoretical_loss": 3.7967156758783416, + "tokens_seen": 720830464 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 267624, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.417776346206665, + "objective/train/theoretical_loss": 3.7966810730988114, + "objective/train/tokens_used": 741356000, + "theoretical_loss": 3.7966810730988114, + "tokens_seen": 720896000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947442326980943, + "loss": 1.4087, + "theoretical_loss": 3.7966810730988114, + "tokens_seen": 720896000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947342026078235, + "loss": 1.4411, + "theoretical_loss": 3.7966464743455526, + "tokens_seen": 720961536 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039472417251755265, + "loss": 1.5054, + "theoretical_loss": 3.79661187961773, + "tokens_seen": 721027072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003947141424272819, + "loss": 1.4277, + "theoretical_loss": 3.79657728891451, + "tokens_seen": 721092608 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394704112337011, + "loss": 1.5197, + "theoretical_loss": 3.796542702235059, + "tokens_seen": 721158144 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039469408224674025, + "loss": 1.4742, + "theoretical_loss": 3.7965081195785424, + "tokens_seen": 721223680 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039468405215646943, + "loss": 1.522, + "theoretical_loss": 3.7964735409441275, + "tokens_seen": 721289216 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003946740220661986, + "loss": 1.4595, + "theoretical_loss": 3.7964389663309808, + "tokens_seen": 721354752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003946639919759278, + "loss": 1.5607, + "theoretical_loss": 3.796404395738269, + "tokens_seen": 721420288 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039465396188565697, + "loss": 1.488, + "theoretical_loss": 3.79636982916516, + "tokens_seen": 721485824 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039464393179538615, + "loss": 1.5221, + "theoretical_loss": 3.7963352666108214, + "tokens_seen": 721551360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003946339017051154, + "loss": 1.6115, + "theoretical_loss": 3.796300708074421, + "tokens_seen": 721616896 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003946238716148445, + "loss": 1.5935, + "theoretical_loss": 3.796266153555127, + "tokens_seen": 721682432 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039461384152457375, + "loss": 1.5963, + "theoretical_loss": 3.7962316030521066, + "tokens_seen": 721747968 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039460381143430293, + "loss": 1.4857, + "theoretical_loss": 3.79619705656453, + "tokens_seen": 721813504 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945937813440321, + "loss": 1.5041, + "theoretical_loss": 3.7961625140915647, + "tokens_seen": 721879040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945837512537613, + "loss": 1.4512, + "theoretical_loss": 3.7961279756323805, + "tokens_seen": 721944576 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945737211634905, + "loss": 1.5226, + "theoretical_loss": 3.7960934411861462, + "tokens_seen": 722010112 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039456369107321966, + "loss": 1.5771, + "theoretical_loss": 3.7960589107520324, + "tokens_seen": 722075648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945536609829489, + "loss": 1.4804, + "theoretical_loss": 3.7960243843292076, + "tokens_seen": 722141184 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394543630892678, + "loss": 1.5741, + "theoretical_loss": 3.795989861916843, + "tokens_seen": 722206720 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039453360080240725, + "loss": 1.4852, + "theoretical_loss": 3.7959553435141076, + "tokens_seen": 722272256 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945235707121364, + "loss": 1.5415, + "theoretical_loss": 3.795920829120173, + "tokens_seen": 722337792 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945135406218656, + "loss": 1.4113, + "theoretical_loss": 3.79588631873421, + "tokens_seen": 722403328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003945035105315948, + "loss": 1.4202, + "theoretical_loss": 3.7958518123553886, + "tokens_seen": 722468864 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394493480441324, + "loss": 1.504, + "theoretical_loss": 3.7958173099828816, + "tokens_seen": 722534400 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039448345035105316, + "loss": 1.5552, + "theoretical_loss": 3.795782811615859, + "tokens_seen": 722599936 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944734202607824, + "loss": 1.5105, + "theoretical_loss": 3.7957483172534934, + "tokens_seen": 722665472 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944633901705115, + "loss": 1.4843, + "theoretical_loss": 3.795713826894957, + "tokens_seen": 722731008 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039445336008024076, + "loss": 1.6662, + "theoretical_loss": 3.7956793405394214, + "tokens_seen": 722796544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944433299899699, + "loss": 1.6062, + "theoretical_loss": 3.79564485818606, + "tokens_seen": 722862080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944332998996991, + "loss": 1.5564, + "theoretical_loss": 3.7956103798340446, + "tokens_seen": 722927616 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944232698094283, + "loss": 1.5298, + "theoretical_loss": 3.795575905482549, + "tokens_seen": 722993152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003944132397191575, + "loss": 1.4987, + "theoretical_loss": 3.7955414351307457, + "tokens_seen": 723058688 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039440320962888666, + "loss": 1.5616, + "theoretical_loss": 3.7955069687778087, + "tokens_seen": 723124224 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039439317953861584, + "loss": 1.6023, + "theoretical_loss": 3.795472506422912, + "tokens_seen": 723189760 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394383149448345, + "loss": 1.5597, + "theoretical_loss": 3.795438048065229, + "tokens_seen": 723255296 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039437311935807426, + "loss": 1.5794, + "theoretical_loss": 3.7954035937039348, + "tokens_seen": 723320832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003943630892678034, + "loss": 1.5608, + "theoretical_loss": 3.7953691433382026, + "tokens_seen": 723386368 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003943530591775326, + "loss": 1.5689, + "theoretical_loss": 3.795334696967208, + "tokens_seen": 723451904 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039434302908726175, + "loss": 1.5174, + "theoretical_loss": 3.795300254590126, + "tokens_seen": 723517440 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394332998996991, + "loss": 1.4986, + "theoretical_loss": 3.7952658162061317, + "tokens_seen": 723582976 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039432296890672017, + "loss": 1.3988, + "theoretical_loss": 3.7952313818144, + "tokens_seen": 723648512 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039431293881644935, + "loss": 1.5825, + "theoretical_loss": 3.7951969514141073, + "tokens_seen": 723714048 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039430290872617853, + "loss": 1.4942, + "theoretical_loss": 3.795162525004429, + "tokens_seen": 723779584 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039429287863590776, + "loss": 1.4793, + "theoretical_loss": 3.795128102584542, + "tokens_seen": 723845120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003942828485456369, + "loss": 1.4874, + "theoretical_loss": 3.7950936841536222, + "tokens_seen": 723910656 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003942728184553661, + "loss": 1.5351, + "theoretical_loss": 3.7950592697108467, + "tokens_seen": 723976192 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039426278836509525, + "loss": 1.5316, + "theoretical_loss": 3.795024859255392, + "tokens_seen": 724041728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003942527582748245, + "loss": 1.5055, + "theoretical_loss": 3.7949904527864353, + "tokens_seen": 724107264 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 268786, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.531082034111023, + "objective/train/theoretical_loss": 3.7949560503031545, + "objective/train/tokens_used": 744632800, + "theoretical_loss": 3.7949560503031545, + "tokens_seen": 724172800 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039424272818455367, + "loss": 1.5606, + "theoretical_loss": 3.7949560503031545, + "tokens_seen": 724172800 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039423269809428285, + "loss": 1.5689, + "theoretical_loss": 3.794921651804727, + "tokens_seen": 724238336 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039422266800401203, + "loss": 1.7265, + "theoretical_loss": 3.7948872572903305, + "tokens_seen": 724303872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003942126379137412, + "loss": 1.4542, + "theoretical_loss": 3.7948528667591432, + "tokens_seen": 724369408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003942026078234704, + "loss": 1.4818, + "theoretical_loss": 3.7948184802103437, + "tokens_seen": 724434944 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039419257773319963, + "loss": 1.5405, + "theoretical_loss": 3.7947840976431104, + "tokens_seen": 724500480 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039418254764292876, + "loss": 1.4424, + "theoretical_loss": 3.7947497190566226, + "tokens_seen": 724566016 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394172517552658, + "loss": 1.4916, + "theoretical_loss": 3.7947153444500588, + "tokens_seen": 724631552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003941624874623871, + "loss": 1.5577, + "theoretical_loss": 3.794680973822599, + "tokens_seen": 724697088 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039415245737211635, + "loss": 1.4234, + "theoretical_loss": 3.7946466071734224, + "tokens_seen": 724762624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003941424272818456, + "loss": 1.5512, + "theoretical_loss": 3.794612244501709, + "tokens_seen": 724828160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003941323971915747, + "loss": 1.4969, + "theoretical_loss": 3.794577885806639, + "tokens_seen": 724893696 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039412236710130395, + "loss": 1.4917, + "theoretical_loss": 3.7945435310873927, + "tokens_seen": 724959232 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039411233701103313, + "loss": 1.5808, + "theoretical_loss": 3.7945091803431503, + "tokens_seen": 725024768 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003941023069207623, + "loss": 1.5147, + "theoretical_loss": 3.7944748335730933, + "tokens_seen": 725090304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940922768304915, + "loss": 1.5503, + "theoretical_loss": 3.794440490776402, + "tokens_seen": 725155840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940822467402207, + "loss": 1.5176, + "theoretical_loss": 3.7944061519522583, + "tokens_seen": 725221376 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039407221664994986, + "loss": 1.4347, + "theoretical_loss": 3.7943718170998437, + "tokens_seen": 725286912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940621865596791, + "loss": 1.4442, + "theoretical_loss": 3.79433748621834, + "tokens_seen": 725352448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940521564694082, + "loss": 1.5578, + "theoretical_loss": 3.7943031593069287, + "tokens_seen": 725417984 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039404212637913745, + "loss": 1.5089, + "theoretical_loss": 3.794268836364793, + "tokens_seen": 725483520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940320962888666, + "loss": 1.4921, + "theoretical_loss": 3.7942345173911147, + "tokens_seen": 725549056 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940220661985958, + "loss": 1.4478, + "theoretical_loss": 3.7942002023850767, + "tokens_seen": 725614592 + }, + { + "epoch": 0.22, + "learning_rate": 0.000394012036108325, + "loss": 1.4128, + "theoretical_loss": 3.7941658913458625, + "tokens_seen": 725680128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003940020060180542, + "loss": 1.5038, + "theoretical_loss": 3.794131584272655, + "tokens_seen": 725745664 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039399197592778336, + "loss": 1.4212, + "theoretical_loss": 3.794097281164637, + "tokens_seen": 725811200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939819458375126, + "loss": 1.5254, + "theoretical_loss": 3.7940629820209937, + "tokens_seen": 725876736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939719157472417, + "loss": 1.517, + "theoretical_loss": 3.794028686840908, + "tokens_seen": 725942272 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039396188565697096, + "loss": 1.5037, + "theoretical_loss": 3.7939943956235647, + "tokens_seen": 726007808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939518555667001, + "loss": 1.4749, + "theoretical_loss": 3.7939601083681476, + "tokens_seen": 726073344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939418254764293, + "loss": 1.567, + "theoretical_loss": 3.793925825073842, + "tokens_seen": 726138880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939317953861585, + "loss": 1.5029, + "theoretical_loss": 3.7938915457398323, + "tokens_seen": 726204416 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003939217652958877, + "loss": 1.5357, + "theoretical_loss": 3.793857270365304, + "tokens_seen": 726269952 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039391173520561686, + "loss": 1.5328, + "theoretical_loss": 3.7938229989494427, + "tokens_seen": 726335488 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039390170511534604, + "loss": 1.4514, + "theoretical_loss": 3.793788731491434, + "tokens_seen": 726401024 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003938916750250752, + "loss": 1.5503, + "theoretical_loss": 3.7937544679904636, + "tokens_seen": 726466560 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039388164493480446, + "loss": 1.5813, + "theoretical_loss": 3.793720208445718, + "tokens_seen": 726532096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003938716148445336, + "loss": 1.4293, + "theoretical_loss": 3.793685952856383, + "tokens_seen": 726597632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003938615847542628, + "loss": 1.3642, + "theoretical_loss": 3.7936517012216457, + "tokens_seen": 726663168 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039385155466399195, + "loss": 1.5715, + "theoretical_loss": 3.7936174535406932, + "tokens_seen": 726728704 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003938415245737212, + "loss": 1.4726, + "theoretical_loss": 3.793583209812712, + "tokens_seen": 726794240 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039383149448345037, + "loss": 1.4515, + "theoretical_loss": 3.7935489700368894, + "tokens_seen": 726859776 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039382146439317955, + "loss": 1.5186, + "theoretical_loss": 3.7935147342124136, + "tokens_seen": 726925312 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039381143430290873, + "loss": 1.4528, + "theoretical_loss": 3.7934805023384723, + "tokens_seen": 726990848 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039380140421263796, + "loss": 1.6193, + "theoretical_loss": 3.7934462744142534, + "tokens_seen": 727056384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003937913741223671, + "loss": 1.6508, + "theoretical_loss": 3.7934120504389455, + "tokens_seen": 727121920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003937813440320963, + "loss": 1.5223, + "theoretical_loss": 3.7933778304117363, + "tokens_seen": 727187456 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039377131394182545, + "loss": 1.5733, + "theoretical_loss": 3.7933436143318153, + "tokens_seen": 727252992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003937612838515547, + "loss": 1.5237, + "theoretical_loss": 3.7933094021983713, + "tokens_seen": 727318528 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039375125376128387, + "loss": 1.4987, + "theoretical_loss": 3.793275194010594, + "tokens_seen": 727384064 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 269904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5349664688110352, + "objective/train/theoretical_loss": 3.7932409897676727, + "objective/train/tokens_used": 747909600, + "theoretical_loss": 3.7932409897676727, + "tokens_seen": 727449600 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039374122367101305, + "loss": 1.5457, + "theoretical_loss": 3.7932409897676727, + "tokens_seen": 727449600 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039373119358074223, + "loss": 1.4102, + "theoretical_loss": 3.793206789468797, + "tokens_seen": 727515136 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003937211634904714, + "loss": 1.5596, + "theoretical_loss": 3.793172593113157, + "tokens_seen": 727580672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003937111334002006, + "loss": 1.4925, + "theoretical_loss": 3.7931384006999425, + "tokens_seen": 727646208 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039370110330992983, + "loss": 1.473, + "theoretical_loss": 3.7931042122283447, + "tokens_seen": 727711744 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039369107321965896, + "loss": 1.5692, + "theoretical_loss": 3.7930700276975537, + "tokens_seen": 727777280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003936810431293882, + "loss": 1.4736, + "theoretical_loss": 3.793035847106761, + "tokens_seen": 727842816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003936710130391173, + "loss": 1.5667, + "theoretical_loss": 3.7930016704551575, + "tokens_seen": 727908352 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039366098294884655, + "loss": 1.426, + "theoretical_loss": 3.7929674977419343, + "tokens_seen": 727973888 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039365095285857573, + "loss": 1.4418, + "theoretical_loss": 3.792933328966283, + "tokens_seen": 728039424 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003936409227683049, + "loss": 1.4573, + "theoretical_loss": 3.792899164127397, + "tokens_seen": 728104960 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003936308926780341, + "loss": 1.4695, + "theoretical_loss": 3.792865003224467, + "tokens_seen": 728170496 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039362086258776333, + "loss": 1.5472, + "theoretical_loss": 3.792830846256685, + "tokens_seen": 728236032 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039361083249749246, + "loss": 1.4792, + "theoretical_loss": 3.792796693223245, + "tokens_seen": 728301568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003936008024072217, + "loss": 1.4597, + "theoretical_loss": 3.792762544123339, + "tokens_seen": 728367104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935907723169508, + "loss": 1.4483, + "theoretical_loss": 3.79272839895616, + "tokens_seen": 728432640 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039358074222668006, + "loss": 1.532, + "theoretical_loss": 3.792694257720902, + "tokens_seen": 728498176 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039357071213640924, + "loss": 1.5761, + "theoretical_loss": 3.792660120416758, + "tokens_seen": 728563712 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935606820461384, + "loss": 1.5043, + "theoretical_loss": 3.792625987042922, + "tokens_seen": 728629248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935506519558676, + "loss": 1.4704, + "theoretical_loss": 3.7925918575985884, + "tokens_seen": 728694784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935406218655968, + "loss": 1.49, + "theoretical_loss": 3.7925577320829507, + "tokens_seen": 728760320 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039353059177532596, + "loss": 1.5017, + "theoretical_loss": 3.792523610495204, + "tokens_seen": 728825856 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935205616850552, + "loss": 1.468, + "theoretical_loss": 3.7924894928345427, + "tokens_seen": 728891392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003935105315947843, + "loss": 1.5041, + "theoretical_loss": 3.7924553791001623, + "tokens_seen": 728956928 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039350050150451356, + "loss": 1.4807, + "theoretical_loss": 3.792421269291258, + "tokens_seen": 729022464 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934904714142427, + "loss": 1.5047, + "theoretical_loss": 3.7923871634070245, + "tokens_seen": 729088000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934804413239719, + "loss": 1.5676, + "theoretical_loss": 3.792353061446658, + "tokens_seen": 729153536 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934704112337011, + "loss": 1.4449, + "theoretical_loss": 3.792318963409355, + "tokens_seen": 729219072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934603811434303, + "loss": 1.5366, + "theoretical_loss": 3.7922848692943103, + "tokens_seen": 729284608 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039345035105315946, + "loss": 1.4421, + "theoretical_loss": 3.792250779100722, + "tokens_seen": 729350144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934403209628887, + "loss": 1.4778, + "theoretical_loss": 3.792216692827785, + "tokens_seen": 729415680 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039343029087261783, + "loss": 1.4297, + "theoretical_loss": 3.792182610474698, + "tokens_seen": 729481216 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039342026078234706, + "loss": 1.5122, + "theoretical_loss": 3.792148532040657, + "tokens_seen": 729546752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934102306920762, + "loss": 1.5958, + "theoretical_loss": 3.7921144575248595, + "tokens_seen": 729612288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003934002006018054, + "loss": 1.3635, + "theoretical_loss": 3.7920803869265027, + "tokens_seen": 729677824 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039339017051153466, + "loss": 1.489, + "theoretical_loss": 3.792046320244786, + "tokens_seen": 729743360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003933801404212638, + "loss": 1.4509, + "theoretical_loss": 3.792012257478906, + "tokens_seen": 729808896 + }, + { + "epoch": 0.22, + "learning_rate": 0.000393370110330993, + "loss": 1.3684, + "theoretical_loss": 3.791978198628061, + "tokens_seen": 729874432 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039336008024072215, + "loss": 1.4606, + "theoretical_loss": 3.7919441436914507, + "tokens_seen": 729939968 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003933500501504514, + "loss": 1.4337, + "theoretical_loss": 3.7919100926682727, + "tokens_seen": 730005504 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039334002006018057, + "loss": 1.5243, + "theoretical_loss": 3.7918760455577267, + "tokens_seen": 730071040 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039332998996990975, + "loss": 1.4568, + "theoretical_loss": 3.791842002359012, + "tokens_seen": 730136576 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039331995987963893, + "loss": 1.4305, + "theoretical_loss": 3.7918079630713275, + "tokens_seen": 730202112 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039330992978936816, + "loss": 1.5507, + "theoretical_loss": 3.7917739276938733, + "tokens_seen": 730267648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003932998996990973, + "loss": 1.438, + "theoretical_loss": 3.7917398962258493, + "tokens_seen": 730333184 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003932898696088265, + "loss": 1.5348, + "theoretical_loss": 3.791705868666456, + "tokens_seen": 730398720 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039327983951855565, + "loss": 1.3839, + "theoretical_loss": 3.791671845014893, + "tokens_seen": 730464256 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003932698094282849, + "loss": 1.3751, + "theoretical_loss": 3.7916378252703624, + "tokens_seen": 730529792 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039325977933801407, + "loss": 1.6164, + "theoretical_loss": 3.791603809432064, + "tokens_seen": 730595328 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039324974924774325, + "loss": 1.4685, + "theoretical_loss": 3.7915697974991986, + "tokens_seen": 730660864 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 271086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4659218788146973, + "objective/train/theoretical_loss": 3.7915357894709687, + "objective/train/tokens_used": 751186400, + "theoretical_loss": 3.7915357894709687, + "tokens_seen": 730726400 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039323971915747243, + "loss": 1.4561, + "theoretical_loss": 3.7915357894709687, + "tokens_seen": 730726400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003932296890672016, + "loss": 1.4017, + "theoretical_loss": 3.791501785346575, + "tokens_seen": 730791936 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003932196589769308, + "loss": 1.466, + "theoretical_loss": 3.79146778512522, + "tokens_seen": 730857472 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039320962888666003, + "loss": 1.5207, + "theoretical_loss": 3.7914337888061054, + "tokens_seen": 730923008 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039319959879638916, + "loss": 1.4496, + "theoretical_loss": 3.7913997963884336, + "tokens_seen": 730988544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003931895687061184, + "loss": 1.4741, + "theoretical_loss": 3.791365807871407, + "tokens_seen": 731054080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003931795386158475, + "loss": 1.5812, + "theoretical_loss": 3.791331823254229, + "tokens_seen": 731119616 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039316950852557675, + "loss": 1.5202, + "theoretical_loss": 3.791297842536102, + "tokens_seen": 731185152 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039315947843530593, + "loss": 1.4048, + "theoretical_loss": 3.791263865716229, + "tokens_seen": 731250688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003931494483450351, + "loss": 1.5316, + "theoretical_loss": 3.7912298927938144, + "tokens_seen": 731316224 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003931394182547643, + "loss": 1.5131, + "theoretical_loss": 3.7911959237680612, + "tokens_seen": 731381760 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039312938816449353, + "loss": 1.4444, + "theoretical_loss": 3.791161958638173, + "tokens_seen": 731447296 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039311935807422266, + "loss": 1.5441, + "theoretical_loss": 3.7911279974033554, + "tokens_seen": 731512832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003931093279839519, + "loss": 1.4239, + "theoretical_loss": 3.7910940400628115, + "tokens_seen": 731578368 + }, + { + "epoch": 0.22, + "learning_rate": 0.000393099297893681, + "loss": 1.3892, + "theoretical_loss": 3.7910600866157464, + "tokens_seen": 731643904 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039308926780341026, + "loss": 1.4547, + "theoretical_loss": 3.7910261370613654, + "tokens_seen": 731709440 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039307923771313944, + "loss": 1.4892, + "theoretical_loss": 3.790992191398873, + "tokens_seen": 731774976 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003930692076228686, + "loss": 1.5204, + "theoretical_loss": 3.790958249627475, + "tokens_seen": 731840512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003930591775325978, + "loss": 1.5741, + "theoretical_loss": 3.7909243117463767, + "tokens_seen": 731906048 + }, + { + "epoch": 0.22, + "learning_rate": 0.000393049147442327, + "loss": 1.4272, + "theoretical_loss": 3.7908903777547835, + "tokens_seen": 731971584 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039303911735205616, + "loss": 1.4294, + "theoretical_loss": 3.7908564476519024, + "tokens_seen": 732037120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003930290872617854, + "loss": 1.507, + "theoretical_loss": 3.7908225214369393, + "tokens_seen": 732102656 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003930190571715145, + "loss": 1.5361, + "theoretical_loss": 3.7907885991091, + "tokens_seen": 732168192 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039300902708124376, + "loss": 1.4546, + "theoretical_loss": 3.790754680667593, + "tokens_seen": 732233728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929989969909729, + "loss": 1.5004, + "theoretical_loss": 3.7907207661116233, + "tokens_seen": 732299264 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929889669007021, + "loss": 1.536, + "theoretical_loss": 3.7906868554403994, + "tokens_seen": 732364800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929789368104313, + "loss": 1.3986, + "theoretical_loss": 3.7906529486531286, + "tokens_seen": 732430336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929689067201605, + "loss": 1.4848, + "theoretical_loss": 3.7906190457490183, + "tokens_seen": 732495872 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039295887662988967, + "loss": 1.4363, + "theoretical_loss": 3.790585146727276, + "tokens_seen": 732561408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929488465396189, + "loss": 1.4604, + "theoretical_loss": 3.790551251587111, + "tokens_seen": 732626944 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039293881644934803, + "loss": 1.498, + "theoretical_loss": 3.7905173603277307, + "tokens_seen": 732692480 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039292878635907726, + "loss": 1.4238, + "theoretical_loss": 3.790483472948344, + "tokens_seen": 732758016 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929187562688064, + "loss": 1.3846, + "theoretical_loss": 3.79044958944816, + "tokens_seen": 732823552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003929087261785356, + "loss": 1.6318, + "theoretical_loss": 3.7904157098263873, + "tokens_seen": 732889088 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003928986960882648, + "loss": 1.5715, + "theoretical_loss": 3.7903818340822353, + "tokens_seen": 732954624 + }, + { + "epoch": 0.22, + "learning_rate": 0.000392888665997994, + "loss": 1.5173, + "theoretical_loss": 3.790347962214914, + "tokens_seen": 733020160 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039287863590772317, + "loss": 1.427, + "theoretical_loss": 3.790314094223633, + "tokens_seen": 733085696 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039286860581745235, + "loss": 1.5234, + "theoretical_loss": 3.7902802301076024, + "tokens_seen": 733151232 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039285857572718153, + "loss": 1.4997, + "theoretical_loss": 3.790246369866032, + "tokens_seen": 733216768 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039284854563691077, + "loss": 1.4647, + "theoretical_loss": 3.790212513498132, + "tokens_seen": 733282304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003928385155466399, + "loss": 1.578, + "theoretical_loss": 3.7901786610031145, + "tokens_seen": 733347840 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039282848545636913, + "loss": 1.4576, + "theoretical_loss": 3.7901448123801886, + "tokens_seen": 733413376 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039281845536609825, + "loss": 1.5203, + "theoretical_loss": 3.790110967628567, + "tokens_seen": 733478912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003928084252758275, + "loss": 1.5591, + "theoretical_loss": 3.7900771267474607, + "tokens_seen": 733544448 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039279839518555667, + "loss": 1.4701, + "theoretical_loss": 3.7900432897360803, + "tokens_seen": 733609984 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039278836509528585, + "loss": 1.4557, + "theoretical_loss": 3.790009456593639, + "tokens_seen": 733675520 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039277833500501503, + "loss": 1.4241, + "theoretical_loss": 3.7899756273193486, + "tokens_seen": 733741056 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039276830491474427, + "loss": 1.5142, + "theoretical_loss": 3.7899418019124207, + "tokens_seen": 733806592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003927582748244734, + "loss": 1.4893, + "theoretical_loss": 3.7899079803720688, + "tokens_seen": 733872128 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039274824473420263, + "loss": 1.5551, + "theoretical_loss": 3.7898741626975054, + "tokens_seen": 733937664 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 272255, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5827839374542236, + "objective/train/theoretical_loss": 3.7898403488879424, + "objective/train/tokens_used": 754463200, + "theoretical_loss": 3.7898403488879424, + "tokens_seen": 734003200 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039273821464393176, + "loss": 1.6175, + "theoretical_loss": 3.7898403488879424, + "tokens_seen": 734003200 + }, + { + "epoch": 0.22, + "learning_rate": 0.000392728184553661, + "loss": 1.553, + "theoretical_loss": 3.7898065389425946, + "tokens_seen": 734068736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003927181544633902, + "loss": 1.3181, + "theoretical_loss": 3.789772732860675, + "tokens_seen": 734134272 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039270812437311936, + "loss": 1.5134, + "theoretical_loss": 3.7897389306413967, + "tokens_seen": 734199808 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039269809428284854, + "loss": 1.5716, + "theoretical_loss": 3.7897051322839745, + "tokens_seen": 734265344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003926880641925777, + "loss": 1.4762, + "theoretical_loss": 3.7896713377876217, + "tokens_seen": 734330880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003926780341023069, + "loss": 1.5016, + "theoretical_loss": 3.7896375471515533, + "tokens_seen": 734396416 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039266800401203613, + "loss": 1.5009, + "theoretical_loss": 3.789603760374984, + "tokens_seen": 734461952 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039265797392176526, + "loss": 1.4847, + "theoretical_loss": 3.789569977457128, + "tokens_seen": 734527488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003926479438314945, + "loss": 1.5072, + "theoretical_loss": 3.789536198397201, + "tokens_seen": 734593024 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039263791374122373, + "loss": 1.6654, + "theoretical_loss": 3.789502423194418, + "tokens_seen": 734658560 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039262788365095286, + "loss": 1.5088, + "theoretical_loss": 3.7894686518479945, + "tokens_seen": 734724096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003926178535606821, + "loss": 1.4699, + "theoretical_loss": 3.789434884357147, + "tokens_seen": 734789632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003926078234704112, + "loss": 1.5756, + "theoretical_loss": 3.7894011207210907, + "tokens_seen": 734855168 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039259779338014046, + "loss": 1.4345, + "theoretical_loss": 3.7893673609390417, + "tokens_seen": 734920704 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039258776328986964, + "loss": 1.488, + "theoretical_loss": 3.789333605010217, + "tokens_seen": 734986240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003925777331995988, + "loss": 1.3993, + "theoretical_loss": 3.7892998529338326, + "tokens_seen": 735051776 + }, + { + "epoch": 0.22, + "learning_rate": 0.000392567703109328, + "loss": 1.5786, + "theoretical_loss": 3.7892661047091067, + "tokens_seen": 735117312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003925576730190572, + "loss": 1.4668, + "theoretical_loss": 3.789232360335255, + "tokens_seen": 735182848 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039254764292878636, + "loss": 1.5146, + "theoretical_loss": 3.789198619811496, + "tokens_seen": 735248384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003925376128385156, + "loss": 1.4396, + "theoretical_loss": 3.7891648831370466, + "tokens_seen": 735313920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003925275827482447, + "loss": 1.5806, + "theoretical_loss": 3.7891311503111247, + "tokens_seen": 735379456 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039251755265797396, + "loss": 1.3482, + "theoretical_loss": 3.789097421332949, + "tokens_seen": 735444992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003925075225677031, + "loss": 1.5207, + "theoretical_loss": 3.7890636962017368, + "tokens_seen": 735510528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924974924774323, + "loss": 1.5115, + "theoretical_loss": 3.7890299749167076, + "tokens_seen": 735576064 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924874623871615, + "loss": 1.4979, + "theoretical_loss": 3.7889962574770792, + "tokens_seen": 735641600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924774322968907, + "loss": 1.4638, + "theoretical_loss": 3.7889625438820715, + "tokens_seen": 735707136 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039246740220661987, + "loss": 1.4184, + "theoretical_loss": 3.788928834130903, + "tokens_seen": 735772672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924573721163491, + "loss": 1.4739, + "theoretical_loss": 3.7888951282227934, + "tokens_seen": 735838208 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039244734202607823, + "loss": 1.5246, + "theoretical_loss": 3.788861426156962, + "tokens_seen": 735903744 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039243731193580746, + "loss": 1.4674, + "theoretical_loss": 3.7888277279326292, + "tokens_seen": 735969280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924272818455366, + "loss": 1.612, + "theoretical_loss": 3.7887940335490153, + "tokens_seen": 736034816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003924172517552658, + "loss": 1.494, + "theoretical_loss": 3.7887603430053396, + "tokens_seen": 736100352 + }, + { + "epoch": 0.22, + "learning_rate": 0.000392407221664995, + "loss": 1.4731, + "theoretical_loss": 3.788726656300824, + "tokens_seen": 736165888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003923971915747242, + "loss": 1.4505, + "theoretical_loss": 3.7886929734346886, + "tokens_seen": 736231424 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039238716148445337, + "loss": 1.5629, + "theoretical_loss": 3.788659294406154, + "tokens_seen": 736296960 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039237713139418255, + "loss": 1.5394, + "theoretical_loss": 3.788625619214442, + "tokens_seen": 736362496 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039236710130391173, + "loss": 1.4248, + "theoretical_loss": 3.7885919478587744, + "tokens_seen": 736428032 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039235707121364097, + "loss": 1.4413, + "theoretical_loss": 3.788558280338372, + "tokens_seen": 736493568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003923470411233701, + "loss": 1.5109, + "theoretical_loss": 3.7885246166524578, + "tokens_seen": 736559104 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039233701103309933, + "loss": 1.387, + "theoretical_loss": 3.7884909568002536, + "tokens_seen": 736624640 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039232698094282846, + "loss": 1.4946, + "theoretical_loss": 3.7884573007809808, + "tokens_seen": 736690176 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003923169508525577, + "loss": 1.5377, + "theoretical_loss": 3.788423648593863, + "tokens_seen": 736755712 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039230692076228687, + "loss": 1.4395, + "theoretical_loss": 3.788390000238123, + "tokens_seen": 736821248 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039229689067201605, + "loss": 1.5208, + "theoretical_loss": 3.788356355712984, + "tokens_seen": 736886784 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039228686058174523, + "loss": 1.5285, + "theoretical_loss": 3.788322715017668, + "tokens_seen": 736952320 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039227683049147447, + "loss": 1.4932, + "theoretical_loss": 3.7882890781514007, + "tokens_seen": 737017856 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003922668004012036, + "loss": 1.5869, + "theoretical_loss": 3.7882554451134043, + "tokens_seen": 737083392 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039225677031093283, + "loss": 1.3959, + "theoretical_loss": 3.7882218159029026, + "tokens_seen": 737148928 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039224674022066196, + "loss": 1.5002, + "theoretical_loss": 3.7881881905191213, + "tokens_seen": 737214464 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 273341, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5457980632781982, + "objective/train/theoretical_loss": 3.788154568961283, + "objective/train/tokens_used": 757740000, + "theoretical_loss": 3.788154568961283, + "tokens_seen": 737280000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003922367101303912, + "loss": 1.5276, + "theoretical_loss": 3.788154568961283, + "tokens_seen": 737280000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003922266800401204, + "loss": 1.5394, + "theoretical_loss": 3.788120951228614, + "tokens_seen": 737345536 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039221664994984956, + "loss": 1.5174, + "theoretical_loss": 3.788087337320338, + "tokens_seen": 737411072 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039220661985957874, + "loss": 1.4216, + "theoretical_loss": 3.7880537272356802, + "tokens_seen": 737476608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921965897693079, + "loss": 1.5487, + "theoretical_loss": 3.788020120973867, + "tokens_seen": 737542144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921865596790371, + "loss": 1.5105, + "theoretical_loss": 3.7879865185341224, + "tokens_seen": 737607680 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039217652958876633, + "loss": 1.4632, + "theoretical_loss": 3.787952919915673, + "tokens_seen": 737673216 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039216649949849546, + "loss": 1.4817, + "theoretical_loss": 3.7879193251177457, + "tokens_seen": 737738752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921564694082247, + "loss": 1.5322, + "theoretical_loss": 3.7878857341395653, + "tokens_seen": 737804288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921464393179539, + "loss": 1.4663, + "theoretical_loss": 3.7878521469803585, + "tokens_seen": 737869824 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039213640922768306, + "loss": 1.5491, + "theoretical_loss": 3.7878185636393527, + "tokens_seen": 737935360 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039212637913741224, + "loss": 1.5147, + "theoretical_loss": 3.787784984115774, + "tokens_seen": 738000896 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921163490471414, + "loss": 1.4224, + "theoretical_loss": 3.78775140840885, + "tokens_seen": 738066432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003921063189568706, + "loss": 1.4927, + "theoretical_loss": 3.7877178365178086, + "tokens_seen": 738131968 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039209628886659984, + "loss": 1.4741, + "theoretical_loss": 3.787684268441876, + "tokens_seen": 738197504 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039208625877632896, + "loss": 1.4851, + "theoretical_loss": 3.787650704180281, + "tokens_seen": 738263040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003920762286860582, + "loss": 1.5811, + "theoretical_loss": 3.787617143732251, + "tokens_seen": 738328576 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003920661985957873, + "loss": 1.5051, + "theoretical_loss": 3.787583587097015, + "tokens_seen": 738394112 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039205616850551656, + "loss": 1.4073, + "theoretical_loss": 3.7875500342738015, + "tokens_seen": 738459648 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039204613841524574, + "loss": 1.468, + "theoretical_loss": 3.787516485261839, + "tokens_seen": 738525184 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003920361083249749, + "loss": 1.4812, + "theoretical_loss": 3.7874829400603556, + "tokens_seen": 738590720 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003920260782347041, + "loss": 1.37, + "theoretical_loss": 3.7874493986685813, + "tokens_seen": 738656256 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003920160481444333, + "loss": 1.5107, + "theoretical_loss": 3.787415861085745, + "tokens_seen": 738721792 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039200601805416247, + "loss": 1.3923, + "theoretical_loss": 3.787382327311077, + "tokens_seen": 738787328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003919959879638917, + "loss": 1.4777, + "theoretical_loss": 3.7873487973438067, + "tokens_seen": 738852864 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039198595787362083, + "loss": 1.5555, + "theoretical_loss": 3.7873152711831644, + "tokens_seen": 738918400 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039197592778335007, + "loss": 1.4585, + "theoretical_loss": 3.7872817488283803, + "tokens_seen": 738983936 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039196589769307925, + "loss": 1.4401, + "theoretical_loss": 3.7872482302786845, + "tokens_seen": 739049472 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039195586760280843, + "loss": 1.4621, + "theoretical_loss": 3.787214715533308, + "tokens_seen": 739115008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003919458375125376, + "loss": 1.5625, + "theoretical_loss": 3.787181204591482, + "tokens_seen": 739180544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003919358074222668, + "loss": 1.4001, + "theoretical_loss": 3.787147697452437, + "tokens_seen": 739246080 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039192577733199597, + "loss": 1.671, + "theoretical_loss": 3.7871141941154054, + "tokens_seen": 739311616 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003919157472417252, + "loss": 1.4796, + "theoretical_loss": 3.7870806945796174, + "tokens_seen": 739377152 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039190571715145433, + "loss": 1.3694, + "theoretical_loss": 3.7870471988443066, + "tokens_seen": 739442688 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039189568706118357, + "loss": 1.4162, + "theoretical_loss": 3.7870137069087035, + "tokens_seen": 739508224 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039188565697091275, + "loss": 1.5044, + "theoretical_loss": 3.786980218772041, + "tokens_seen": 739573760 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039187562688064193, + "loss": 1.6418, + "theoretical_loss": 3.7869467344335526, + "tokens_seen": 739639296 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039186559679037117, + "loss": 1.4574, + "theoretical_loss": 3.7869132538924695, + "tokens_seen": 739704832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003918555667001003, + "loss": 1.5167, + "theoretical_loss": 3.786879777148025, + "tokens_seen": 739770368 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039184553660982953, + "loss": 1.439, + "theoretical_loss": 3.786846304199453, + "tokens_seen": 739835904 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039183550651955866, + "loss": 1.5125, + "theoretical_loss": 3.7868128350459864, + "tokens_seen": 739901440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003918254764292879, + "loss": 1.4021, + "theoretical_loss": 3.7867793696868586, + "tokens_seen": 739966976 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039181544633901707, + "loss": 1.5228, + "theoretical_loss": 3.7867459081213037, + "tokens_seen": 740032512 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039180541624874625, + "loss": 1.5452, + "theoretical_loss": 3.7867124503485563, + "tokens_seen": 740098048 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039179538615847543, + "loss": 1.4338, + "theoretical_loss": 3.7866789963678498, + "tokens_seen": 740163584 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039178535606820467, + "loss": 1.5555, + "theoretical_loss": 3.786645546178419, + "tokens_seen": 740229120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003917753259779338, + "loss": 1.4856, + "theoretical_loss": 3.7866120997794983, + "tokens_seen": 740294656 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039176529588766303, + "loss": 1.6589, + "theoretical_loss": 3.786578657170324, + "tokens_seen": 740360192 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039175526579739216, + "loss": 1.4716, + "theoretical_loss": 3.7865452183501302, + "tokens_seen": 740425728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003917452357071214, + "loss": 1.5638, + "theoretical_loss": 3.7865117833181525, + "tokens_seen": 740491264 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 274540, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6254640817642212, + "objective/train/theoretical_loss": 3.786478352073626, + "objective/train/tokens_used": 761016800, + "theoretical_loss": 3.786478352073626, + "tokens_seen": 740556800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003917352056168506, + "loss": 1.5948, + "theoretical_loss": 3.786478352073626, + "tokens_seen": 740556800 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039172517552657976, + "loss": 1.4563, + "theoretical_loss": 3.786444924615788, + "tokens_seen": 740622336 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039171514543630894, + "loss": 1.5249, + "theoretical_loss": 3.7864115009438724, + "tokens_seen": 740687872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003917051153460381, + "loss": 1.4378, + "theoretical_loss": 3.7863780810571175, + "tokens_seen": 740753408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003916950852557673, + "loss": 1.4341, + "theoretical_loss": 3.7863446649547594, + "tokens_seen": 740818944 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039168505516549653, + "loss": 1.5184, + "theoretical_loss": 3.786311252636034, + "tokens_seen": 740884480 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039167502507522566, + "loss": 1.4167, + "theoretical_loss": 3.7862778441001783, + "tokens_seen": 740950016 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003916649949849549, + "loss": 1.3487, + "theoretical_loss": 3.7862444393464307, + "tokens_seen": 741015552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003916549648946841, + "loss": 1.6736, + "theoretical_loss": 3.7862110383740273, + "tokens_seen": 741081088 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039164493480441326, + "loss": 1.3781, + "theoretical_loss": 3.7861776411822063, + "tokens_seen": 741146624 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039163490471414244, + "loss": 1.3911, + "theoretical_loss": 3.7861442477702063, + "tokens_seen": 741212160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003916248746238716, + "loss": 1.5156, + "theoretical_loss": 3.7861108581372633, + "tokens_seen": 741277696 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003916148445336008, + "loss": 1.4423, + "theoretical_loss": 3.786077472282618, + "tokens_seen": 741343232 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039160481444333004, + "loss": 1.5196, + "theoretical_loss": 3.786044090205507, + "tokens_seen": 741408768 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039159478435305916, + "loss": 1.6, + "theoretical_loss": 3.7860107119051696, + "tokens_seen": 741474304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915847542627884, + "loss": 1.5107, + "theoretical_loss": 3.785977337380846, + "tokens_seen": 741539840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915747241725175, + "loss": 1.4997, + "theoretical_loss": 3.7859439666317733, + "tokens_seen": 741605376 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039156469408224676, + "loss": 1.5042, + "theoretical_loss": 3.7859105996571922, + "tokens_seen": 741670912 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039155466399197594, + "loss": 1.4984, + "theoretical_loss": 3.7858772364563418, + "tokens_seen": 741736448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915446339017051, + "loss": 1.4357, + "theoretical_loss": 3.785843877028462, + "tokens_seen": 741801984 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915346038114343, + "loss": 1.5566, + "theoretical_loss": 3.7858105213727935, + "tokens_seen": 741867520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915245737211635, + "loss": 1.4754, + "theoretical_loss": 3.7857771694885756, + "tokens_seen": 741933056 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039151454363089267, + "loss": 1.5126, + "theoretical_loss": 3.785743821375049, + "tokens_seen": 741998592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003915045135406219, + "loss": 1.4228, + "theoretical_loss": 3.785710477031455, + "tokens_seen": 742064128 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039149448345035103, + "loss": 1.5097, + "theoretical_loss": 3.785677136457034, + "tokens_seen": 742129664 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039148445336008027, + "loss": 1.4498, + "theoretical_loss": 3.7856437996510275, + "tokens_seen": 742195200 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039147442326980945, + "loss": 1.45, + "theoretical_loss": 3.7856104666126766, + "tokens_seen": 742260736 + }, + { + "epoch": 0.22, + "learning_rate": 0.00039146439317953863, + "loss": 1.5699, + "theoretical_loss": 3.785577137341223, + "tokens_seen": 742326272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0003914543630892678, + "loss": 1.4723, + "theoretical_loss": 3.7855438118359084, + "tokens_seen": 742391808 + }, + { + "epoch": 0.22, + "learning_rate": 0.000391444332998997, + "loss": 1.5019, + "theoretical_loss": 3.785510490095975, + "tokens_seen": 742457344 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039143430290872617, + "loss": 1.5589, + "theoretical_loss": 3.7854771721206646, + "tokens_seen": 742522880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003914242728184554, + "loss": 1.4779, + "theoretical_loss": 3.78544385790922, + "tokens_seen": 742588416 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039141424272818453, + "loss": 1.5209, + "theoretical_loss": 3.7854105474608843, + "tokens_seen": 742653952 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039140421263791377, + "loss": 1.593, + "theoretical_loss": 3.785377240774899, + "tokens_seen": 742719488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003913941825476429, + "loss": 1.2939, + "theoretical_loss": 3.7853439378505094, + "tokens_seen": 742785024 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039138415245737213, + "loss": 1.3371, + "theoretical_loss": 3.7853106386869575, + "tokens_seen": 742850560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003913741223671013, + "loss": 1.5452, + "theoretical_loss": 3.7852773432834868, + "tokens_seen": 742916096 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003913640922768305, + "loss": 1.5228, + "theoretical_loss": 3.7852440516393413, + "tokens_seen": 742981632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003913540621865597, + "loss": 1.5265, + "theoretical_loss": 3.785210763753765, + "tokens_seen": 743047168 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039134403209628886, + "loss": 1.5178, + "theoretical_loss": 3.785177479626002, + "tokens_seen": 743112704 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039133400200601804, + "loss": 1.513, + "theoretical_loss": 3.785144199255297, + "tokens_seen": 743178240 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039132397191574727, + "loss": 1.3671, + "theoretical_loss": 3.7851109226408943, + "tokens_seen": 743243776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003913139418254764, + "loss": 1.5212, + "theoretical_loss": 3.785077649782039, + "tokens_seen": 743309312 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039130391173520563, + "loss": 1.4988, + "theoretical_loss": 3.7850443806779763, + "tokens_seen": 743374848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003912938816449348, + "loss": 1.5945, + "theoretical_loss": 3.7850111153279515, + "tokens_seen": 743440384 + }, + { + "epoch": 0.23, + "learning_rate": 0.000391283851554664, + "loss": 1.4628, + "theoretical_loss": 3.78497785373121, + "tokens_seen": 743505920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003912738214643932, + "loss": 1.4697, + "theoretical_loss": 3.784944595886998, + "tokens_seen": 743571456 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039126379137412236, + "loss": 1.3576, + "theoretical_loss": 3.7849113417945603, + "tokens_seen": 743636992 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039125376128385154, + "loss": 1.5531, + "theoretical_loss": 3.7848780914531437, + "tokens_seen": 743702528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003912437311935808, + "loss": 1.6566, + "theoretical_loss": 3.784844844861995, + "tokens_seen": 743768064 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 275779, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6011989116668701, + "objective/train/theoretical_loss": 3.784811602020361, + "objective/train/tokens_used": 764293600, + "theoretical_loss": 3.784811602020361, + "tokens_seen": 743833600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003912337011033099, + "loss": 1.5224, + "theoretical_loss": 3.784811602020361, + "tokens_seen": 743833600 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039122367101303914, + "loss": 1.4771, + "theoretical_loss": 3.7847783629274874, + "tokens_seen": 743899136 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039121364092276826, + "loss": 1.573, + "theoretical_loss": 3.7847451275826227, + "tokens_seen": 743964672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003912036108324975, + "loss": 1.54, + "theoretical_loss": 3.784711895985013, + "tokens_seen": 744030208 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003911935807422267, + "loss": 1.5818, + "theoretical_loss": 3.784678668133906, + "tokens_seen": 744095744 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039118355065195586, + "loss": 1.4522, + "theoretical_loss": 3.7846454440285493, + "tokens_seen": 744161280 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039117352056168504, + "loss": 1.4705, + "theoretical_loss": 3.7846122236681916, + "tokens_seen": 744226816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003911634904714143, + "loss": 1.4444, + "theoretical_loss": 3.7845790070520806, + "tokens_seen": 744292352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003911534603811434, + "loss": 1.4271, + "theoretical_loss": 3.784545794179464, + "tokens_seen": 744357888 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039114343029087264, + "loss": 1.4467, + "theoretical_loss": 3.7845125850495918, + "tokens_seen": 744423424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003911334002006018, + "loss": 1.5077, + "theoretical_loss": 3.7844793796617116, + "tokens_seen": 744488960 + }, + { + "epoch": 0.23, + "learning_rate": 0.000391123370110331, + "loss": 1.4382, + "theoretical_loss": 3.784446178015073, + "tokens_seen": 744554496 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039111334002006024, + "loss": 1.4893, + "theoretical_loss": 3.7844129801089244, + "tokens_seen": 744620032 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039110330992978936, + "loss": 1.5079, + "theoretical_loss": 3.784379785942517, + "tokens_seen": 744685568 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003910932798395186, + "loss": 1.486, + "theoretical_loss": 3.784346595515098, + "tokens_seen": 744751104 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039108324974924773, + "loss": 1.5126, + "theoretical_loss": 3.784313408825919, + "tokens_seen": 744816640 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039107321965897696, + "loss": 1.4779, + "theoretical_loss": 3.78428022587423, + "tokens_seen": 744882176 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039106318956870614, + "loss": 1.392, + "theoretical_loss": 3.7842470466592806, + "tokens_seen": 744947712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003910531594784353, + "loss": 1.4633, + "theoretical_loss": 3.784213871180322, + "tokens_seen": 745013248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003910431293881645, + "loss": 1.4994, + "theoretical_loss": 3.7841806994366047, + "tokens_seen": 745078784 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003910330992978937, + "loss": 1.4495, + "theoretical_loss": 3.7841475314273794, + "tokens_seen": 745144320 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039102306920762287, + "loss": 1.3199, + "theoretical_loss": 3.7841143671518975, + "tokens_seen": 745209856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003910130391173521, + "loss": 1.6759, + "theoretical_loss": 3.7840812066094105, + "tokens_seen": 745275392 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039100300902708123, + "loss": 1.4959, + "theoretical_loss": 3.78404804979917, + "tokens_seen": 745340928 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039099297893681047, + "loss": 1.4846, + "theoretical_loss": 3.7840148967204277, + "tokens_seen": 745406464 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039098294884653965, + "loss": 1.457, + "theoretical_loss": 3.7839817473724353, + "tokens_seen": 745472000 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039097291875626883, + "loss": 1.5371, + "theoretical_loss": 3.783948601754445, + "tokens_seen": 745537536 + }, + { + "epoch": 0.23, + "learning_rate": 0.000390962888665998, + "loss": 1.5451, + "theoretical_loss": 3.7839154598657103, + "tokens_seen": 745603072 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003909528585757272, + "loss": 1.4438, + "theoretical_loss": 3.7838823217054833, + "tokens_seen": 745668608 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039094282848545637, + "loss": 1.5309, + "theoretical_loss": 3.7838491872730167, + "tokens_seen": 745734144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003909327983951856, + "loss": 1.5105, + "theoretical_loss": 3.783816056567564, + "tokens_seen": 745799680 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039092276830491473, + "loss": 1.4935, + "theoretical_loss": 3.783782929588378, + "tokens_seen": 745865216 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039091273821464397, + "loss": 1.4456, + "theoretical_loss": 3.783749806334712, + "tokens_seen": 745930752 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003909027081243731, + "loss": 1.5415, + "theoretical_loss": 3.783716686805821, + "tokens_seen": 745996288 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039089267803410233, + "loss": 1.4535, + "theoretical_loss": 3.7836835710009584, + "tokens_seen": 746061824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003908826479438315, + "loss": 1.5246, + "theoretical_loss": 3.783650458919378, + "tokens_seen": 746127360 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003908726178535607, + "loss": 1.4244, + "theoretical_loss": 3.783617350560334, + "tokens_seen": 746192896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003908625877632899, + "loss": 1.4727, + "theoretical_loss": 3.7835842459230817, + "tokens_seen": 746258432 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039085255767301906, + "loss": 1.4007, + "theoretical_loss": 3.783551145006876, + "tokens_seen": 746323968 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039084252758274824, + "loss": 1.4527, + "theoretical_loss": 3.783518047810971, + "tokens_seen": 746389504 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039083249749247747, + "loss": 1.5609, + "theoretical_loss": 3.7834849543346234, + "tokens_seen": 746455040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003908224674022066, + "loss": 1.4893, + "theoretical_loss": 3.7834518645770876, + "tokens_seen": 746520576 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039081243731193583, + "loss": 1.4971, + "theoretical_loss": 3.7834187785376194, + "tokens_seen": 746586112 + }, + { + "epoch": 0.23, + "learning_rate": 0.000390802407221665, + "loss": 1.5173, + "theoretical_loss": 3.7833856962154755, + "tokens_seen": 746651648 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003907923771313942, + "loss": 1.5076, + "theoretical_loss": 3.7833526176099106, + "tokens_seen": 746717184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003907823470411234, + "loss": 1.3818, + "theoretical_loss": 3.7833195427201822, + "tokens_seen": 746782720 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039077231695085256, + "loss": 1.4696, + "theoretical_loss": 3.783286471545547, + "tokens_seen": 746848256 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039076228686058174, + "loss": 1.5379, + "theoretical_loss": 3.7832534040852606, + "tokens_seen": 746913792 + }, + { + "epoch": 0.23, + "learning_rate": 0.000390752256770311, + "loss": 1.4476, + "theoretical_loss": 3.7832203403385805, + "tokens_seen": 746979328 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003907422266800401, + "loss": 1.5183, + "theoretical_loss": 3.783187280304764, + "tokens_seen": 747044864 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 277055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4990746974945068, + "objective/train/theoretical_loss": 3.783154223983069, + "objective/train/tokens_used": 767570400, + "theoretical_loss": 3.783154223983069, + "tokens_seen": 747110400 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039073219658976934, + "loss": 1.576, + "theoretical_loss": 3.783154223983069, + "tokens_seen": 747110400 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039072216649949846, + "loss": 1.4154, + "theoretical_loss": 3.783121171372753, + "tokens_seen": 747175936 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003907121364092277, + "loss": 1.45, + "theoretical_loss": 3.783088122473073, + "tokens_seen": 747241472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003907021063189569, + "loss": 1.5443, + "theoretical_loss": 3.7830550772832874, + "tokens_seen": 747307008 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039069207622868606, + "loss": 1.5403, + "theoretical_loss": 3.7830220358026545, + "tokens_seen": 747372544 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039068204613841524, + "loss": 1.4256, + "theoretical_loss": 3.782988998030433, + "tokens_seen": 747438080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003906720160481445, + "loss": 1.4867, + "theoretical_loss": 3.7829559639658816, + "tokens_seen": 747503616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003906619859578736, + "loss": 1.4177, + "theoretical_loss": 3.7829229336082584, + "tokens_seen": 747569152 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039065195586760284, + "loss": 1.4502, + "theoretical_loss": 3.7828899069568243, + "tokens_seen": 747634688 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039064192577733197, + "loss": 1.4873, + "theoretical_loss": 3.7828568840108363, + "tokens_seen": 747700224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003906318956870612, + "loss": 1.4276, + "theoretical_loss": 3.782823864769556, + "tokens_seen": 747765760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003906218655967904, + "loss": 1.4245, + "theoretical_loss": 3.7827908492322413, + "tokens_seen": 747831296 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039061183550651957, + "loss": 1.473, + "theoretical_loss": 3.7827578373981536, + "tokens_seen": 747896832 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039060180541624875, + "loss": 1.5159, + "theoretical_loss": 3.782724829266553, + "tokens_seen": 747962368 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039059177532597793, + "loss": 1.3778, + "theoretical_loss": 3.782691824836699, + "tokens_seen": 748027904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003905817452357071, + "loss": 1.5711, + "theoretical_loss": 3.7826588241078527, + "tokens_seen": 748093440 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039057171514543634, + "loss": 1.4195, + "theoretical_loss": 3.7826258270792747, + "tokens_seen": 748158976 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039056168505516547, + "loss": 1.468, + "theoretical_loss": 3.7825928337502264, + "tokens_seen": 748224512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003905516549648947, + "loss": 1.6578, + "theoretical_loss": 3.7825598441199695, + "tokens_seen": 748290048 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039054162487462383, + "loss": 1.4807, + "theoretical_loss": 3.7825268581877642, + "tokens_seen": 748355584 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039053159478435307, + "loss": 1.5042, + "theoretical_loss": 3.7824938759528726, + "tokens_seen": 748421120 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039052156469408225, + "loss": 1.508, + "theoretical_loss": 3.7824608974145573, + "tokens_seen": 748486656 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039051153460381143, + "loss": 1.4777, + "theoretical_loss": 3.782427922572079, + "tokens_seen": 748552192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003905015045135406, + "loss": 1.506, + "theoretical_loss": 3.7823949514247017, + "tokens_seen": 748617728 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039049147442326985, + "loss": 1.4562, + "theoretical_loss": 3.782361983971686, + "tokens_seen": 748683264 + }, + { + "epoch": 0.23, + "learning_rate": 0.000390481444332999, + "loss": 1.5416, + "theoretical_loss": 3.7823290202122966, + "tokens_seen": 748748800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003904714142427282, + "loss": 1.3645, + "theoretical_loss": 3.7822960601457947, + "tokens_seen": 748814336 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039046138415245734, + "loss": 1.4267, + "theoretical_loss": 3.782263103771445, + "tokens_seen": 748879872 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039045135406218657, + "loss": 1.5942, + "theoretical_loss": 3.78223015108851, + "tokens_seen": 748945408 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039044132397191575, + "loss": 1.4641, + "theoretical_loss": 3.7821972020962527, + "tokens_seen": 749010944 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039043129388164493, + "loss": 1.5104, + "theoretical_loss": 3.782164256793938, + "tokens_seen": 749076480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003904212637913741, + "loss": 1.4795, + "theoretical_loss": 3.782131315180829, + "tokens_seen": 749142016 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003904112337011033, + "loss": 1.5131, + "theoretical_loss": 3.782098377256191, + "tokens_seen": 749207552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003904012036108325, + "loss": 1.4008, + "theoretical_loss": 3.7820654430192873, + "tokens_seen": 749273088 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903911735205617, + "loss": 1.4519, + "theoretical_loss": 3.782032512469383, + "tokens_seen": 749338624 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903811434302909, + "loss": 1.5325, + "theoretical_loss": 3.781999585605743, + "tokens_seen": 749404160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903711133400201, + "loss": 1.5802, + "theoretical_loss": 3.781966662427632, + "tokens_seen": 749469696 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039036108324974926, + "loss": 1.5327, + "theoretical_loss": 3.7819337429343154, + "tokens_seen": 749535232 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039035105315947844, + "loss": 1.4857, + "theoretical_loss": 3.7819008271250594, + "tokens_seen": 749600768 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039034102306920767, + "loss": 1.567, + "theoretical_loss": 3.7818679149991286, + "tokens_seen": 749666304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903309929789368, + "loss": 1.5539, + "theoretical_loss": 3.7818350065557893, + "tokens_seen": 749731840 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039032096288866603, + "loss": 1.5954, + "theoretical_loss": 3.7818021017943075, + "tokens_seen": 749797376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903109327983952, + "loss": 1.4185, + "theoretical_loss": 3.78176920071395, + "tokens_seen": 749862912 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003903009027081244, + "loss": 1.4567, + "theoretical_loss": 3.7817363033139824, + "tokens_seen": 749928448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003902908726178536, + "loss": 1.4798, + "theoretical_loss": 3.781703409593672, + "tokens_seen": 749993984 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039028084252758276, + "loss": 1.4973, + "theoretical_loss": 3.7816705195522866, + "tokens_seen": 750059520 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039027081243731194, + "loss": 1.5044, + "theoretical_loss": 3.7816376331890913, + "tokens_seen": 750125056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003902607823470412, + "loss": 1.569, + "theoretical_loss": 3.7816047505033556, + "tokens_seen": 750190592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003902507522567703, + "loss": 1.5542, + "theoretical_loss": 3.7815718714943456, + "tokens_seen": 750256128 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039024072216649954, + "loss": 1.4322, + "theoretical_loss": 3.7815389961613297, + "tokens_seen": 750321664 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 278273, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4322211742401123, + "objective/train/theoretical_loss": 3.7815061245035757, + "objective/train/tokens_used": 770847200, + "theoretical_loss": 3.7815061245035757, + "tokens_seen": 750387200 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039023069207622866, + "loss": 1.4645, + "theoretical_loss": 3.7815061245035757, + "tokens_seen": 750387200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003902206619859579, + "loss": 1.481, + "theoretical_loss": 3.7814732565203517, + "tokens_seen": 750452736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003902106318956871, + "loss": 1.5221, + "theoretical_loss": 3.7814403922109263, + "tokens_seen": 750518272 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039020060180541626, + "loss": 1.3395, + "theoretical_loss": 3.781407531574568, + "tokens_seen": 750583808 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039019057171514544, + "loss": 1.451, + "theoretical_loss": 3.7813746746105465, + "tokens_seen": 750649344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003901805416248747, + "loss": 1.4203, + "theoretical_loss": 3.7813418213181293, + "tokens_seen": 750714880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003901705115346038, + "loss": 1.3748, + "theoretical_loss": 3.7813089716965864, + "tokens_seen": 750780416 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039016048144433304, + "loss": 1.4152, + "theoretical_loss": 3.7812761257451872, + "tokens_seen": 750845952 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039015045135406217, + "loss": 1.5226, + "theoretical_loss": 3.7812432834632017, + "tokens_seen": 750911488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003901404212637914, + "loss": 1.445, + "theoretical_loss": 3.7812104448498998, + "tokens_seen": 750977024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003901303911735206, + "loss": 1.3966, + "theoretical_loss": 3.7811776099045504, + "tokens_seen": 751042560 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039012036108324977, + "loss": 1.5279, + "theoretical_loss": 3.781144778626425, + "tokens_seen": 751108096 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039011033099297895, + "loss": 1.5055, + "theoretical_loss": 3.781111951014794, + "tokens_seen": 751173632 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039010030090270813, + "loss": 1.5411, + "theoretical_loss": 3.781079127068928, + "tokens_seen": 751239168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003900902708124373, + "loss": 1.5109, + "theoretical_loss": 3.781046306788097, + "tokens_seen": 751304704 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039008024072216654, + "loss": 1.5594, + "theoretical_loss": 3.7810134901715733, + "tokens_seen": 751370240 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039007021063189567, + "loss": 1.5366, + "theoretical_loss": 3.7809806772186283, + "tokens_seen": 751435776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003900601805416249, + "loss": 1.5269, + "theoretical_loss": 3.780947867928532, + "tokens_seen": 751501312 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039005015045135403, + "loss": 1.4867, + "theoretical_loss": 3.7809150623005583, + "tokens_seen": 751566848 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039004012036108327, + "loss": 1.5226, + "theoretical_loss": 3.7808822603339776, + "tokens_seen": 751632384 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039003009027081245, + "loss": 1.5327, + "theoretical_loss": 3.780849462028063, + "tokens_seen": 751697920 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039002006018054163, + "loss": 1.4591, + "theoretical_loss": 3.780816667382086, + "tokens_seen": 751763456 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003900100300902708, + "loss": 1.523, + "theoretical_loss": 3.7807838763953194, + "tokens_seen": 751828992 + }, + { + "epoch": 0.23, + "learning_rate": 0.00039000000000000005, + "loss": 1.4022, + "theoretical_loss": 3.780751089067037, + "tokens_seen": 751894528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003899899699097292, + "loss": 1.4933, + "theoretical_loss": 3.78071830539651, + "tokens_seen": 751960064 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003899799398194584, + "loss": 1.5199, + "theoretical_loss": 3.7806855253830136, + "tokens_seen": 752025600 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038996990972918754, + "loss": 1.5395, + "theoretical_loss": 3.78065274902582, + "tokens_seen": 752091136 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038995987963891677, + "loss": 1.4901, + "theoretical_loss": 3.7806199763242025, + "tokens_seen": 752156672 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038994984954864595, + "loss": 1.465, + "theoretical_loss": 3.780587207277436, + "tokens_seen": 752222208 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038993981945837513, + "loss": 1.4743, + "theoretical_loss": 3.7805544418847936, + "tokens_seen": 752287744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003899297893681043, + "loss": 1.5349, + "theoretical_loss": 3.7805216801455503, + "tokens_seen": 752353280 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003899197592778335, + "loss": 1.385, + "theoretical_loss": 3.7804889220589804, + "tokens_seen": 752418816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003899097291875627, + "loss": 1.6085, + "theoretical_loss": 3.780456167624358, + "tokens_seen": 752484352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898996990972919, + "loss": 1.5005, + "theoretical_loss": 3.780423416840958, + "tokens_seen": 752549888 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038988966900702104, + "loss": 1.6492, + "theoretical_loss": 3.7803906697080567, + "tokens_seen": 752615424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898796389167503, + "loss": 1.4518, + "theoretical_loss": 3.780357926224928, + "tokens_seen": 752680960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898696088264794, + "loss": 1.5315, + "theoretical_loss": 3.780325186390848, + "tokens_seen": 752746496 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038985957873620864, + "loss": 1.6136, + "theoretical_loss": 3.780292450205092, + "tokens_seen": 752812032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898495486459378, + "loss": 1.482, + "theoretical_loss": 3.7802597176669366, + "tokens_seen": 752877568 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389839518555667, + "loss": 1.3913, + "theoretical_loss": 3.7802269887756577, + "tokens_seen": 752943104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898294884653962, + "loss": 1.4003, + "theoretical_loss": 3.7801942635305306, + "tokens_seen": 753008640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003898194583751254, + "loss": 1.5405, + "theoretical_loss": 3.780161541930833, + "tokens_seen": 753074176 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038980942828485454, + "loss": 1.4722, + "theoretical_loss": 3.7801288239758417, + "tokens_seen": 753139712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897993981945838, + "loss": 1.5282, + "theoretical_loss": 3.7800961096648327, + "tokens_seen": 753205248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897893681043129, + "loss": 1.4869, + "theoretical_loss": 3.7800633989970835, + "tokens_seen": 753270784 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038977933801404214, + "loss": 1.4425, + "theoretical_loss": 3.780030691971872, + "tokens_seen": 753336320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897693079237713, + "loss": 1.5412, + "theoretical_loss": 3.7799979885884754, + "tokens_seen": 753401856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897592778335005, + "loss": 1.5697, + "theoretical_loss": 3.7799652888461712, + "tokens_seen": 753467392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897492477432297, + "loss": 1.5531, + "theoretical_loss": 3.7799325927442373, + "tokens_seen": 753532928 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038973921765295886, + "loss": 1.4926, + "theoretical_loss": 3.7798999002819524, + "tokens_seen": 753598464 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 279492, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5427801609039307, + "objective/train/theoretical_loss": 3.779867211458595, + "objective/train/tokens_used": 774124000, + "theoretical_loss": 3.779867211458595, + "tokens_seen": 753664000 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038972918756268805, + "loss": 1.4864, + "theoretical_loss": 3.779867211458595, + "tokens_seen": 753664000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897191574724173, + "loss": 1.502, + "theoretical_loss": 3.7798345262734427, + "tokens_seen": 753729536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003897091273821464, + "loss": 1.5002, + "theoretical_loss": 3.779801844725775, + "tokens_seen": 753795072 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038969909729187564, + "loss": 1.4768, + "theoretical_loss": 3.779769166814871, + "tokens_seen": 753860608 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038968906720160477, + "loss": 1.5, + "theoretical_loss": 3.7797364925400094, + "tokens_seen": 753926144 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389679037111334, + "loss": 1.507, + "theoretical_loss": 3.77970382190047, + "tokens_seen": 753991680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003896690070210632, + "loss": 1.4842, + "theoretical_loss": 3.7796711548955324, + "tokens_seen": 754057216 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038965897693079237, + "loss": 1.5324, + "theoretical_loss": 3.7796384915244765, + "tokens_seen": 754122752 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038964894684052155, + "loss": 1.459, + "theoretical_loss": 3.7796058317865815, + "tokens_seen": 754188288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003896389167502508, + "loss": 1.4859, + "theoretical_loss": 3.7795731756811284, + "tokens_seen": 754253824 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038962888665997997, + "loss": 1.5902, + "theoretical_loss": 3.779540523207398, + "tokens_seen": 754319360 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038961885656970915, + "loss": 1.5356, + "theoretical_loss": 3.77950787436467, + "tokens_seen": 754384896 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038960882647943833, + "loss": 1.4305, + "theoretical_loss": 3.779475229152226, + "tokens_seen": 754450432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003895987963891675, + "loss": 1.5528, + "theoretical_loss": 3.7794425875693465, + "tokens_seen": 754515968 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038958876629889674, + "loss": 1.4959, + "theoretical_loss": 3.779409949615313, + "tokens_seen": 754581504 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038957873620862587, + "loss": 1.5135, + "theoretical_loss": 3.7793773152894072, + "tokens_seen": 754647040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003895687061183551, + "loss": 1.5369, + "theoretical_loss": 3.77934468459091, + "tokens_seen": 754712576 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038955867602808423, + "loss": 1.43, + "theoretical_loss": 3.779312057519104, + "tokens_seen": 754778112 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038954864593781347, + "loss": 1.5037, + "theoretical_loss": 3.7792794340732714, + "tokens_seen": 754843648 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038953861584754265, + "loss": 1.5461, + "theoretical_loss": 3.7792468142526934, + "tokens_seen": 754909184 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038952858575727183, + "loss": 1.4348, + "theoretical_loss": 3.7792141980566534, + "tokens_seen": 754974720 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389518555667001, + "loss": 1.4225, + "theoretical_loss": 3.7791815854844337, + "tokens_seen": 755040256 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038950852557673025, + "loss": 1.5144, + "theoretical_loss": 3.7791489765353177, + "tokens_seen": 755105792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894984954864594, + "loss": 1.4945, + "theoretical_loss": 3.7791163712085885, + "tokens_seen": 755171328 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894884653961886, + "loss": 1.4292, + "theoretical_loss": 3.7790837695035284, + "tokens_seen": 755236864 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038947843530591774, + "loss": 1.449, + "theoretical_loss": 3.7790511714194217, + "tokens_seen": 755302400 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038946840521564697, + "loss": 1.5527, + "theoretical_loss": 3.779018576955552, + "tokens_seen": 755367936 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038945837512537615, + "loss": 1.4729, + "theoretical_loss": 3.778985986111203, + "tokens_seen": 755433472 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038944834503510533, + "loss": 1.6047, + "theoretical_loss": 3.778953398885659, + "tokens_seen": 755499008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894383149448345, + "loss": 1.5422, + "theoretical_loss": 3.7789208152782043, + "tokens_seen": 755564544 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894282848545637, + "loss": 1.5034, + "theoretical_loss": 3.7788882352881235, + "tokens_seen": 755630080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894182547642929, + "loss": 1.4318, + "theoretical_loss": 3.7788556589147015, + "tokens_seen": 755695616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003894082246740221, + "loss": 1.5298, + "theoretical_loss": 3.7788230861572227, + "tokens_seen": 755761152 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038939819458375124, + "loss": 1.4481, + "theoretical_loss": 3.778790517014972, + "tokens_seen": 755826688 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003893881644934805, + "loss": 1.5009, + "theoretical_loss": 3.7787579514872354, + "tokens_seen": 755892224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003893781344032096, + "loss": 1.4319, + "theoretical_loss": 3.7787253895732986, + "tokens_seen": 755957760 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038936810431293884, + "loss": 1.4449, + "theoretical_loss": 3.778692831272447, + "tokens_seen": 756023296 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389358074222668, + "loss": 1.5266, + "theoretical_loss": 3.778660276583966, + "tokens_seen": 756088832 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003893480441323972, + "loss": 1.6073, + "theoretical_loss": 3.7786277255071425, + "tokens_seen": 756154368 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003893380140421264, + "loss": 1.5304, + "theoretical_loss": 3.7785951780412628, + "tokens_seen": 756219904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003893279839518556, + "loss": 1.4666, + "theoretical_loss": 3.7785626341856133, + "tokens_seen": 756285440 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038931795386158474, + "loss": 1.3954, + "theoretical_loss": 3.77853009393948, + "tokens_seen": 756350976 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389307923771314, + "loss": 1.5698, + "theoretical_loss": 3.7784975573021518, + "tokens_seen": 756416512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892978936810431, + "loss": 1.3972, + "theoretical_loss": 3.7784650242729136, + "tokens_seen": 756482048 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038928786359077234, + "loss": 1.5049, + "theoretical_loss": 3.7784324948510544, + "tokens_seen": 756547584 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892778335005015, + "loss": 1.4602, + "theoretical_loss": 3.7783999690358607, + "tokens_seen": 756613120 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892678034102307, + "loss": 1.4271, + "theoretical_loss": 3.778367446826621, + "tokens_seen": 756678656 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892577733199599, + "loss": 1.5172, + "theoretical_loss": 3.778334928222623, + "tokens_seen": 756744192 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038924774322968906, + "loss": 1.395, + "theoretical_loss": 3.7783024132231544, + "tokens_seen": 756809728 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038923771313941825, + "loss": 1.5384, + "theoretical_loss": 3.7782699018275046, + "tokens_seen": 756875264 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 280659, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.603344202041626, + "objective/train/theoretical_loss": 3.778237394034961, + "objective/train/tokens_used": 777400800, + "theoretical_loss": 3.778237394034961, + "tokens_seen": 756940800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892276830491475, + "loss": 1.5267, + "theoretical_loss": 3.778237394034961, + "tokens_seen": 756940800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003892176529588766, + "loss": 1.4786, + "theoretical_loss": 3.7782048898448135, + "tokens_seen": 757006336 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038920762286860584, + "loss": 1.555, + "theoretical_loss": 3.7781723892563503, + "tokens_seen": 757071872 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038919759277833497, + "loss": 1.4155, + "theoretical_loss": 3.778139892268861, + "tokens_seen": 757137408 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003891875626880642, + "loss": 1.5732, + "theoretical_loss": 3.7781073988816347, + "tokens_seen": 757202944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003891775325977934, + "loss": 1.4555, + "theoretical_loss": 3.7780749090939616, + "tokens_seen": 757268480 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038916750250752257, + "loss": 1.5542, + "theoretical_loss": 3.7780424229051306, + "tokens_seen": 757334016 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038915747241725175, + "loss": 1.4762, + "theoretical_loss": 3.778009940314432, + "tokens_seen": 757399552 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389147442326981, + "loss": 1.4648, + "theoretical_loss": 3.7779774613211563, + "tokens_seen": 757465088 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003891374122367101, + "loss": 1.5214, + "theoretical_loss": 3.7779449859245933, + "tokens_seen": 757530624 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038912738214643935, + "loss": 1.4832, + "theoretical_loss": 3.777912514124034, + "tokens_seen": 757596160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003891173520561685, + "loss": 1.5012, + "theoretical_loss": 3.7778800459187694, + "tokens_seen": 757661696 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003891073219658977, + "loss": 1.5068, + "theoretical_loss": 3.77784758130809, + "tokens_seen": 757727232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003890972918756269, + "loss": 1.4672, + "theoretical_loss": 3.7778151202912875, + "tokens_seen": 757792768 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038908726178535607, + "loss": 1.5263, + "theoretical_loss": 3.777782662867653, + "tokens_seen": 757858304 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038907723169508525, + "loss": 1.509, + "theoretical_loss": 3.777750209036478, + "tokens_seen": 757923840 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038906720160481443, + "loss": 1.5183, + "theoretical_loss": 3.7777177587970545, + "tokens_seen": 757989376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003890571715145436, + "loss": 1.5467, + "theoretical_loss": 3.777685312148674, + "tokens_seen": 758054912 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038904714142427285, + "loss": 1.5268, + "theoretical_loss": 3.77765286909063, + "tokens_seen": 758120448 + }, + { + "epoch": 0.23, + "learning_rate": 0.000389037111334002, + "loss": 1.3561, + "theoretical_loss": 3.777620429622213, + "tokens_seen": 758185984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003890270812437312, + "loss": 1.4766, + "theoretical_loss": 3.7775879937427175, + "tokens_seen": 758251520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003890170511534604, + "loss": 1.4893, + "theoretical_loss": 3.777555561451435, + "tokens_seen": 758317056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003890070210631896, + "loss": 1.5867, + "theoretical_loss": 3.777523132747659, + "tokens_seen": 758382592 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038899699097291876, + "loss": 1.5302, + "theoretical_loss": 3.7774907076306823, + "tokens_seen": 758448128 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038898696088264794, + "loss": 1.4563, + "theoretical_loss": 3.777458286099799, + "tokens_seen": 758513664 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003889769307923771, + "loss": 1.537, + "theoretical_loss": 3.7774258681543027, + "tokens_seen": 758579200 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038896690070210635, + "loss": 1.6247, + "theoretical_loss": 3.7773934537934863, + "tokens_seen": 758644736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003889568706118355, + "loss": 1.5978, + "theoretical_loss": 3.7773610430166444, + "tokens_seen": 758710272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003889468405215647, + "loss": 1.6618, + "theoretical_loss": 3.7773286358230713, + "tokens_seen": 758775808 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038893681043129384, + "loss": 1.4983, + "theoretical_loss": 3.777296232212061, + "tokens_seen": 758841344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003889267803410231, + "loss": 1.5022, + "theoretical_loss": 3.777263832182909, + "tokens_seen": 758906880 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038891675025075226, + "loss": 1.4111, + "theoretical_loss": 3.7772314357349086, + "tokens_seen": 758972416 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038890672016048144, + "loss": 1.5428, + "theoretical_loss": 3.7771990428673563, + "tokens_seen": 759037952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888966900702106, + "loss": 1.5401, + "theoretical_loss": 3.7771666535795463, + "tokens_seen": 759103488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888866599799398, + "loss": 1.5907, + "theoretical_loss": 3.777134267870774, + "tokens_seen": 759169024 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038887662988966904, + "loss": 1.4017, + "theoretical_loss": 3.7771018857403362, + "tokens_seen": 759234560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888665997993982, + "loss": 1.5797, + "theoretical_loss": 3.7770695071875275, + "tokens_seen": 759300096 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888565697091274, + "loss": 1.5302, + "theoretical_loss": 3.7770371322116443, + "tokens_seen": 759365632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888465396188566, + "loss": 1.5441, + "theoretical_loss": 3.7770047608119826, + "tokens_seen": 759431168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888365095285858, + "loss": 1.5775, + "theoretical_loss": 3.7769723929878394, + "tokens_seen": 759496704 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038882647943831494, + "loss": 1.4448, + "theoretical_loss": 3.77694002873851, + "tokens_seen": 759562240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888164493480442, + "loss": 1.466, + "theoretical_loss": 3.7769076680632927, + "tokens_seen": 759627776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003888064192577733, + "loss": 1.4933, + "theoretical_loss": 3.7768753109614837, + "tokens_seen": 759693312 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038879638916750254, + "loss": 1.4151, + "theoretical_loss": 3.7768429574323803, + "tokens_seen": 759758848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003887863590772317, + "loss": 1.3862, + "theoretical_loss": 3.77681060747528, + "tokens_seen": 759824384 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003887763289869609, + "loss": 1.489, + "theoretical_loss": 3.7767782610894804, + "tokens_seen": 759889920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003887662988966901, + "loss": 1.5317, + "theoretical_loss": 3.776745918274279, + "tokens_seen": 759955456 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038875626880641927, + "loss": 1.5319, + "theoretical_loss": 3.776713579028974, + "tokens_seen": 760020992 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038874623871614845, + "loss": 1.5359, + "theoretical_loss": 3.776681243352864, + "tokens_seen": 760086528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003887362086258777, + "loss": 1.6488, + "theoretical_loss": 3.7766489112452466, + "tokens_seen": 760152064 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 281753, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5441899299621582, + "objective/train/theoretical_loss": 3.776616582705421, + "objective/train/tokens_used": 780677600, + "theoretical_loss": 3.776616582705421, + "tokens_seen": 760217600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003887261785356068, + "loss": 1.4906, + "theoretical_loss": 3.776616582705421, + "tokens_seen": 760217600 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038871614844533604, + "loss": 1.4461, + "theoretical_loss": 3.7765842577326856, + "tokens_seen": 760283136 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038870611835506517, + "loss": 1.6624, + "theoretical_loss": 3.7765519363263396, + "tokens_seen": 760348672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886960882647944, + "loss": 1.4615, + "theoretical_loss": 3.7765196184856826, + "tokens_seen": 760414208 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886860581745236, + "loss": 1.5878, + "theoretical_loss": 3.776487304210012, + "tokens_seen": 760479744 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038867602808425277, + "loss": 1.4414, + "theoretical_loss": 3.7764549934986302, + "tokens_seen": 760545280 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038866599799398195, + "loss": 1.5521, + "theoretical_loss": 3.776422686350836, + "tokens_seen": 760610816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886559679037112, + "loss": 1.4688, + "theoretical_loss": 3.7763903827659275, + "tokens_seen": 760676352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886459378134403, + "loss": 1.5698, + "theoretical_loss": 3.776358082743208, + "tokens_seen": 760741888 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038863590772316955, + "loss": 1.4222, + "theoretical_loss": 3.776325786281975, + "tokens_seen": 760807424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886258776328987, + "loss": 1.5024, + "theoretical_loss": 3.776293493381531, + "tokens_seen": 760872960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886158475426279, + "loss": 1.5683, + "theoretical_loss": 3.776261204041176, + "tokens_seen": 760938496 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003886058174523571, + "loss": 1.4002, + "theoretical_loss": 3.7762289182602107, + "tokens_seen": 761004032 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038859578736208627, + "loss": 1.4359, + "theoretical_loss": 3.7761966360379366, + "tokens_seen": 761069568 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038858575727181545, + "loss": 1.475, + "theoretical_loss": 3.776164357373655, + "tokens_seen": 761135104 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038857572718154463, + "loss": 1.5591, + "theoretical_loss": 3.7761320822666677, + "tokens_seen": 761200640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003885656970912738, + "loss": 1.5671, + "theoretical_loss": 3.7760998107162767, + "tokens_seen": 761266176 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038855566700100305, + "loss": 1.4455, + "theoretical_loss": 3.7760675427217825, + "tokens_seen": 761331712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003885456369107322, + "loss": 1.4747, + "theoretical_loss": 3.776035278282489, + "tokens_seen": 761397248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003885356068204614, + "loss": 1.4285, + "theoretical_loss": 3.7760030173976977, + "tokens_seen": 761462784 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003885255767301906, + "loss": 1.4714, + "theoretical_loss": 3.775970760066711, + "tokens_seen": 761528320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003885155466399198, + "loss": 1.3531, + "theoretical_loss": 3.7759385062888318, + "tokens_seen": 761593856 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038850551654964896, + "loss": 1.4288, + "theoretical_loss": 3.7759062560633634, + "tokens_seen": 761659392 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038849548645937814, + "loss": 1.5187, + "theoretical_loss": 3.775874009389608, + "tokens_seen": 761724928 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003884854563691073, + "loss": 1.552, + "theoretical_loss": 3.77584176626687, + "tokens_seen": 761790464 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038847542627883655, + "loss": 1.5118, + "theoretical_loss": 3.7758095266944522, + "tokens_seen": 761856000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003884653961885657, + "loss": 1.4721, + "theoretical_loss": 3.775777290671659, + "tokens_seen": 761921536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003884553660982949, + "loss": 1.4877, + "theoretical_loss": 3.775745058197794, + "tokens_seen": 761987072 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038844533600802404, + "loss": 1.5785, + "theoretical_loss": 3.7757128292721607, + "tokens_seen": 762052608 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003884353059177533, + "loss": 1.615, + "theoretical_loss": 3.775680603894064, + "tokens_seen": 762118144 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038842527582748246, + "loss": 1.4743, + "theoretical_loss": 3.7756483820628084, + "tokens_seen": 762183680 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038841524573721164, + "loss": 1.4895, + "theoretical_loss": 3.7756161637776984, + "tokens_seen": 762249216 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003884052156469408, + "loss": 1.5175, + "theoretical_loss": 3.775583949038039, + "tokens_seen": 762314752 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038839518555667, + "loss": 1.4997, + "theoretical_loss": 3.775551737843135, + "tokens_seen": 762380288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883851554663992, + "loss": 1.5539, + "theoretical_loss": 3.775519530192293, + "tokens_seen": 762445824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883751253761284, + "loss": 1.4989, + "theoretical_loss": 3.775487326084816, + "tokens_seen": 762511360 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038836509528585755, + "loss": 1.4506, + "theoretical_loss": 3.775455125520012, + "tokens_seen": 762576896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883550651955868, + "loss": 1.4458, + "theoretical_loss": 3.7754229284971856, + "tokens_seen": 762642432 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038834503510531596, + "loss": 1.6122, + "theoretical_loss": 3.775390735015644, + "tokens_seen": 762707968 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038833500501504514, + "loss": 1.4715, + "theoretical_loss": 3.7753585450746927, + "tokens_seen": 762773504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883249749247743, + "loss": 1.5038, + "theoretical_loss": 3.775326358673637, + "tokens_seen": 762839040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883149448345035, + "loss": 1.5073, + "theoretical_loss": 3.775294175811786, + "tokens_seen": 762904576 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003883049147442327, + "loss": 1.5834, + "theoretical_loss": 3.775261996488445, + "tokens_seen": 762970112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003882948846539619, + "loss": 1.3794, + "theoretical_loss": 3.775229820702921, + "tokens_seen": 763035648 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038828485456369105, + "loss": 1.4909, + "theoretical_loss": 3.7751976484545224, + "tokens_seen": 763101184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003882748244734203, + "loss": 1.5212, + "theoretical_loss": 3.7751654797425553, + "tokens_seen": 763166720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003882647943831494, + "loss": 1.5518, + "theoretical_loss": 3.775133314566328, + "tokens_seen": 763232256 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038825476429287865, + "loss": 1.4526, + "theoretical_loss": 3.7751011529251475, + "tokens_seen": 763297792 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038824473420260783, + "loss": 1.5305, + "theoretical_loss": 3.7750689948183234, + "tokens_seen": 763363328 + }, + { + "epoch": 0.23, + "learning_rate": 0.000388234704112337, + "loss": 1.6197, + "theoretical_loss": 3.7750368402451624, + "tokens_seen": 763428864 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 282838, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.403523325920105, + "objective/train/theoretical_loss": 3.775004689204974, + "objective/train/tokens_used": 783954400, + "theoretical_loss": 3.775004689204974, + "tokens_seen": 763494400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003882246740220662, + "loss": 1.5055, + "theoretical_loss": 3.775004689204974, + "tokens_seen": 763494400 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038821464393179537, + "loss": 1.5976, + "theoretical_loss": 3.7749725416970668, + "tokens_seen": 763559936 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038820461384152455, + "loss": 1.526, + "theoretical_loss": 3.7749403977207483, + "tokens_seen": 763625472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881945837512538, + "loss": 1.4676, + "theoretical_loss": 3.774908257275329, + "tokens_seen": 763691008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881845536609829, + "loss": 1.5189, + "theoretical_loss": 3.7748761203601164, + "tokens_seen": 763756544 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038817452357071215, + "loss": 1.5175, + "theoretical_loss": 3.774843986974422, + "tokens_seen": 763822080 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038816449348044133, + "loss": 1.4959, + "theoretical_loss": 3.774811857117554, + "tokens_seen": 763887616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881544633901705, + "loss": 1.4706, + "theoretical_loss": 3.7747797307888225, + "tokens_seen": 763953152 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881444332998997, + "loss": 1.5357, + "theoretical_loss": 3.7747476079875373, + "tokens_seen": 764018688 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881344032096289, + "loss": 1.4876, + "theoretical_loss": 3.7747154887130088, + "tokens_seen": 764084224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881243731193581, + "loss": 1.548, + "theoretical_loss": 3.7746833729645473, + "tokens_seen": 764149760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003881143430290873, + "loss": 1.5247, + "theoretical_loss": 3.774651260741463, + "tokens_seen": 764215296 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038810431293881647, + "loss": 1.5968, + "theoretical_loss": 3.7746191520430674, + "tokens_seen": 764280832 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038809428284854565, + "loss": 1.3304, + "theoretical_loss": 3.774587046868671, + "tokens_seen": 764346368 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038808425275827483, + "loss": 1.4819, + "theoretical_loss": 3.7745549452175844, + "tokens_seen": 764411904 + }, + { + "epoch": 0.23, + "learning_rate": 0.000388074222668004, + "loss": 1.4726, + "theoretical_loss": 3.77452284708912, + "tokens_seen": 764477440 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038806419257773325, + "loss": 1.4602, + "theoretical_loss": 3.7744907524825884, + "tokens_seen": 764542976 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003880541624874624, + "loss": 1.5295, + "theoretical_loss": 3.774458661397302, + "tokens_seen": 764608512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003880441323971916, + "loss": 1.5515, + "theoretical_loss": 3.774426573832572, + "tokens_seen": 764674048 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003880341023069208, + "loss": 1.5795, + "theoretical_loss": 3.7743944897877113, + "tokens_seen": 764739584 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038802407221665, + "loss": 1.504, + "theoretical_loss": 3.7743624092620314, + "tokens_seen": 764805120 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038801404212637916, + "loss": 1.4738, + "theoretical_loss": 3.774330332254846, + "tokens_seen": 764870656 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038800401203610834, + "loss": 1.4408, + "theoretical_loss": 3.774298258765466, + "tokens_seen": 764936192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003879939819458375, + "loss": 1.5198, + "theoretical_loss": 3.774266188793206, + "tokens_seen": 765001728 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038798395185556675, + "loss": 1.5607, + "theoretical_loss": 3.7742341223373774, + "tokens_seen": 765067264 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003879739217652959, + "loss": 1.4818, + "theoretical_loss": 3.774202059397295, + "tokens_seen": 765132800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003879638916750251, + "loss": 1.4749, + "theoretical_loss": 3.774169999972271, + "tokens_seen": 765198336 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038795386158475424, + "loss": 1.4166, + "theoretical_loss": 3.77413794406162, + "tokens_seen": 765263872 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003879438314944835, + "loss": 1.4354, + "theoretical_loss": 3.7741058916646555, + "tokens_seen": 765329408 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038793380140421266, + "loss": 1.5181, + "theoretical_loss": 3.7740738427806915, + "tokens_seen": 765394944 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038792377131394184, + "loss": 1.4966, + "theoretical_loss": 3.7740417974090423, + "tokens_seen": 765460480 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387913741223671, + "loss": 1.5516, + "theoretical_loss": 3.774009755549022, + "tokens_seen": 765526016 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003879037111334002, + "loss": 1.6018, + "theoretical_loss": 3.773977717199946, + "tokens_seen": 765591552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878936810431294, + "loss": 1.5121, + "theoretical_loss": 3.773945682361128, + "tokens_seen": 765657088 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878836509528586, + "loss": 1.5749, + "theoretical_loss": 3.7739136510318834, + "tokens_seen": 765722624 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038787362086258775, + "loss": 1.3558, + "theoretical_loss": 3.7738816232115275, + "tokens_seen": 765788160 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387863590772317, + "loss": 1.641, + "theoretical_loss": 3.773849598899376, + "tokens_seen": 765853696 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038785356068204616, + "loss": 1.5727, + "theoretical_loss": 3.773817578094744, + "tokens_seen": 765919232 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038784353059177534, + "loss": 1.5792, + "theoretical_loss": 3.7737855607969477, + "tokens_seen": 765984768 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878335005015045, + "loss": 1.4881, + "theoretical_loss": 3.7737535470053025, + "tokens_seen": 766050304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878234704112337, + "loss": 1.4968, + "theoretical_loss": 3.7737215367191252, + "tokens_seen": 766115840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878134403209629, + "loss": 1.5179, + "theoretical_loss": 3.773689529937731, + "tokens_seen": 766181376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003878034102306921, + "loss": 1.4452, + "theoretical_loss": 3.773657526660438, + "tokens_seen": 766246912 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038779338014042125, + "loss": 1.4668, + "theoretical_loss": 3.773625526886562, + "tokens_seen": 766312448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003877833500501505, + "loss": 1.5959, + "theoretical_loss": 3.7735935306154196, + "tokens_seen": 766377984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003877733199598796, + "loss": 1.4761, + "theoretical_loss": 3.7735615378463283, + "tokens_seen": 766443520 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038776328986960885, + "loss": 1.6545, + "theoretical_loss": 3.7735295485786056, + "tokens_seen": 766509056 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038775325977933803, + "loss": 1.4934, + "theoretical_loss": 3.773497562811569, + "tokens_seen": 766574592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003877432296890672, + "loss": 1.5174, + "theoretical_loss": 3.773465580544536, + "tokens_seen": 766640128 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003877331995987964, + "loss": 1.568, + "theoretical_loss": 3.7734336017768237, + "tokens_seen": 766705664 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 283988, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4980601072311401, + "objective/train/theoretical_loss": 3.773401626507752, + "objective/train/tokens_used": 787231200, + "theoretical_loss": 3.773401626507752, + "tokens_seen": 766771200 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038772316950852557, + "loss": 1.4565, + "theoretical_loss": 3.773401626507752, + "tokens_seen": 766771200 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038771313941825475, + "loss": 1.529, + "theoretical_loss": 3.773369654736637, + "tokens_seen": 766836736 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387703109327984, + "loss": 1.4265, + "theoretical_loss": 3.773337686462799, + "tokens_seen": 766902272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876930792377131, + "loss": 1.5359, + "theoretical_loss": 3.7733057216855554, + "tokens_seen": 766967808 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038768304914744235, + "loss": 1.5484, + "theoretical_loss": 3.7732737604042255, + "tokens_seen": 767033344 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038767301905717153, + "loss": 1.4521, + "theoretical_loss": 3.773241802618129, + "tokens_seen": 767098880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876629889669007, + "loss": 1.4571, + "theoretical_loss": 3.7732098483265837, + "tokens_seen": 767164416 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876529588766299, + "loss": 1.533, + "theoretical_loss": 3.7731778975289094, + "tokens_seen": 767229952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876429287863591, + "loss": 1.5022, + "theoretical_loss": 3.7731459502244262, + "tokens_seen": 767295488 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038763289869608826, + "loss": 1.5071, + "theoretical_loss": 3.773114006412454, + "tokens_seen": 767361024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876228686058175, + "loss": 1.4214, + "theoretical_loss": 3.773082066092312, + "tokens_seen": 767426560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003876128385155466, + "loss": 1.5837, + "theoretical_loss": 3.773050129263321, + "tokens_seen": 767492096 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038760280842527585, + "loss": 1.505, + "theoretical_loss": 3.7730181959248013, + "tokens_seen": 767557632 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387592778335005, + "loss": 1.4806, + "theoretical_loss": 3.7729862660760727, + "tokens_seen": 767623168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875827482447342, + "loss": 1.4952, + "theoretical_loss": 3.7729543397164567, + "tokens_seen": 767688704 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875727181544634, + "loss": 1.5072, + "theoretical_loss": 3.772922416845274, + "tokens_seen": 767754240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875626880641926, + "loss": 1.4847, + "theoretical_loss": 3.7728904974618462, + "tokens_seen": 767819776 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038755265797392176, + "loss": 1.5029, + "theoretical_loss": 3.7728585815654934, + "tokens_seen": 767885312 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387542627883651, + "loss": 1.4762, + "theoretical_loss": 3.7728266691555383, + "tokens_seen": 767950848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875325977933801, + "loss": 1.4755, + "theoretical_loss": 3.7727947602313017, + "tokens_seen": 768016384 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038752256770310936, + "loss": 1.4552, + "theoretical_loss": 3.772762854792106, + "tokens_seen": 768081920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875125376128385, + "loss": 1.4583, + "theoretical_loss": 3.772730952837273, + "tokens_seen": 768147456 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003875025075225677, + "loss": 1.4556, + "theoretical_loss": 3.7726990543661256, + "tokens_seen": 768212992 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003874924774322969, + "loss": 1.4989, + "theoretical_loss": 3.7726671593779852, + "tokens_seen": 768278528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003874824473420261, + "loss": 1.5619, + "theoretical_loss": 3.7726352678721753, + "tokens_seen": 768344064 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038747241725175526, + "loss": 1.563, + "theoretical_loss": 3.772603379848018, + "tokens_seen": 768409600 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038746238716148444, + "loss": 1.4642, + "theoretical_loss": 3.772571495304837, + "tokens_seen": 768475136 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003874523570712136, + "loss": 1.4698, + "theoretical_loss": 3.7725396142419547, + "tokens_seen": 768540672 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038744232698094286, + "loss": 1.4544, + "theoretical_loss": 3.772507736658695, + "tokens_seen": 768606208 + }, + { + "epoch": 0.23, + "learning_rate": 0.000387432296890672, + "loss": 1.519, + "theoretical_loss": 3.7724758625543817, + "tokens_seen": 768671744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003874222668004012, + "loss": 1.5844, + "theoretical_loss": 3.7724439919283386, + "tokens_seen": 768737280 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038741223671013035, + "loss": 1.6178, + "theoretical_loss": 3.7724121247798887, + "tokens_seen": 768802816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003874022066198596, + "loss": 1.538, + "theoretical_loss": 3.7723802611083572, + "tokens_seen": 768868352 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038739217652958876, + "loss": 1.417, + "theoretical_loss": 3.7723484009130677, + "tokens_seen": 768933888 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038738214643931795, + "loss": 1.6482, + "theoretical_loss": 3.772316544193345, + "tokens_seen": 768999424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003873721163490472, + "loss": 1.5341, + "theoretical_loss": 3.772284690948514, + "tokens_seen": 769064960 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038736208625877636, + "loss": 1.4604, + "theoretical_loss": 3.7722528411778997, + "tokens_seen": 769130496 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038735205616850554, + "loss": 1.5258, + "theoretical_loss": 3.7722209948808265, + "tokens_seen": 769196032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003873420260782347, + "loss": 1.5498, + "theoretical_loss": 3.77218915205662, + "tokens_seen": 769261568 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003873319959879639, + "loss": 1.5505, + "theoretical_loss": 3.772157312704606, + "tokens_seen": 769327104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003873219658976931, + "loss": 1.489, + "theoretical_loss": 3.7721254768241104, + "tokens_seen": 769392640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003873119358074223, + "loss": 1.4791, + "theoretical_loss": 3.772093644414458, + "tokens_seen": 769458176 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038730190571715145, + "loss": 1.4327, + "theoretical_loss": 3.7720618154749754, + "tokens_seen": 769523712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872918756268807, + "loss": 1.4834, + "theoretical_loss": 3.772029990004989, + "tokens_seen": 769589248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872818455366098, + "loss": 1.5559, + "theoretical_loss": 3.771998168003825, + "tokens_seen": 769654784 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038727181544633905, + "loss": 1.4027, + "theoretical_loss": 3.77196634947081, + "tokens_seen": 769720320 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038726178535606823, + "loss": 1.5228, + "theoretical_loss": 3.7719345344052706, + "tokens_seen": 769785856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872517552657974, + "loss": 1.5109, + "theoretical_loss": 3.771902722806534, + "tokens_seen": 769851392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872417251755266, + "loss": 1.4, + "theoretical_loss": 3.7718709146739275, + "tokens_seen": 769916928 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038723169508525577, + "loss": 1.5199, + "theoretical_loss": 3.771839110006778, + "tokens_seen": 769982464 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 285055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5373198986053467, + "objective/train/theoretical_loss": 3.7718073088044135, + "objective/train/tokens_used": 790508000, + "theoretical_loss": 3.7718073088044135, + "tokens_seen": 770048000 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038722166499498495, + "loss": 1.5317, + "theoretical_loss": 3.7718073088044135, + "tokens_seen": 770048000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872116349047142, + "loss": 1.5171, + "theoretical_loss": 3.7717755110661617, + "tokens_seen": 770113536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003872016048144433, + "loss": 1.4533, + "theoretical_loss": 3.7717437167913497, + "tokens_seen": 770179072 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038719157472417255, + "loss": 1.5096, + "theoretical_loss": 3.771711925979307, + "tokens_seen": 770244608 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038718154463390173, + "loss": 1.5514, + "theoretical_loss": 3.7716801386293604, + "tokens_seen": 770310144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871715145436309, + "loss": 1.3903, + "theoretical_loss": 3.7716483547408393, + "tokens_seen": 770375680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871614844533601, + "loss": 1.4908, + "theoretical_loss": 3.7716165743130725, + "tokens_seen": 770441216 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871514543630893, + "loss": 1.3771, + "theoretical_loss": 3.7715847973453878, + "tokens_seen": 770506752 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038714142427281846, + "loss": 1.4629, + "theoretical_loss": 3.7715530238371153, + "tokens_seen": 770572288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871313941825477, + "loss": 1.4623, + "theoretical_loss": 3.7715212537875837, + "tokens_seen": 770637824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871213640922768, + "loss": 1.4575, + "theoretical_loss": 3.7714894871961233, + "tokens_seen": 770703360 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038711133400200605, + "loss": 1.4974, + "theoretical_loss": 3.7714577240620617, + "tokens_seen": 770768896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003871013039117352, + "loss": 1.4627, + "theoretical_loss": 3.771425964384731, + "tokens_seen": 770834432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870912738214644, + "loss": 1.5236, + "theoretical_loss": 3.7713942081634593, + "tokens_seen": 770899968 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870812437311936, + "loss": 1.4606, + "theoretical_loss": 3.7713624553975778, + "tokens_seen": 770965504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870712136409228, + "loss": 1.527, + "theoretical_loss": 3.7713307060864167, + "tokens_seen": 771031040 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038706118355065196, + "loss": 1.54, + "theoretical_loss": 3.771298960229306, + "tokens_seen": 771096576 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870511534603812, + "loss": 1.5399, + "theoretical_loss": 3.7712672178255775, + "tokens_seen": 771162112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870411233701103, + "loss": 1.6086, + "theoretical_loss": 3.7712354788745612, + "tokens_seen": 771227648 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038703109327983956, + "loss": 1.3813, + "theoretical_loss": 3.771203743375588, + "tokens_seen": 771293184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870210631895687, + "loss": 1.5066, + "theoretical_loss": 3.77117201132799, + "tokens_seen": 771358720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870110330992979, + "loss": 1.4982, + "theoretical_loss": 3.771140282731098, + "tokens_seen": 771424256 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003870010030090271, + "loss": 1.4831, + "theoretical_loss": 3.7711085575842436, + "tokens_seen": 771489792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003869909729187563, + "loss": 1.4816, + "theoretical_loss": 3.7710768358867597, + "tokens_seen": 771555328 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038698094282848546, + "loss": 1.5478, + "theoretical_loss": 3.771045117637977, + "tokens_seen": 771620864 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038697091273821464, + "loss": 1.4865, + "theoretical_loss": 3.7710134028372284, + "tokens_seen": 771686400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003869608826479438, + "loss": 1.6344, + "theoretical_loss": 3.770981691483846, + "tokens_seen": 771751936 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038695085255767306, + "loss": 1.4862, + "theoretical_loss": 3.7709499835771627, + "tokens_seen": 771817472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003869408224674022, + "loss": 1.447, + "theoretical_loss": 3.7709182791165112, + "tokens_seen": 771883008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003869307923771314, + "loss": 1.4782, + "theoretical_loss": 3.770886578101224, + "tokens_seen": 771948544 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038692076228686055, + "loss": 1.5177, + "theoretical_loss": 3.770854880530635, + "tokens_seen": 772014080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003869107321965898, + "loss": 1.4635, + "theoretical_loss": 3.770823186404077, + "tokens_seen": 772079616 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038690070210631896, + "loss": 1.5169, + "theoretical_loss": 3.770791495720883, + "tokens_seen": 772145152 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038689067201604815, + "loss": 1.4615, + "theoretical_loss": 3.770759808480388, + "tokens_seen": 772210688 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038688064192577733, + "loss": 1.3406, + "theoretical_loss": 3.770728124681925, + "tokens_seen": 772276224 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038687061183550656, + "loss": 1.6315, + "theoretical_loss": 3.7706964443248276, + "tokens_seen": 772341760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003868605817452357, + "loss": 1.5207, + "theoretical_loss": 3.7706647674084315, + "tokens_seen": 772407296 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003868505516549649, + "loss": 1.4101, + "theoretical_loss": 3.77063309393207, + "tokens_seen": 772472832 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038684052156469405, + "loss": 1.5182, + "theoretical_loss": 3.7706014238950782, + "tokens_seen": 772538368 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003868304914744233, + "loss": 1.3914, + "theoretical_loss": 3.7705697572967902, + "tokens_seen": 772603904 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038682046138415247, + "loss": 1.5447, + "theoretical_loss": 3.7705380941365423, + "tokens_seen": 772669440 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038681043129388165, + "loss": 1.4, + "theoretical_loss": 3.770506434413668, + "tokens_seen": 772734976 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038680040120361083, + "loss": 1.5606, + "theoretical_loss": 3.770474778127504, + "tokens_seen": 772800512 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038679037111334, + "loss": 1.5079, + "theoretical_loss": 3.7704431252773856, + "tokens_seen": 772866048 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003867803410230692, + "loss": 1.6078, + "theoretical_loss": 3.770411475862648, + "tokens_seen": 772931584 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038677031093279843, + "loss": 1.4999, + "theoretical_loss": 3.7703798298826277, + "tokens_seen": 772997120 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038676028084252755, + "loss": 1.681, + "theoretical_loss": 3.77034818733666, + "tokens_seen": 773062656 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003867502507522568, + "loss": 1.5594, + "theoretical_loss": 3.7703165482240824, + "tokens_seen": 773128192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003867402206619859, + "loss": 1.4764, + "theoretical_loss": 3.77028491254423, + "tokens_seen": 773193728 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038673019057171515, + "loss": 1.3997, + "theoretical_loss": 3.7702532802964406, + "tokens_seen": 773259264 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 286255, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.400518774986267, + "objective/train/theoretical_loss": 3.7702216514800506, + "objective/train/tokens_used": 793784800, + "theoretical_loss": 3.7702216514800506, + "tokens_seen": 773324800 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038672016048144433, + "loss": 1.4969, + "theoretical_loss": 3.7702216514800506, + "tokens_seen": 773324800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003867101303911735, + "loss": 1.4971, + "theoretical_loss": 3.770190026094397, + "tokens_seen": 773390336 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003867001003009027, + "loss": 1.4424, + "theoretical_loss": 3.7701584041388165, + "tokens_seen": 773455872 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038669007021063193, + "loss": 1.5406, + "theoretical_loss": 3.7701267856126472, + "tokens_seen": 773521408 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038668004012036106, + "loss": 1.4364, + "theoretical_loss": 3.7700951705152264, + "tokens_seen": 773586944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003866700100300903, + "loss": 1.574, + "theoretical_loss": 3.7700635588458926, + "tokens_seen": 773652480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003866599799398194, + "loss": 1.5288, + "theoretical_loss": 3.7700319506039826, + "tokens_seen": 773718016 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038664994984954866, + "loss": 1.5699, + "theoretical_loss": 3.7700003457888345, + "tokens_seen": 773783552 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038663991975927784, + "loss": 1.5091, + "theoretical_loss": 3.7699687443997876, + "tokens_seen": 773849088 + }, + { + "epoch": 0.23, + "learning_rate": 0.000386629889669007, + "loss": 1.5552, + "theoretical_loss": 3.7699371464361797, + "tokens_seen": 773914624 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038661985957873625, + "loss": 1.5121, + "theoretical_loss": 3.7699055518973497, + "tokens_seen": 773980160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003866098294884654, + "loss": 1.4537, + "theoretical_loss": 3.769873960782636, + "tokens_seen": 774045696 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865997993981946, + "loss": 1.4439, + "theoretical_loss": 3.769842373091379, + "tokens_seen": 774111232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865897693079238, + "loss": 1.6406, + "theoretical_loss": 3.769810788822916, + "tokens_seen": 774176768 + }, + { + "epoch": 0.23, + "learning_rate": 0.000386579739217653, + "loss": 1.4824, + "theoretical_loss": 3.769779207976588, + "tokens_seen": 774242304 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038656970912738216, + "loss": 1.6147, + "theoretical_loss": 3.7697476305517337, + "tokens_seen": 774307840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865596790371114, + "loss": 1.4789, + "theoretical_loss": 3.7697160565476935, + "tokens_seen": 774373376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865496489468405, + "loss": 1.4408, + "theoretical_loss": 3.769684485963807, + "tokens_seen": 774438912 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038653961885656976, + "loss": 1.5844, + "theoretical_loss": 3.7696529187994137, + "tokens_seen": 774504448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865295887662989, + "loss": 1.5368, + "theoretical_loss": 3.769621355053855, + "tokens_seen": 774569984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865195586760281, + "loss": 1.5947, + "theoretical_loss": 3.769589794726471, + "tokens_seen": 774635520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003865095285857573, + "loss": 1.4907, + "theoretical_loss": 3.7695582378166024, + "tokens_seen": 774701056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003864994984954865, + "loss": 1.605, + "theoretical_loss": 3.7695266843235897, + "tokens_seen": 774766592 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038648946840521566, + "loss": 1.5432, + "theoretical_loss": 3.7694951342467746, + "tokens_seen": 774832128 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038647943831494484, + "loss": 1.5231, + "theoretical_loss": 3.7694635875854985, + "tokens_seen": 774897664 + }, + { + "epoch": 0.23, + "learning_rate": 0.000386469408224674, + "loss": 1.5337, + "theoretical_loss": 3.7694320443391023, + "tokens_seen": 774963200 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038645937813440326, + "loss": 1.4519, + "theoretical_loss": 3.769400504506927, + "tokens_seen": 775028736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003864493480441324, + "loss": 1.6015, + "theoretical_loss": 3.7693689680883153, + "tokens_seen": 775094272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0003864393179538616, + "loss": 1.4105, + "theoretical_loss": 3.769337435082609, + "tokens_seen": 775159808 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038642928786359075, + "loss": 1.5318, + "theoretical_loss": 3.769305905489151, + "tokens_seen": 775225344 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038641925777332, + "loss": 1.6661, + "theoretical_loss": 3.769274379307282, + "tokens_seen": 775290880 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038640922768304917, + "loss": 1.4702, + "theoretical_loss": 3.7692428565363456, + "tokens_seen": 775356416 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038639919759277835, + "loss": 1.5441, + "theoretical_loss": 3.769211337175684, + "tokens_seen": 775421952 + }, + { + "epoch": 0.23, + "learning_rate": 0.00038638916750250753, + "loss": 1.4823, + "theoretical_loss": 3.7691798212246406, + "tokens_seen": 775487488 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038637913741223676, + "loss": 1.4547, + "theoretical_loss": 3.769148308682558, + "tokens_seen": 775553024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003863691073219659, + "loss": 1.5766, + "theoretical_loss": 3.76911679954878, + "tokens_seen": 775618560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003863590772316951, + "loss": 1.494, + "theoretical_loss": 3.7690852938226493, + "tokens_seen": 775684096 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038634904714142425, + "loss": 1.5004, + "theoretical_loss": 3.7690537915035103, + "tokens_seen": 775749632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003863390170511535, + "loss": 1.6863, + "theoretical_loss": 3.769022292590706, + "tokens_seen": 775815168 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038632898696088267, + "loss": 1.561, + "theoretical_loss": 3.768990797083581, + "tokens_seen": 775880704 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038631895687061185, + "loss": 1.4531, + "theoretical_loss": 3.7689593049814794, + "tokens_seen": 775946240 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038630892678034103, + "loss": 1.4222, + "theoretical_loss": 3.768927816283745, + "tokens_seen": 776011776 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003862988966900702, + "loss": 1.533, + "theoretical_loss": 3.768896330989723, + "tokens_seen": 776077312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003862888665997994, + "loss": 1.5991, + "theoretical_loss": 3.7688648490987577, + "tokens_seen": 776142848 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038627883650952863, + "loss": 1.5297, + "theoretical_loss": 3.768833370610194, + "tokens_seen": 776208384 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038626880641925775, + "loss": 1.424, + "theoretical_loss": 3.768801895523377, + "tokens_seen": 776273920 + }, + { + "epoch": 0.24, + "learning_rate": 0.000386258776328987, + "loss": 1.4702, + "theoretical_loss": 3.768770423837652, + "tokens_seen": 776339456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003862487462387161, + "loss": 1.6189, + "theoretical_loss": 3.7687389555523643, + "tokens_seen": 776404992 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038623871614844535, + "loss": 1.4753, + "theoretical_loss": 3.7687074906668596, + "tokens_seen": 776470528 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038622868605817453, + "loss": 1.4958, + "theoretical_loss": 3.768676029180484, + "tokens_seen": 776536064 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 287303, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5367624759674072, + "objective/train/theoretical_loss": 3.768644571092583, + "objective/train/tokens_used": 797061600, + "theoretical_loss": 3.768644571092583, + "tokens_seen": 776601600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003862186559679037, + "loss": 1.5442, + "theoretical_loss": 3.768644571092583, + "tokens_seen": 776601600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003862086258776329, + "loss": 1.7246, + "theoretical_loss": 3.768613116402503, + "tokens_seen": 776667136 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038619859578736213, + "loss": 1.4368, + "theoretical_loss": 3.76858166510959, + "tokens_seen": 776732672 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038618856569709126, + "loss": 1.5486, + "theoretical_loss": 3.768550217213191, + "tokens_seen": 776798208 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003861785356068205, + "loss": 1.5736, + "theoretical_loss": 3.7685187727126523, + "tokens_seen": 776863744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003861685055165496, + "loss": 1.441, + "theoretical_loss": 3.768487331607321, + "tokens_seen": 776929280 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038615847542627886, + "loss": 1.4606, + "theoretical_loss": 3.7684558938965442, + "tokens_seen": 776994816 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038614844533600804, + "loss": 1.4788, + "theoretical_loss": 3.768424459579669, + "tokens_seen": 777060352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003861384152457372, + "loss": 1.6038, + "theoretical_loss": 3.7683930286560425, + "tokens_seen": 777125888 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003861283851554664, + "loss": 1.5255, + "theoretical_loss": 3.7683616011250134, + "tokens_seen": 777191424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003861183550651956, + "loss": 1.4175, + "theoretical_loss": 3.768330176985928, + "tokens_seen": 777256960 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038610832497492476, + "loss": 1.5209, + "theoretical_loss": 3.7682987562381354, + "tokens_seen": 777322496 + }, + { + "epoch": 0.24, + "learning_rate": 0.000386098294884654, + "loss": 1.4917, + "theoretical_loss": 3.768267338880983, + "tokens_seen": 777388032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860882647943831, + "loss": 1.5327, + "theoretical_loss": 3.7682359249138195, + "tokens_seen": 777453568 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038607823470411236, + "loss": 1.4371, + "theoretical_loss": 3.7682045143359937, + "tokens_seen": 777519104 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860682046138415, + "loss": 1.5718, + "theoretical_loss": 3.7681731071468536, + "tokens_seen": 777584640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860581745235707, + "loss": 1.5665, + "theoretical_loss": 3.7681417033457487, + "tokens_seen": 777650176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860481444332999, + "loss": 1.5465, + "theoretical_loss": 3.768110302932027, + "tokens_seen": 777715712 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860381143430291, + "loss": 1.5013, + "theoretical_loss": 3.768078905905039, + "tokens_seen": 777781248 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038602808425275826, + "loss": 1.4824, + "theoretical_loss": 3.7680475122641335, + "tokens_seen": 777846784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860180541624875, + "loss": 1.5703, + "theoretical_loss": 3.7680161220086603, + "tokens_seen": 777912320 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003860080240722166, + "loss": 1.455, + "theoretical_loss": 3.7679847351379685, + "tokens_seen": 777977856 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038599799398194586, + "loss": 1.4239, + "theoretical_loss": 3.7679533516514088, + "tokens_seen": 778043392 + }, + { + "epoch": 0.24, + "learning_rate": 0.000385987963891675, + "loss": 1.5704, + "theoretical_loss": 3.767921971548331, + "tokens_seen": 778108928 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003859779338014042, + "loss": 1.4802, + "theoretical_loss": 3.767890594828085, + "tokens_seen": 778174464 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003859679037111334, + "loss": 1.3704, + "theoretical_loss": 3.7678592214900224, + "tokens_seen": 778240000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003859578736208626, + "loss": 1.4121, + "theoretical_loss": 3.7678278515334926, + "tokens_seen": 778305536 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038594784353059177, + "loss": 1.391, + "theoretical_loss": 3.7677964849578474, + "tokens_seen": 778371072 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038593781344032095, + "loss": 1.3472, + "theoretical_loss": 3.7677651217624373, + "tokens_seen": 778436608 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038592778335005013, + "loss": 1.467, + "theoretical_loss": 3.767733761946613, + "tokens_seen": 778502144 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038591775325977937, + "loss": 1.5009, + "theoretical_loss": 3.767702405509727, + "tokens_seen": 778567680 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003859077231695085, + "loss": 1.4689, + "theoretical_loss": 3.7676710524511305, + "tokens_seen": 778633216 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038589769307923773, + "loss": 1.4745, + "theoretical_loss": 3.767639702770175, + "tokens_seen": 778698752 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038588766298896685, + "loss": 1.3309, + "theoretical_loss": 3.767608356466212, + "tokens_seen": 778764288 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003858776328986961, + "loss": 1.3999, + "theoretical_loss": 3.7675770135385944, + "tokens_seen": 778829824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003858676028084253, + "loss": 1.5368, + "theoretical_loss": 3.7675456739866746, + "tokens_seen": 778895360 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038585757271815445, + "loss": 1.5363, + "theoretical_loss": 3.7675143378098035, + "tokens_seen": 778960896 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003858475426278837, + "loss": 1.4552, + "theoretical_loss": 3.7674830050073354, + "tokens_seen": 779026432 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038583751253761287, + "loss": 1.4871, + "theoretical_loss": 3.7674516755786227, + "tokens_seen": 779091968 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038582748244734205, + "loss": 1.4566, + "theoretical_loss": 3.767420349523018, + "tokens_seen": 779157504 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038581745235707123, + "loss": 1.5175, + "theoretical_loss": 3.7673890268398744, + "tokens_seen": 779223040 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003858074222668004, + "loss": 1.5718, + "theoretical_loss": 3.7673577075285456, + "tokens_seen": 779288576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003857973921765296, + "loss": 1.4168, + "theoretical_loss": 3.767326391588385, + "tokens_seen": 779354112 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038578736208625883, + "loss": 1.5012, + "theoretical_loss": 3.767295079018746, + "tokens_seen": 779419648 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038577733199598796, + "loss": 1.4838, + "theoretical_loss": 3.7672637698189835, + "tokens_seen": 779485184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003857673019057172, + "loss": 1.4946, + "theoretical_loss": 3.7672324639884502, + "tokens_seen": 779550720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003857572718154463, + "loss": 1.5701, + "theoretical_loss": 3.7672011615265015, + "tokens_seen": 779616256 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038574724172517555, + "loss": 1.5431, + "theoretical_loss": 3.7671698624324907, + "tokens_seen": 779681792 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038573721163490473, + "loss": 1.5087, + "theoretical_loss": 3.767138566705773, + "tokens_seen": 779747328 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003857271815446339, + "loss": 1.5515, + "theoretical_loss": 3.7671072743457032, + "tokens_seen": 779812864 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 288522, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4792721271514893, + "objective/train/theoretical_loss": 3.767075985351636, + "objective/train/tokens_used": 800338400, + "theoretical_loss": 3.767075985351636, + "tokens_seen": 779878400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003857171514543631, + "loss": 1.5901, + "theoretical_loss": 3.767075985351636, + "tokens_seen": 779878400 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038570712136409233, + "loss": 1.6366, + "theoretical_loss": 3.767044699722927, + "tokens_seen": 779943936 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038569709127382146, + "loss": 1.5637, + "theoretical_loss": 3.767013417458931, + "tokens_seen": 780009472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856870611835507, + "loss": 1.5356, + "theoretical_loss": 3.766982138559004, + "tokens_seen": 780075008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856770310932798, + "loss": 1.4695, + "theoretical_loss": 3.7669508630225006, + "tokens_seen": 780140544 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038566700100300906, + "loss": 1.4879, + "theoretical_loss": 3.7669195908487776, + "tokens_seen": 780206080 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038565697091273824, + "loss": 1.412, + "theoretical_loss": 3.766888322037191, + "tokens_seen": 780271616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856469408224674, + "loss": 1.5434, + "theoretical_loss": 3.7668570565870967, + "tokens_seen": 780337152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856369107321966, + "loss": 1.6588, + "theoretical_loss": 3.7668257944978505, + "tokens_seen": 780402688 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856268806419258, + "loss": 1.5363, + "theoretical_loss": 3.7667945357688097, + "tokens_seen": 780468224 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038561685055165496, + "loss": 1.5135, + "theoretical_loss": 3.7667632803993314, + "tokens_seen": 780533760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003856068204613842, + "loss": 1.5077, + "theoretical_loss": 3.7667320283887715, + "tokens_seen": 780599296 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855967903711133, + "loss": 1.5807, + "theoretical_loss": 3.7667007797364875, + "tokens_seen": 780664832 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038558676028084256, + "loss": 1.4505, + "theoretical_loss": 3.766669534441837, + "tokens_seen": 780730368 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855767301905717, + "loss": 1.4987, + "theoretical_loss": 3.766638292504177, + "tokens_seen": 780795904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855667001003009, + "loss": 1.5721, + "theoretical_loss": 3.766607053922865, + "tokens_seen": 780861440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855566700100301, + "loss": 1.4618, + "theoretical_loss": 3.7665758186972584, + "tokens_seen": 780926976 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855466399197593, + "loss": 1.5063, + "theoretical_loss": 3.7665445868267162, + "tokens_seen": 780992512 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038553660982948846, + "loss": 1.4553, + "theoretical_loss": 3.766513358310596, + "tokens_seen": 781058048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855265797392177, + "loss": 1.5556, + "theoretical_loss": 3.7664821331482568, + "tokens_seen": 781123584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003855165496489468, + "loss": 1.4467, + "theoretical_loss": 3.7664509113390556, + "tokens_seen": 781189120 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038550651955867606, + "loss": 1.4531, + "theoretical_loss": 3.7664196928823523, + "tokens_seen": 781254656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003854964894684052, + "loss": 1.4698, + "theoretical_loss": 3.7663884777775047, + "tokens_seen": 781320192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003854864593781344, + "loss": 1.5453, + "theoretical_loss": 3.7663572660238733, + "tokens_seen": 781385728 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003854764292878636, + "loss": 1.4899, + "theoretical_loss": 3.766326057620816, + "tokens_seen": 781451264 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003854663991975928, + "loss": 1.4971, + "theoretical_loss": 3.7662948525676927, + "tokens_seen": 781516800 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038545636910732197, + "loss": 1.4166, + "theoretical_loss": 3.7662636508638627, + "tokens_seen": 781582336 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038544633901705115, + "loss": 1.4798, + "theoretical_loss": 3.7662324525086865, + "tokens_seen": 781647872 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038543630892678033, + "loss": 1.6717, + "theoretical_loss": 3.766201257501523, + "tokens_seen": 781713408 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038542627883650957, + "loss": 1.4851, + "theoretical_loss": 3.7661700658417328, + "tokens_seen": 781778944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003854162487462387, + "loss": 1.3817, + "theoretical_loss": 3.7661388775286753, + "tokens_seen": 781844480 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038540621865596793, + "loss": 1.5267, + "theoretical_loss": 3.7661076925617123, + "tokens_seen": 781910016 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038539618856569705, + "loss": 1.4264, + "theoretical_loss": 3.7660765109402035, + "tokens_seen": 781975552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003853861584754263, + "loss": 1.5141, + "theoretical_loss": 3.76604533266351, + "tokens_seen": 782041088 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038537612838515547, + "loss": 1.471, + "theoretical_loss": 3.7660141577309925, + "tokens_seen": 782106624 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038536609829488465, + "loss": 1.5088, + "theoretical_loss": 3.7659829861420127, + "tokens_seen": 782172160 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038535606820461383, + "loss": 1.5986, + "theoretical_loss": 3.765951817895931, + "tokens_seen": 782237696 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038534603811434307, + "loss": 1.4848, + "theoretical_loss": 3.765920652992109, + "tokens_seen": 782303232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003853360080240722, + "loss": 1.4962, + "theoretical_loss": 3.7658894914299097, + "tokens_seen": 782368768 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038532597793380143, + "loss": 1.3945, + "theoretical_loss": 3.7658583332086932, + "tokens_seen": 782434304 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038531594784353056, + "loss": 1.4257, + "theoretical_loss": 3.7658271783278225, + "tokens_seen": 782499840 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003853059177532598, + "loss": 1.5469, + "theoretical_loss": 3.765796026786659, + "tokens_seen": 782565376 + }, + { + "epoch": 0.24, + "learning_rate": 0.000385295887662989, + "loss": 1.4302, + "theoretical_loss": 3.7657648785845663, + "tokens_seen": 782630912 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038528585757271816, + "loss": 1.4895, + "theoretical_loss": 3.765733733720906, + "tokens_seen": 782696448 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038527582748244734, + "loss": 1.4884, + "theoretical_loss": 3.7657025921950407, + "tokens_seen": 782761984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003852657973921765, + "loss": 1.3524, + "theoretical_loss": 3.765671454006334, + "tokens_seen": 782827520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003852557673019057, + "loss": 1.5586, + "theoretical_loss": 3.7656403191541483, + "tokens_seen": 782893056 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038524573721163493, + "loss": 1.5679, + "theoretical_loss": 3.765609187637847, + "tokens_seen": 782958592 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038523570712136406, + "loss": 1.3994, + "theoretical_loss": 3.765578059456794, + "tokens_seen": 783024128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003852256770310933, + "loss": 1.4549, + "theoretical_loss": 3.765546934610352, + "tokens_seen": 783089664 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 289672, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6125402450561523, + "objective/train/theoretical_loss": 3.7655158130978856, + "objective/train/tokens_used": 803615200, + "theoretical_loss": 3.7655158130978856, + "tokens_seen": 783155200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003852156469408225, + "loss": 1.5055, + "theoretical_loss": 3.7655158130978856, + "tokens_seen": 783155200 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038520561685055166, + "loss": 1.5345, + "theoretical_loss": 3.765484694918758, + "tokens_seen": 783220736 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038519558676028084, + "loss": 1.4781, + "theoretical_loss": 3.7654535800723337, + "tokens_seen": 783286272 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038518555667001, + "loss": 1.3971, + "theoretical_loss": 3.7654224685579765, + "tokens_seen": 783351808 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003851755265797392, + "loss": 1.5309, + "theoretical_loss": 3.765391360375052, + "tokens_seen": 783417344 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038516549648946844, + "loss": 1.5293, + "theoretical_loss": 3.765360255522924, + "tokens_seen": 783482880 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038515546639919756, + "loss": 1.4597, + "theoretical_loss": 3.7653291540009572, + "tokens_seen": 783548416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003851454363089268, + "loss": 1.5675, + "theoretical_loss": 3.7652980558085165, + "tokens_seen": 783613952 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003851354062186559, + "loss": 1.6057, + "theoretical_loss": 3.7652669609449676, + "tokens_seen": 783679488 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038512537612838516, + "loss": 1.6268, + "theoretical_loss": 3.7652358694096755, + "tokens_seen": 783745024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003851153460381144, + "loss": 1.4642, + "theoretical_loss": 3.765204781202006, + "tokens_seen": 783810560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003851053159478435, + "loss": 1.4422, + "theoretical_loss": 3.765173696321324, + "tokens_seen": 783876096 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038509528585757276, + "loss": 1.4416, + "theoretical_loss": 3.7651426147669964, + "tokens_seen": 783941632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850852557673019, + "loss": 1.4971, + "theoretical_loss": 3.765111536538388, + "tokens_seen": 784007168 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850752256770311, + "loss": 1.4437, + "theoretical_loss": 3.7650804616348665, + "tokens_seen": 784072704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850651955867603, + "loss": 1.3384, + "theoretical_loss": 3.765049390055797, + "tokens_seen": 784138240 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850551654964895, + "loss": 1.505, + "theoretical_loss": 3.765018321800546, + "tokens_seen": 784203776 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038504513540621866, + "loss": 1.5515, + "theoretical_loss": 3.7649872568684812, + "tokens_seen": 784269312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850351053159479, + "loss": 1.5232, + "theoretical_loss": 3.764956195258969, + "tokens_seen": 784334848 + }, + { + "epoch": 0.24, + "learning_rate": 0.000385025075225677, + "loss": 1.3463, + "theoretical_loss": 3.7649251369713763, + "tokens_seen": 784400384 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038501504513540626, + "loss": 1.4143, + "theoretical_loss": 3.7648940820050707, + "tokens_seen": 784465920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003850050150451354, + "loss": 1.3744, + "theoretical_loss": 3.764863030359419, + "tokens_seen": 784531456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003849949849548646, + "loss": 1.3773, + "theoretical_loss": 3.7648319820337894, + "tokens_seen": 784596992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003849849548645938, + "loss": 1.4903, + "theoretical_loss": 3.7648009370275495, + "tokens_seen": 784662528 + }, + { + "epoch": 0.24, + "learning_rate": 0.000384974924774323, + "loss": 1.4484, + "theoretical_loss": 3.7647698953400672, + "tokens_seen": 784728064 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038496489468405217, + "loss": 1.4858, + "theoretical_loss": 3.76473885697071, + "tokens_seen": 784793600 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038495486459378135, + "loss": 1.4459, + "theoretical_loss": 3.7647078219188472, + "tokens_seen": 784859136 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038494483450351053, + "loss": 1.5471, + "theoretical_loss": 3.7646767901838465, + "tokens_seen": 784924672 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038493480441323977, + "loss": 1.4954, + "theoretical_loss": 3.764645761765077, + "tokens_seen": 784990208 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003849247743229689, + "loss": 1.4367, + "theoretical_loss": 3.7646147366619074, + "tokens_seen": 785055744 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038491474423269813, + "loss": 1.5555, + "theoretical_loss": 3.764583714873706, + "tokens_seen": 785121280 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038490471414242725, + "loss": 1.4344, + "theoretical_loss": 3.764552696399843, + "tokens_seen": 785186816 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003848946840521565, + "loss": 1.439, + "theoretical_loss": 3.764521681239687, + "tokens_seen": 785252352 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038488465396188567, + "loss": 1.4702, + "theoretical_loss": 3.7644906693926075, + "tokens_seen": 785317888 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038487462387161485, + "loss": 1.6403, + "theoretical_loss": 3.7644596608579746, + "tokens_seen": 785383424 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038486459378134403, + "loss": 1.5034, + "theoretical_loss": 3.764428655635158, + "tokens_seen": 785448960 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038485456369107327, + "loss": 1.4219, + "theoretical_loss": 3.764397653723527, + "tokens_seen": 785514496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003848445336008024, + "loss": 1.4578, + "theoretical_loss": 3.7643666551224526, + "tokens_seen": 785580032 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038483450351053163, + "loss": 1.457, + "theoretical_loss": 3.764335659831305, + "tokens_seen": 785645568 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038482447342026076, + "loss": 1.4151, + "theoretical_loss": 3.7643046678494545, + "tokens_seen": 785711104 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038481444332999, + "loss": 1.4361, + "theoretical_loss": 3.764273679176272, + "tokens_seen": 785776640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003848044132397192, + "loss": 1.6029, + "theoretical_loss": 3.764242693811128, + "tokens_seen": 785842176 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038479438314944836, + "loss": 1.5098, + "theoretical_loss": 3.764211711753394, + "tokens_seen": 785907712 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038478435305917754, + "loss": 1.368, + "theoretical_loss": 3.7641807330024406, + "tokens_seen": 785973248 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003847743229689067, + "loss": 1.3929, + "theoretical_loss": 3.7641497575576404, + "tokens_seen": 786038784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003847642928786359, + "loss": 1.4644, + "theoretical_loss": 3.764118785418363, + "tokens_seen": 786104320 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038475426278836513, + "loss": 1.579, + "theoretical_loss": 3.7640878165839817, + "tokens_seen": 786169856 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038474423269809426, + "loss": 1.4584, + "theoretical_loss": 3.7640568510538683, + "tokens_seen": 786235392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003847342026078235, + "loss": 1.4377, + "theoretical_loss": 3.764025888827394, + "tokens_seen": 786300928 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003847241725175527, + "loss": 1.4595, + "theoretical_loss": 3.763994929903932, + "tokens_seen": 786366464 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 290810, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.403365135192871, + "objective/train/theoretical_loss": 3.763963974282854, + "objective/train/tokens_used": 806892000, + "theoretical_loss": 3.763963974282854, + "tokens_seen": 786432000 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038471414242728186, + "loss": 1.4659, + "theoretical_loss": 3.763963974282854, + "tokens_seen": 786432000 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038470411233701104, + "loss": 1.4453, + "theoretical_loss": 3.763933021963533, + "tokens_seen": 786497536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003846940822467402, + "loss": 1.4661, + "theoretical_loss": 3.763902072945341, + "tokens_seen": 786563072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003846840521564694, + "loss": 1.4585, + "theoretical_loss": 3.7638711272276524, + "tokens_seen": 786628608 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038467402206619864, + "loss": 1.4847, + "theoretical_loss": 3.7638401848098386, + "tokens_seen": 786694144 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038466399197592776, + "loss": 1.6001, + "theoretical_loss": 3.763809245691274, + "tokens_seen": 786759680 + }, + { + "epoch": 0.24, + "learning_rate": 0.000384653961885657, + "loss": 1.4329, + "theoretical_loss": 3.7637783098713316, + "tokens_seen": 786825216 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003846439317953861, + "loss": 1.4853, + "theoretical_loss": 3.7637473773493855, + "tokens_seen": 786890752 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038463390170511536, + "loss": 1.4549, + "theoretical_loss": 3.7637164481248084, + "tokens_seen": 786956288 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038462387161484454, + "loss": 1.4188, + "theoretical_loss": 3.763685522196975, + "tokens_seen": 787021824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003846138415245737, + "loss": 1.5265, + "theoretical_loss": 3.7636545995652604, + "tokens_seen": 787087360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003846038114343029, + "loss": 1.5854, + "theoretical_loss": 3.7636236802290366, + "tokens_seen": 787152896 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003845937813440321, + "loss": 1.5282, + "theoretical_loss": 3.7635927641876803, + "tokens_seen": 787218432 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038458375125376127, + "loss": 1.5239, + "theoretical_loss": 3.7635618514405644, + "tokens_seen": 787283968 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003845737211634905, + "loss": 1.5566, + "theoretical_loss": 3.7635309419870646, + "tokens_seen": 787349504 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038456369107321963, + "loss": 1.4214, + "theoretical_loss": 3.763500035826556, + "tokens_seen": 787415040 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038455366098294886, + "loss": 1.409, + "theoretical_loss": 3.7634691329584133, + "tokens_seen": 787480576 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038454363089267805, + "loss": 1.4766, + "theoretical_loss": 3.763438233382012, + "tokens_seen": 787546112 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038453360080240723, + "loss": 1.5539, + "theoretical_loss": 3.7634073370967274, + "tokens_seen": 787611648 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003845235707121364, + "loss": 1.4146, + "theoretical_loss": 3.7633764441019357, + "tokens_seen": 787677184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003845135406218656, + "loss": 1.3656, + "theoretical_loss": 3.763345554397012, + "tokens_seen": 787742720 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038450351053159477, + "loss": 1.4866, + "theoretical_loss": 3.763314667981333, + "tokens_seen": 787808256 + }, + { + "epoch": 0.24, + "learning_rate": 0.000384493480441324, + "loss": 1.5892, + "theoretical_loss": 3.7632837848542744, + "tokens_seen": 787873792 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038448345035105313, + "loss": 1.3174, + "theoretical_loss": 3.7632529050152126, + "tokens_seen": 787939328 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038447342026078237, + "loss": 1.4672, + "theoretical_loss": 3.7632220284635247, + "tokens_seen": 788004864 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003844633901705115, + "loss": 1.4248, + "theoretical_loss": 3.763191155198586, + "tokens_seen": 788070400 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038445336008024073, + "loss": 1.4787, + "theoretical_loss": 3.7631602852197745, + "tokens_seen": 788135936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003844433299899699, + "loss": 1.5365, + "theoretical_loss": 3.763129418526467, + "tokens_seen": 788201472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003844332998996991, + "loss": 1.3987, + "theoretical_loss": 3.7630985551180407, + "tokens_seen": 788267008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003844232698094283, + "loss": 1.4189, + "theoretical_loss": 3.7630676949938726, + "tokens_seen": 788332544 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038441323971915745, + "loss": 1.4319, + "theoretical_loss": 3.763036838153341, + "tokens_seen": 788398080 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038440320962888664, + "loss": 1.5471, + "theoretical_loss": 3.763005984595823, + "tokens_seen": 788463616 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038439317953861587, + "loss": 1.5097, + "theoretical_loss": 3.762975134320696, + "tokens_seen": 788529152 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038438314944834505, + "loss": 1.4999, + "theoretical_loss": 3.762944287327339, + "tokens_seen": 788594688 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038437311935807423, + "loss": 1.4089, + "theoretical_loss": 3.7629134436151297, + "tokens_seen": 788660224 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038436308926780347, + "loss": 1.5394, + "theoretical_loss": 3.7628826031834466, + "tokens_seen": 788725760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003843530591775326, + "loss": 1.5246, + "theoretical_loss": 3.7628517660316687, + "tokens_seen": 788791296 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038434302908726183, + "loss": 1.5159, + "theoretical_loss": 3.7628209321591735, + "tokens_seen": 788856832 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038433299899699096, + "loss": 1.5577, + "theoretical_loss": 3.762790101565341, + "tokens_seen": 788922368 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003843229689067202, + "loss": 1.4362, + "theoretical_loss": 3.76275927424955, + "tokens_seen": 788987904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003843129388164494, + "loss": 1.4301, + "theoretical_loss": 3.7627284502111795, + "tokens_seen": 789053440 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038430290872617856, + "loss": 1.5473, + "theoretical_loss": 3.762697629449609, + "tokens_seen": 789118976 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038429287863590774, + "loss": 1.4428, + "theoretical_loss": 3.762666811964218, + "tokens_seen": 789184512 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003842828485456369, + "loss": 1.519, + "theoretical_loss": 3.762635997754386, + "tokens_seen": 789250048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003842728184553661, + "loss": 1.4213, + "theoretical_loss": 3.762605186819494, + "tokens_seen": 789315584 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038426278836509533, + "loss": 1.5292, + "theoretical_loss": 3.76257437915892, + "tokens_seen": 789381120 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038425275827482446, + "loss": 1.4039, + "theoretical_loss": 3.762543574772047, + "tokens_seen": 789446656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003842427281845537, + "loss": 1.4099, + "theoretical_loss": 3.7625127736582527, + "tokens_seen": 789512192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003842326980942829, + "loss": 1.4438, + "theoretical_loss": 3.7624819758169195, + "tokens_seen": 789577728 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038422266800401206, + "loss": 1.4637, + "theoretical_loss": 3.7624511812474273, + "tokens_seen": 789643264 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 292000, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6679941415786743, + "objective/train/theoretical_loss": 3.762420389949157, + "objective/train/tokens_used": 810168800, + "theoretical_loss": 3.762420389949157, + "tokens_seen": 789708800 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038421263791374124, + "loss": 1.5624, + "theoretical_loss": 3.762420389949157, + "tokens_seen": 789708800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003842026078234704, + "loss": 1.5139, + "theoretical_loss": 3.76238960192149, + "tokens_seen": 789774336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841925777331996, + "loss": 1.4547, + "theoretical_loss": 3.762358817163808, + "tokens_seen": 789839872 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038418254764292884, + "loss": 1.5227, + "theoretical_loss": 3.762328035675491, + "tokens_seen": 789905408 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038417251755265796, + "loss": 1.456, + "theoretical_loss": 3.7622972574559226, + "tokens_seen": 789970944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841624874623872, + "loss": 1.4907, + "theoretical_loss": 3.762266482504483, + "tokens_seen": 790036480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841524573721163, + "loss": 1.51, + "theoretical_loss": 3.762235710820554, + "tokens_seen": 790102016 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038414242728184556, + "loss": 1.5062, + "theoretical_loss": 3.7622049424035184, + "tokens_seen": 790167552 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038413239719157474, + "loss": 1.5578, + "theoretical_loss": 3.762174177252758, + "tokens_seen": 790233088 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841223671013039, + "loss": 1.4697, + "theoretical_loss": 3.7621434153676563, + "tokens_seen": 790298624 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841123370110331, + "loss": 1.6173, + "theoretical_loss": 3.7621126567475947, + "tokens_seen": 790364160 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003841023069207623, + "loss": 1.4967, + "theoretical_loss": 3.762081901391956, + "tokens_seen": 790429696 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038409227683049147, + "loss": 1.4436, + "theoretical_loss": 3.7620511493001243, + "tokens_seen": 790495232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003840822467402207, + "loss": 1.4949, + "theoretical_loss": 3.7620204004714815, + "tokens_seen": 790560768 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038407221664994983, + "loss": 1.4692, + "theoretical_loss": 3.761989654905411, + "tokens_seen": 790626304 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038406218655967907, + "loss": 1.5031, + "theoretical_loss": 3.7619589126012967, + "tokens_seen": 790691840 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038405215646940825, + "loss": 1.6304, + "theoretical_loss": 3.761928173558522, + "tokens_seen": 790757376 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038404212637913743, + "loss": 1.5308, + "theoretical_loss": 3.7618974377764705, + "tokens_seen": 790822912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003840320962888666, + "loss": 1.485, + "theoretical_loss": 3.7618667052545263, + "tokens_seen": 790888448 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003840220661985958, + "loss": 1.4932, + "theoretical_loss": 3.7618359759920734, + "tokens_seen": 790953984 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038401203610832497, + "loss": 1.5972, + "theoretical_loss": 3.7618052499884964, + "tokens_seen": 791019520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003840020060180542, + "loss": 1.4495, + "theoretical_loss": 3.7617745272431797, + "tokens_seen": 791085056 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038399197592778333, + "loss": 1.487, + "theoretical_loss": 3.761743807755507, + "tokens_seen": 791150592 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038398194583751257, + "loss": 1.4892, + "theoretical_loss": 3.7617130915248644, + "tokens_seen": 791216128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003839719157472417, + "loss": 1.5279, + "theoretical_loss": 3.7616823785506357, + "tokens_seen": 791281664 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038396188565697093, + "loss": 1.5986, + "theoretical_loss": 3.761651668832207, + "tokens_seen": 791347200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003839518555667001, + "loss": 1.4669, + "theoretical_loss": 3.761620962368963, + "tokens_seen": 791412736 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003839418254764293, + "loss": 1.4646, + "theoretical_loss": 3.7615902591602888, + "tokens_seen": 791478272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003839317953861585, + "loss": 1.4606, + "theoretical_loss": 3.761559559205571, + "tokens_seen": 791543808 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038392176529588765, + "loss": 1.4858, + "theoretical_loss": 3.7615288625041945, + "tokens_seen": 791609344 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038391173520561684, + "loss": 1.392, + "theoretical_loss": 3.7614981690555456, + "tokens_seen": 791674880 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038390170511534607, + "loss": 1.4851, + "theoretical_loss": 3.7614674788590103, + "tokens_seen": 791740416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003838916750250752, + "loss": 1.5743, + "theoretical_loss": 3.761436791913975, + "tokens_seen": 791805952 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038388164493480443, + "loss": 1.4901, + "theoretical_loss": 3.761406108219826, + "tokens_seen": 791871488 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003838716148445336, + "loss": 1.3892, + "theoretical_loss": 3.76137542777595, + "tokens_seen": 791937024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003838615847542628, + "loss": 1.5149, + "theoretical_loss": 3.761344750581734, + "tokens_seen": 792002560 + }, + { + "epoch": 0.24, + "learning_rate": 0.000383851554663992, + "loss": 1.399, + "theoretical_loss": 3.7613140766365643, + "tokens_seen": 792068096 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038384152457372116, + "loss": 1.4406, + "theoretical_loss": 3.7612834059398286, + "tokens_seen": 792133632 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038383149448345034, + "loss": 1.5579, + "theoretical_loss": 3.7612527384909136, + "tokens_seen": 792199168 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003838214643931796, + "loss": 1.4045, + "theoretical_loss": 3.7612220742892077, + "tokens_seen": 792264704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003838114343029087, + "loss": 1.4921, + "theoretical_loss": 3.761191413334098, + "tokens_seen": 792330240 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038380140421263794, + "loss": 1.493, + "theoretical_loss": 3.7611607556249718, + "tokens_seen": 792395776 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038379137412236706, + "loss": 1.3986, + "theoretical_loss": 3.761130101161217, + "tokens_seen": 792461312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003837813440320963, + "loss": 1.5047, + "theoretical_loss": 3.761099449942223, + "tokens_seen": 792526848 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003837713139418255, + "loss": 1.4206, + "theoretical_loss": 3.7610688019673764, + "tokens_seen": 792592384 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038376128385155466, + "loss": 1.4812, + "theoretical_loss": 3.761038157236067, + "tokens_seen": 792657920 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038375125376128384, + "loss": 1.4176, + "theoretical_loss": 3.7610075157476826, + "tokens_seen": 792723456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003837412236710131, + "loss": 1.5008, + "theoretical_loss": 3.760976877501612, + "tokens_seen": 792788992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003837311935807422, + "loss": 1.516, + "theoretical_loss": 3.7609462424972446, + "tokens_seen": 792854528 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038372116349047144, + "loss": 1.4722, + "theoretical_loss": 3.760915610733969, + "tokens_seen": 792920064 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 293128, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4166289567947388, + "objective/train/theoretical_loss": 3.7608849822111754, + "objective/train/tokens_used": 813445600, + "theoretical_loss": 3.7608849822111754, + "tokens_seen": 792985600 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038371113340020057, + "loss": 1.5292, + "theoretical_loss": 3.7608849822111754, + "tokens_seen": 792985600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003837011033099298, + "loss": 1.5002, + "theoretical_loss": 3.7608543569282515, + "tokens_seen": 793051136 + }, + { + "epoch": 0.24, + "learning_rate": 0.000383691073219659, + "loss": 1.456, + "theoretical_loss": 3.7608237348845885, + "tokens_seen": 793116672 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038368104312938816, + "loss": 1.6006, + "theoretical_loss": 3.760793116079576, + "tokens_seen": 793182208 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038367101303911735, + "loss": 1.4662, + "theoretical_loss": 3.7607625005126026, + "tokens_seen": 793247744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003836609829488465, + "loss": 1.5194, + "theoretical_loss": 3.7607318881830594, + "tokens_seen": 793313280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003836509528585757, + "loss": 1.5085, + "theoretical_loss": 3.7607012790903367, + "tokens_seen": 793378816 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038364092276830494, + "loss": 1.4549, + "theoretical_loss": 3.7606706732338253, + "tokens_seen": 793444352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003836308926780341, + "loss": 1.6047, + "theoretical_loss": 3.7606400706129146, + "tokens_seen": 793509888 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003836208625877633, + "loss": 1.5424, + "theoretical_loss": 3.760609471226996, + "tokens_seen": 793575424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003836108324974925, + "loss": 1.5953, + "theoretical_loss": 3.7605788750754607, + "tokens_seen": 793640960 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038360080240722167, + "loss": 1.5167, + "theoretical_loss": 3.7605482821577, + "tokens_seen": 793706496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003835907723169509, + "loss": 1.4868, + "theoretical_loss": 3.760517692473104, + "tokens_seen": 793772032 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038358074222668003, + "loss": 1.5312, + "theoretical_loss": 3.7604871060210643, + "tokens_seen": 793837568 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038357071213640927, + "loss": 1.5719, + "theoretical_loss": 3.7604565228009736, + "tokens_seen": 793903104 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038356068204613845, + "loss": 1.4658, + "theoretical_loss": 3.760425942812223, + "tokens_seen": 793968640 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038355065195586763, + "loss": 1.4262, + "theoretical_loss": 3.760395366054204, + "tokens_seen": 794034176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003835406218655968, + "loss": 1.5397, + "theoretical_loss": 3.7603647925263095, + "tokens_seen": 794099712 + }, + { + "epoch": 0.24, + "learning_rate": 0.000383530591775326, + "loss": 1.3422, + "theoretical_loss": 3.7603342222279315, + "tokens_seen": 794165248 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038352056168505517, + "loss": 1.5209, + "theoretical_loss": 3.760303655158461, + "tokens_seen": 794230784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003835105315947844, + "loss": 1.4426, + "theoretical_loss": 3.760273091317293, + "tokens_seen": 794296320 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038350050150451353, + "loss": 1.5273, + "theoretical_loss": 3.7602425307038185, + "tokens_seen": 794361856 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038349047141424277, + "loss": 1.4438, + "theoretical_loss": 3.760211973317431, + "tokens_seen": 794427392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003834804413239719, + "loss": 1.6253, + "theoretical_loss": 3.760181419157523, + "tokens_seen": 794492928 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038347041123370113, + "loss": 1.4098, + "theoretical_loss": 3.7601508682234885, + "tokens_seen": 794558464 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003834603811434303, + "loss": 1.5618, + "theoretical_loss": 3.7601203205147202, + "tokens_seen": 794624000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003834503510531595, + "loss": 1.466, + "theoretical_loss": 3.760089776030613, + "tokens_seen": 794689536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003834403209628887, + "loss": 1.4492, + "theoretical_loss": 3.7600592347705586, + "tokens_seen": 794755072 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038343029087261786, + "loss": 1.5077, + "theoretical_loss": 3.7600286967339525, + "tokens_seen": 794820608 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038342026078234704, + "loss": 1.5071, + "theoretical_loss": 3.759998161920187, + "tokens_seen": 794886144 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038341023069207627, + "loss": 1.502, + "theoretical_loss": 3.7599676303286587, + "tokens_seen": 794951680 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003834002006018054, + "loss": 1.5214, + "theoretical_loss": 3.75993710195876, + "tokens_seen": 795017216 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038339017051153463, + "loss": 1.4783, + "theoretical_loss": 3.759906576809886, + "tokens_seen": 795082752 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003833801404212638, + "loss": 1.4591, + "theoretical_loss": 3.759876054881432, + "tokens_seen": 795148288 + }, + { + "epoch": 0.24, + "learning_rate": 0.000383370110330993, + "loss": 1.476, + "theoretical_loss": 3.759845536172792, + "tokens_seen": 795213824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003833600802407222, + "loss": 1.6343, + "theoretical_loss": 3.7598150206833614, + "tokens_seen": 795279360 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038335005015045136, + "loss": 1.5163, + "theoretical_loss": 3.759784508412536, + "tokens_seen": 795344896 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038334002006018054, + "loss": 1.5169, + "theoretical_loss": 3.7597539993597096, + "tokens_seen": 795410432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003833299899699098, + "loss": 1.4636, + "theoretical_loss": 3.7597234935242794, + "tokens_seen": 795475968 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003833199598796389, + "loss": 1.4753, + "theoretical_loss": 3.7596929909056396, + "tokens_seen": 795541504 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038330992978936814, + "loss": 1.4179, + "theoretical_loss": 3.759662491503187, + "tokens_seen": 795607040 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038329989969909726, + "loss": 1.5258, + "theoretical_loss": 3.759631995316318, + "tokens_seen": 795672576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003832898696088265, + "loss": 1.5267, + "theoretical_loss": 3.7596015023444274, + "tokens_seen": 795738112 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003832798395185557, + "loss": 1.464, + "theoretical_loss": 3.759571012586913, + "tokens_seen": 795803648 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038326980942828486, + "loss": 1.3928, + "theoretical_loss": 3.75954052604317, + "tokens_seen": 795869184 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038325977933801404, + "loss": 1.4702, + "theoretical_loss": 3.759510042712596, + "tokens_seen": 795934720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003832497492477433, + "loss": 1.6585, + "theoretical_loss": 3.7594795625945867, + "tokens_seen": 796000256 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003832397191574724, + "loss": 1.5137, + "theoretical_loss": 3.75944908568854, + "tokens_seen": 796065792 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038322968906720164, + "loss": 1.4527, + "theoretical_loss": 3.759418611993853, + "tokens_seen": 796131328 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038321965897693077, + "loss": 1.5244, + "theoretical_loss": 3.7593881415099233, + "tokens_seen": 796196864 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 294439, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5788428783416748, + "objective/train/theoretical_loss": 3.7593576742361474, + "objective/train/tokens_used": 816722400, + "theoretical_loss": 3.7593576742361474, + "tokens_seen": 796262400 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038320962888666, + "loss": 1.5902, + "theoretical_loss": 3.7593576742361474, + "tokens_seen": 796262400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003831995987963892, + "loss": 1.4593, + "theoretical_loss": 3.759327210171924, + "tokens_seen": 796327936 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038318956870611836, + "loss": 1.5389, + "theoretical_loss": 3.7592967493166496, + "tokens_seen": 796393472 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038317953861584755, + "loss": 1.5701, + "theoretical_loss": 3.7592662916697237, + "tokens_seen": 796459008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003831695085255767, + "loss": 1.5117, + "theoretical_loss": 3.759235837230543, + "tokens_seen": 796524544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003831594784353059, + "loss": 1.5018, + "theoretical_loss": 3.7592053859985066, + "tokens_seen": 796590080 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038314944834503514, + "loss": 1.49, + "theoretical_loss": 3.759174937973013, + "tokens_seen": 796655616 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038313941825476427, + "loss": 1.3653, + "theoretical_loss": 3.7591444931534603, + "tokens_seen": 796721152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003831293881644935, + "loss": 1.4136, + "theoretical_loss": 3.759114051539248, + "tokens_seen": 796786688 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038311935807422263, + "loss": 1.4655, + "theoretical_loss": 3.759083613129774, + "tokens_seen": 796852224 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038310932798395187, + "loss": 1.5109, + "theoretical_loss": 3.7590531779244385, + "tokens_seen": 796917760 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038309929789368105, + "loss": 1.4497, + "theoretical_loss": 3.75902274592264, + "tokens_seen": 796983296 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038308926780341023, + "loss": 1.3665, + "theoretical_loss": 3.758992317123778, + "tokens_seen": 797048832 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003830792377131394, + "loss": 1.4448, + "theoretical_loss": 3.7589618915272522, + "tokens_seen": 797114368 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038306920762286865, + "loss": 1.4167, + "theoretical_loss": 3.7589314691324622, + "tokens_seen": 797179904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003830591775325978, + "loss": 1.4594, + "theoretical_loss": 3.7589010499388085, + "tokens_seen": 797245440 + }, + { + "epoch": 0.24, + "learning_rate": 0.000383049147442327, + "loss": 1.4284, + "theoretical_loss": 3.7588706339456905, + "tokens_seen": 797310976 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038303911735205614, + "loss": 1.4781, + "theoretical_loss": 3.7588402211525085, + "tokens_seen": 797376512 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038302908726178537, + "loss": 1.4588, + "theoretical_loss": 3.7588098115586632, + "tokens_seen": 797442048 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038301905717151455, + "loss": 1.5557, + "theoretical_loss": 3.758779405163555, + "tokens_seen": 797507584 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038300902708124373, + "loss": 1.5673, + "theoretical_loss": 3.7587490019665846, + "tokens_seen": 797573120 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829989969909729, + "loss": 1.5825, + "theoretical_loss": 3.7587186019671526, + "tokens_seen": 797638656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829889669007021, + "loss": 1.5021, + "theoretical_loss": 3.7586882051646606, + "tokens_seen": 797704192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829789368104313, + "loss": 1.5646, + "theoretical_loss": 3.7586578115585096, + "tokens_seen": 797769728 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829689067201605, + "loss": 1.4097, + "theoretical_loss": 3.758627421148101, + "tokens_seen": 797835264 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038295887662988964, + "loss": 1.4436, + "theoretical_loss": 3.758597033932836, + "tokens_seen": 797900800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829488465396189, + "loss": 1.5409, + "theoretical_loss": 3.758566649912116, + "tokens_seen": 797966336 + }, + { + "epoch": 0.24, + "learning_rate": 0.000382938816449348, + "loss": 1.4565, + "theoretical_loss": 3.7585362690853445, + "tokens_seen": 798031872 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038292878635907724, + "loss": 1.6316, + "theoretical_loss": 3.758505891451921, + "tokens_seen": 798097408 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829187562688064, + "loss": 1.4945, + "theoretical_loss": 3.7584755170112496, + "tokens_seen": 798162944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003829087261785356, + "loss": 1.5436, + "theoretical_loss": 3.758445145762732, + "tokens_seen": 798228480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003828986960882648, + "loss": 1.4941, + "theoretical_loss": 3.758414777705771, + "tokens_seen": 798294016 + }, + { + "epoch": 0.24, + "learning_rate": 0.000382888665997994, + "loss": 1.4505, + "theoretical_loss": 3.7583844128397685, + "tokens_seen": 798359552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003828786359077232, + "loss": 1.5137, + "theoretical_loss": 3.758354051164128, + "tokens_seen": 798425088 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003828686058174524, + "loss": 1.4305, + "theoretical_loss": 3.7583236926782515, + "tokens_seen": 798490624 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038285857572718156, + "loss": 1.545, + "theoretical_loss": 3.7582933373815433, + "tokens_seen": 798556160 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038284854563691074, + "loss": 1.4331, + "theoretical_loss": 3.758262985273406, + "tokens_seen": 798621696 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038283851554664, + "loss": 1.4083, + "theoretical_loss": 3.7582326363532443, + "tokens_seen": 798687232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003828284854563691, + "loss": 1.5482, + "theoretical_loss": 3.75820229062046, + "tokens_seen": 798752768 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038281845536609834, + "loss": 1.5527, + "theoretical_loss": 3.758171948074457, + "tokens_seen": 798818304 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038280842527582746, + "loss": 1.4575, + "theoretical_loss": 3.7581416087146406, + "tokens_seen": 798883840 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827983951855567, + "loss": 1.4412, + "theoretical_loss": 3.7581112725404147, + "tokens_seen": 798949376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827883650952859, + "loss": 1.4761, + "theoretical_loss": 3.7580809395511823, + "tokens_seen": 799014912 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038277833500501506, + "loss": 1.5309, + "theoretical_loss": 3.7580506097463484, + "tokens_seen": 799080448 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038276830491474424, + "loss": 1.4506, + "theoretical_loss": 3.758020283125318, + "tokens_seen": 799145984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827582748244735, + "loss": 1.494, + "theoretical_loss": 3.757989959687496, + "tokens_seen": 799211520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827482447342026, + "loss": 1.5642, + "theoretical_loss": 3.7579596394322863, + "tokens_seen": 799277056 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038273821464393184, + "loss": 1.4815, + "theoretical_loss": 3.757929322359095, + "tokens_seen": 799342592 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038272818455366097, + "loss": 1.5128, + "theoretical_loss": 3.7578990084673256, + "tokens_seen": 799408128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827181544633902, + "loss": 1.5323, + "theoretical_loss": 3.7578686977563858, + "tokens_seen": 799473664 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 295724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3859856128692627, + "objective/train/theoretical_loss": 3.75783839022568, + "objective/train/tokens_used": 819999200, + "theoretical_loss": 3.75783839022568, + "tokens_seen": 799539200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003827081243731194, + "loss": 1.4931, + "theoretical_loss": 3.75783839022568, + "tokens_seen": 799539200 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038269809428284856, + "loss": 1.5384, + "theoretical_loss": 3.757808085874613, + "tokens_seen": 799604736 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038268806419257775, + "loss": 1.4731, + "theoretical_loss": 3.7577777847025926, + "tokens_seen": 799670272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003826780341023069, + "loss": 1.4738, + "theoretical_loss": 3.757747486709023, + "tokens_seen": 799735808 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003826680040120361, + "loss": 1.4109, + "theoretical_loss": 3.7577171918933114, + "tokens_seen": 799801344 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038265797392176534, + "loss": 1.4393, + "theoretical_loss": 3.757686900254864, + "tokens_seen": 799866880 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038264794383149447, + "loss": 1.5982, + "theoretical_loss": 3.757656611793087, + "tokens_seen": 799932416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003826379137412237, + "loss": 1.5004, + "theoretical_loss": 3.7576263265073875, + "tokens_seen": 799997952 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038262788365095283, + "loss": 1.4033, + "theoretical_loss": 3.7575960443971717, + "tokens_seen": 800063488 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038261785356068207, + "loss": 1.4948, + "theoretical_loss": 3.7575657654618473, + "tokens_seen": 800129024 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038260782347041125, + "loss": 1.4281, + "theoretical_loss": 3.75753548970082, + "tokens_seen": 800194560 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038259779338014043, + "loss": 1.5016, + "theoretical_loss": 3.757505217113499, + "tokens_seen": 800260096 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003825877632898696, + "loss": 1.4067, + "theoretical_loss": 3.757474947699291, + "tokens_seen": 800325632 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038257773319959885, + "loss": 1.4903, + "theoretical_loss": 3.757444681457603, + "tokens_seen": 800391168 + }, + { + "epoch": 0.24, + "learning_rate": 0.000382567703109328, + "loss": 1.4647, + "theoretical_loss": 3.7574144183878433, + "tokens_seen": 800456704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003825576730190572, + "loss": 1.4583, + "theoretical_loss": 3.75738415848942, + "tokens_seen": 800522240 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038254764292878634, + "loss": 1.4221, + "theoretical_loss": 3.7573539017617406, + "tokens_seen": 800587776 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038253761283851557, + "loss": 1.4947, + "theoretical_loss": 3.7573236482042134, + "tokens_seen": 800653312 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038252758274824475, + "loss": 1.4719, + "theoretical_loss": 3.7572933978162477, + "tokens_seen": 800718848 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038251755265797393, + "loss": 1.4775, + "theoretical_loss": 3.757263150597251, + "tokens_seen": 800784384 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003825075225677031, + "loss": 1.5609, + "theoretical_loss": 3.7572329065466326, + "tokens_seen": 800849920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824974924774323, + "loss": 1.4255, + "theoretical_loss": 3.7572026656638013, + "tokens_seen": 800915456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824874623871615, + "loss": 1.4164, + "theoretical_loss": 3.757172427948166, + "tokens_seen": 800980992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824774322968907, + "loss": 1.5129, + "theoretical_loss": 3.7571421933991362, + "tokens_seen": 801046528 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038246740220661984, + "loss": 1.4802, + "theoretical_loss": 3.7571119620161206, + "tokens_seen": 801112064 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824573721163491, + "loss": 1.5198, + "theoretical_loss": 3.7570817337985294, + "tokens_seen": 801177600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824473420260782, + "loss": 1.4934, + "theoretical_loss": 3.757051508745772, + "tokens_seen": 801243136 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038243731193580744, + "loss": 1.5465, + "theoretical_loss": 3.757021286857258, + "tokens_seen": 801308672 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824272818455366, + "loss": 1.6546, + "theoretical_loss": 3.7569910681323977, + "tokens_seen": 801374208 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003824172517552658, + "loss": 1.5222, + "theoretical_loss": 3.756960852570601, + "tokens_seen": 801439744 + }, + { + "epoch": 0.24, + "learning_rate": 0.000382407221664995, + "loss": 1.5491, + "theoretical_loss": 3.7569306401712783, + "tokens_seen": 801505280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823971915747242, + "loss": 1.5845, + "theoretical_loss": 3.7569004309338405, + "tokens_seen": 801570816 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038238716148445334, + "loss": 1.5743, + "theoretical_loss": 3.756870224857698, + "tokens_seen": 801636352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823771313941826, + "loss": 1.5406, + "theoretical_loss": 3.756840021942261, + "tokens_seen": 801701888 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823671013039117, + "loss": 1.4775, + "theoretical_loss": 3.756809822186941, + "tokens_seen": 801767424 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038235707121364094, + "loss": 1.5298, + "theoretical_loss": 3.7567796255911485, + "tokens_seen": 801832960 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823470411233701, + "loss": 1.5115, + "theoretical_loss": 3.7567494321542956, + "tokens_seen": 801898496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823370110330993, + "loss": 1.5166, + "theoretical_loss": 3.756719241875793, + "tokens_seen": 801964032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003823269809428285, + "loss": 1.5345, + "theoretical_loss": 3.7566890547550527, + "tokens_seen": 802029568 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038231695085255766, + "loss": 1.5591, + "theoretical_loss": 3.7566588707914867, + "tokens_seen": 802095104 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038230692076228685, + "loss": 1.4191, + "theoretical_loss": 3.756628689984506, + "tokens_seen": 802160640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003822968906720161, + "loss": 1.3696, + "theoretical_loss": 3.756598512333523, + "tokens_seen": 802226176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003822868605817452, + "loss": 1.4796, + "theoretical_loss": 3.75656833783795, + "tokens_seen": 802291712 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038227683049147444, + "loss": 1.4368, + "theoretical_loss": 3.7565381664972, + "tokens_seen": 802357248 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038226680040120357, + "loss": 1.4759, + "theoretical_loss": 3.7565079983106844, + "tokens_seen": 802422784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003822567703109328, + "loss": 1.3794, + "theoretical_loss": 3.7564778332778164, + "tokens_seen": 802488320 + }, + { + "epoch": 0.24, + "learning_rate": 0.000382246740220662, + "loss": 1.4342, + "theoretical_loss": 3.756447671398009, + "tokens_seen": 802553856 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038223671013039117, + "loss": 1.4378, + "theoretical_loss": 3.7564175126706743, + "tokens_seen": 802619392 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038222668004012035, + "loss": 1.458, + "theoretical_loss": 3.7563873570952264, + "tokens_seen": 802684928 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003822166499498496, + "loss": 1.3963, + "theoretical_loss": 3.7563572046710787, + "tokens_seen": 802750464 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 296852, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.45094633102417, + "objective/train/theoretical_loss": 3.7563270553976436, + "objective/train/tokens_used": 823276000, + "theoretical_loss": 3.7563270553976436, + "tokens_seen": 802816000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003822066198595787, + "loss": 1.3961, + "theoretical_loss": 3.7563270553976436, + "tokens_seen": 802816000 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038219658976930795, + "loss": 1.3975, + "theoretical_loss": 3.7562969092743357, + "tokens_seen": 802881536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003821865596790371, + "loss": 1.4247, + "theoretical_loss": 3.7562667663005684, + "tokens_seen": 802947072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003821765295887663, + "loss": 1.5157, + "theoretical_loss": 3.7562366264757556, + "tokens_seen": 803012608 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003821664994984955, + "loss": 1.5124, + "theoretical_loss": 3.756206489799311, + "tokens_seen": 803078144 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038215646940822467, + "loss": 1.3991, + "theoretical_loss": 3.7561763562706503, + "tokens_seen": 803143680 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038214643931795385, + "loss": 1.633, + "theoretical_loss": 3.7561462258891862, + "tokens_seen": 803209216 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038213640922768303, + "loss": 1.5096, + "theoretical_loss": 3.756116098654334, + "tokens_seen": 803274752 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038212637913741227, + "loss": 1.5303, + "theoretical_loss": 3.7560859745655084, + "tokens_seen": 803340288 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038211634904714145, + "loss": 1.494, + "theoretical_loss": 3.756055853622124, + "tokens_seen": 803405824 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038210631895687063, + "loss": 1.4829, + "theoretical_loss": 3.756025735823596, + "tokens_seen": 803471360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003820962888665998, + "loss": 1.4227, + "theoretical_loss": 3.7559956211693404, + "tokens_seen": 803536896 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038208625877632905, + "loss": 1.5481, + "theoretical_loss": 3.7559655096587705, + "tokens_seen": 803602432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003820762286860582, + "loss": 1.4755, + "theoretical_loss": 3.7559354012913038, + "tokens_seen": 803667968 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003820661985957874, + "loss": 1.469, + "theoretical_loss": 3.755905296066355, + "tokens_seen": 803733504 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038205616850551654, + "loss": 1.516, + "theoretical_loss": 3.75587519398334, + "tokens_seen": 803799040 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038204613841524577, + "loss": 1.5144, + "theoretical_loss": 3.755845095041675, + "tokens_seen": 803864576 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038203610832497495, + "loss": 1.5777, + "theoretical_loss": 3.755814999240776, + "tokens_seen": 803930112 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038202607823470413, + "loss": 1.6038, + "theoretical_loss": 3.7557849065800593, + "tokens_seen": 803995648 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003820160481444333, + "loss": 1.5794, + "theoretical_loss": 3.755754817058941, + "tokens_seen": 804061184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003820060180541625, + "loss": 1.5838, + "theoretical_loss": 3.755724730676838, + "tokens_seen": 804126720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819959879638917, + "loss": 1.4484, + "theoretical_loss": 3.7556946474331667, + "tokens_seen": 804192256 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819859578736209, + "loss": 1.4599, + "theoretical_loss": 3.7556645673273445, + "tokens_seen": 804257792 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038197592778335004, + "loss": 1.561, + "theoretical_loss": 3.755634490358789, + "tokens_seen": 804323328 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819658976930793, + "loss": 1.5387, + "theoretical_loss": 3.7556044165269156, + "tokens_seen": 804388864 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819558676028084, + "loss": 1.4967, + "theoretical_loss": 3.7555743458311435, + "tokens_seen": 804454400 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038194583751253764, + "loss": 1.6003, + "theoretical_loss": 3.7555442782708885, + "tokens_seen": 804519936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819358074222668, + "loss": 1.4282, + "theoretical_loss": 3.75551421384557, + "tokens_seen": 804585472 + }, + { + "epoch": 0.24, + "learning_rate": 0.000381925777331996, + "loss": 1.3549, + "theoretical_loss": 3.755484152554605, + "tokens_seen": 804651008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819157472417252, + "loss": 1.4015, + "theoretical_loss": 3.755454094397411, + "tokens_seen": 804716544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003819057171514544, + "loss": 1.4676, + "theoretical_loss": 3.7554240393734073, + "tokens_seen": 804782080 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038189568706118354, + "loss": 1.501, + "theoretical_loss": 3.755393987482011, + "tokens_seen": 804847616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818856569709128, + "loss": 1.4992, + "theoretical_loss": 3.755363938722642, + "tokens_seen": 804913152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818756268806419, + "loss": 1.4892, + "theoretical_loss": 3.7553338930947167, + "tokens_seen": 804978688 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038186559679037114, + "loss": 1.448, + "theoretical_loss": 3.7553038505976564, + "tokens_seen": 805044224 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818555667001003, + "loss": 1.4001, + "theoretical_loss": 3.7552738112308783, + "tokens_seen": 805109760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818455366098295, + "loss": 1.5495, + "theoretical_loss": 3.7552437749938017, + "tokens_seen": 805175296 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818355065195587, + "loss": 1.5738, + "theoretical_loss": 3.755213741885847, + "tokens_seen": 805240832 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038182547642928786, + "loss": 1.5431, + "theoretical_loss": 3.755183711906432, + "tokens_seen": 805306368 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038181544633901705, + "loss": 1.3873, + "theoretical_loss": 3.755153685054977, + "tokens_seen": 805371904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003818054162487463, + "loss": 1.4982, + "theoretical_loss": 3.755123661330902, + "tokens_seen": 805437440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003817953861584754, + "loss": 1.5713, + "theoretical_loss": 3.755093640733626, + "tokens_seen": 805502976 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038178535606820464, + "loss": 1.4844, + "theoretical_loss": 3.75506362326257, + "tokens_seen": 805568512 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038177532597793377, + "loss": 1.5856, + "theoretical_loss": 3.755033608917153, + "tokens_seen": 805634048 + }, + { + "epoch": 0.24, + "learning_rate": 0.000381765295887663, + "loss": 1.4575, + "theoretical_loss": 3.7550035976967964, + "tokens_seen": 805699584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003817552657973922, + "loss": 1.4419, + "theoretical_loss": 3.7549735896009206, + "tokens_seen": 805765120 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038174523570712137, + "loss": 1.5256, + "theoretical_loss": 3.754943584628945, + "tokens_seen": 805830656 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038173520561685055, + "loss": 1.5666, + "theoretical_loss": 3.7549135827802917, + "tokens_seen": 805896192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003817251755265798, + "loss": 1.4458, + "theoretical_loss": 3.7548835840543813, + "tokens_seen": 805961728 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003817151454363089, + "loss": 1.391, + "theoretical_loss": 3.7548535884506347, + "tokens_seen": 806027264 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 297892, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.498117446899414, + "objective/train/theoretical_loss": 3.754823595968473, + "objective/train/tokens_used": 826552800, + "theoretical_loss": 3.754823595968473, + "tokens_seen": 806092800 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038170511534603815, + "loss": 1.4107, + "theoretical_loss": 3.754823595968473, + "tokens_seen": 806092800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003816950852557673, + "loss": 1.5195, + "theoretical_loss": 3.7547936066073175, + "tokens_seen": 806158336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003816850551654965, + "loss": 1.5027, + "theoretical_loss": 3.7547636203665906, + "tokens_seen": 806223872 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003816750250752257, + "loss": 1.4243, + "theoretical_loss": 3.7547336372457134, + "tokens_seen": 806289408 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038166499498495487, + "loss": 1.4259, + "theoretical_loss": 3.7547036572441073, + "tokens_seen": 806354944 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038165496489468405, + "loss": 1.6067, + "theoretical_loss": 3.7546736803611953, + "tokens_seen": 806420480 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038164493480441323, + "loss": 1.6174, + "theoretical_loss": 3.754643706596399, + "tokens_seen": 806486016 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003816349047141424, + "loss": 1.5087, + "theoretical_loss": 3.7546137359491407, + "tokens_seen": 806551552 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038162487462387165, + "loss": 1.3621, + "theoretical_loss": 3.7545837684188426, + "tokens_seen": 806617088 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003816148445336008, + "loss": 1.3778, + "theoretical_loss": 3.7545538040049284, + "tokens_seen": 806682624 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038160481444333, + "loss": 1.4662, + "theoretical_loss": 3.7545238427068197, + "tokens_seen": 806748160 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038159478435305914, + "loss": 1.4043, + "theoretical_loss": 3.7544938845239404, + "tokens_seen": 806813696 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003815847542627884, + "loss": 1.4824, + "theoretical_loss": 3.7544639294557127, + "tokens_seen": 806879232 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038157472417251756, + "loss": 1.4897, + "theoretical_loss": 3.7544339775015603, + "tokens_seen": 806944768 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038156469408224674, + "loss": 1.5081, + "theoretical_loss": 3.7544040286609066, + "tokens_seen": 807010304 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003815546639919759, + "loss": 1.3626, + "theoretical_loss": 3.7543740829331744, + "tokens_seen": 807075840 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038154463390170515, + "loss": 1.5171, + "theoretical_loss": 3.754344140317789, + "tokens_seen": 807141376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003815346038114343, + "loss": 1.4336, + "theoretical_loss": 3.754314200814173, + "tokens_seen": 807206912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003815245737211635, + "loss": 1.4549, + "theoretical_loss": 3.754284264421751, + "tokens_seen": 807272448 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038151454363089264, + "loss": 1.4713, + "theoretical_loss": 3.754254331139946, + "tokens_seen": 807337984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003815045135406219, + "loss": 1.3898, + "theoretical_loss": 3.7542244009681847, + "tokens_seen": 807403520 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038149448345035106, + "loss": 1.4643, + "theoretical_loss": 3.7541944739058892, + "tokens_seen": 807469056 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038148445336008024, + "loss": 1.4402, + "theoretical_loss": 3.754164549952485, + "tokens_seen": 807534592 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003814744232698094, + "loss": 1.3972, + "theoretical_loss": 3.754134629107397, + "tokens_seen": 807600128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003814643931795386, + "loss": 1.4242, + "theoretical_loss": 3.7541047113700503, + "tokens_seen": 807665664 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003814543630892678, + "loss": 1.5356, + "theoretical_loss": 3.7540747967398698, + "tokens_seen": 807731200 + }, + { + "epoch": 0.24, + "learning_rate": 0.000381444332998997, + "loss": 1.3864, + "theoretical_loss": 3.75404488521628, + "tokens_seen": 807796736 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038143430290872614, + "loss": 1.4324, + "theoretical_loss": 3.7540149767987074, + "tokens_seen": 807862272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003814242728184554, + "loss": 1.4646, + "theoretical_loss": 3.753985071486577, + "tokens_seen": 807927808 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038141424272818456, + "loss": 1.4598, + "theoretical_loss": 3.7539551692793145, + "tokens_seen": 807993344 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038140421263791374, + "loss": 1.5422, + "theoretical_loss": 3.753925270176346, + "tokens_seen": 808058880 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003813941825476429, + "loss": 1.5045, + "theoretical_loss": 3.7538953741770973, + "tokens_seen": 808124416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003813841524573721, + "loss": 1.5428, + "theoretical_loss": 3.753865481280995, + "tokens_seen": 808189952 + }, + { + "epoch": 0.24, + "learning_rate": 0.00038137412236710134, + "loss": 1.4254, + "theoretical_loss": 3.753835591487464, + "tokens_seen": 808255488 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003813640922768305, + "loss": 1.4615, + "theoretical_loss": 3.7538057047959326, + "tokens_seen": 808321024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003813540621865597, + "loss": 1.5395, + "theoretical_loss": 3.753775821205826, + "tokens_seen": 808386560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0003813440320962889, + "loss": 1.4802, + "theoretical_loss": 3.7537459407165716, + "tokens_seen": 808452096 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038133400200601806, + "loss": 1.4116, + "theoretical_loss": 3.7537160633275963, + "tokens_seen": 808517632 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038132397191574725, + "loss": 1.5268, + "theoretical_loss": 3.7536861890383273, + "tokens_seen": 808583168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003813139418254765, + "loss": 1.572, + "theoretical_loss": 3.7536563178481916, + "tokens_seen": 808648704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003813039117352056, + "loss": 1.4307, + "theoretical_loss": 3.753626449756616, + "tokens_seen": 808714240 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038129388164493484, + "loss": 1.4741, + "theoretical_loss": 3.753596584763029, + "tokens_seen": 808779776 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038128385155466397, + "loss": 1.4272, + "theoretical_loss": 3.753566722866858, + "tokens_seen": 808845312 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003812738214643932, + "loss": 1.5589, + "theoretical_loss": 3.753536864067531, + "tokens_seen": 808910848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003812637913741224, + "loss": 1.6263, + "theoretical_loss": 3.7535070083644753, + "tokens_seen": 808976384 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038125376128385157, + "loss": 1.4584, + "theoretical_loss": 3.7534771557571194, + "tokens_seen": 809041920 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038124373119358075, + "loss": 1.4101, + "theoretical_loss": 3.7534473062448916, + "tokens_seen": 809107456 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038123370110331, + "loss": 1.4628, + "theoretical_loss": 3.7534174598272205, + "tokens_seen": 809172992 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003812236710130391, + "loss": 1.4922, + "theoretical_loss": 3.7533876165035345, + "tokens_seen": 809238528 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038121364092276835, + "loss": 1.5319, + "theoretical_loss": 3.7533577762732624, + "tokens_seen": 809304064 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 298985, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5224976539611816, + "objective/train/theoretical_loss": 3.753327939135833, + "objective/train/tokens_used": 829829600, + "theoretical_loss": 3.753327939135833, + "tokens_seen": 809369600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003812036108324975, + "loss": 1.5484, + "theoretical_loss": 3.753327939135833, + "tokens_seen": 809369600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003811935807422267, + "loss": 1.5375, + "theoretical_loss": 3.7532981050906757, + "tokens_seen": 809435136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003811835506519559, + "loss": 1.5251, + "theoretical_loss": 3.753268274137219, + "tokens_seen": 809500672 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038117352056168507, + "loss": 1.4652, + "theoretical_loss": 3.753238446274893, + "tokens_seen": 809566208 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038116349047141425, + "loss": 1.4331, + "theoretical_loss": 3.7532086215031266, + "tokens_seen": 809631744 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038115346038114343, + "loss": 1.5322, + "theoretical_loss": 3.7531787998213497, + "tokens_seen": 809697280 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003811434302908726, + "loss": 1.4707, + "theoretical_loss": 3.7531489812289927, + "tokens_seen": 809762816 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038113340020060185, + "loss": 1.4028, + "theoretical_loss": 3.753119165725484, + "tokens_seen": 809828352 + }, + { + "epoch": 0.25, + "learning_rate": 0.000381123370110331, + "loss": 1.4166, + "theoretical_loss": 3.753089353310255, + "tokens_seen": 809893888 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003811133400200602, + "loss": 1.4169, + "theoretical_loss": 3.7530595439827357, + "tokens_seen": 809959424 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038110330992978934, + "loss": 1.4417, + "theoretical_loss": 3.7530297377423567, + "tokens_seen": 810024960 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003810932798395186, + "loss": 1.3961, + "theoretical_loss": 3.752999934588548, + "tokens_seen": 810090496 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038108324974924776, + "loss": 1.4169, + "theoretical_loss": 3.7529701345207402, + "tokens_seen": 810156032 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038107321965897694, + "loss": 1.5265, + "theoretical_loss": 3.752940337538365, + "tokens_seen": 810221568 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003810631895687061, + "loss": 1.416, + "theoretical_loss": 3.752910543640853, + "tokens_seen": 810287104 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038105315947843535, + "loss": 1.3929, + "theoretical_loss": 3.752880752827635, + "tokens_seen": 810352640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003810431293881645, + "loss": 1.5075, + "theoretical_loss": 3.752850965098143, + "tokens_seen": 810418176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003810330992978937, + "loss": 1.4406, + "theoretical_loss": 3.752821180451808, + "tokens_seen": 810483712 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038102306920762284, + "loss": 1.4748, + "theoretical_loss": 3.7527913988880615, + "tokens_seen": 810549248 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003810130391173521, + "loss": 1.5781, + "theoretical_loss": 3.7527616204063357, + "tokens_seen": 810614784 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038100300902708126, + "loss": 1.444, + "theoretical_loss": 3.7527318450060623, + "tokens_seen": 810680320 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038099297893681044, + "loss": 1.5998, + "theoretical_loss": 3.752702072686673, + "tokens_seen": 810745856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003809829488465396, + "loss": 1.4031, + "theoretical_loss": 3.7526723034476004, + "tokens_seen": 810811392 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003809729187562688, + "loss": 1.5111, + "theoretical_loss": 3.752642537288277, + "tokens_seen": 810876928 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380962888665998, + "loss": 1.5131, + "theoretical_loss": 3.752612774208135, + "tokens_seen": 810942464 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003809528585757272, + "loss": 1.4154, + "theoretical_loss": 3.752583014206608, + "tokens_seen": 811008000 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038094282848545635, + "loss": 1.5221, + "theoretical_loss": 3.752553257283127, + "tokens_seen": 811073536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003809327983951856, + "loss": 1.5318, + "theoretical_loss": 3.7525235034371267, + "tokens_seen": 811139072 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038092276830491476, + "loss": 1.4696, + "theoretical_loss": 3.7524937526680393, + "tokens_seen": 811204608 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038091273821464394, + "loss": 1.4386, + "theoretical_loss": 3.752464004975298, + "tokens_seen": 811270144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003809027081243731, + "loss": 1.4747, + "theoretical_loss": 3.7524342603583367, + "tokens_seen": 811335680 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808926780341023, + "loss": 1.3702, + "theoretical_loss": 3.752404518816588, + "tokens_seen": 811401216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808826479438315, + "loss": 1.4783, + "theoretical_loss": 3.752374780349488, + "tokens_seen": 811466752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808726178535607, + "loss": 1.4526, + "theoretical_loss": 3.752345044956468, + "tokens_seen": 811532288 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038086258776328985, + "loss": 1.3757, + "theoretical_loss": 3.7523153126369633, + "tokens_seen": 811597824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808525576730191, + "loss": 1.4845, + "theoretical_loss": 3.7522855833904076, + "tokens_seen": 811663360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808425275827482, + "loss": 1.517, + "theoretical_loss": 3.7522558572162352, + "tokens_seen": 811728896 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038083249749247745, + "loss": 1.4546, + "theoretical_loss": 3.752226134113881, + "tokens_seen": 811794432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808224674022066, + "loss": 1.48, + "theoretical_loss": 3.7521964140827793, + "tokens_seen": 811859968 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003808124373119358, + "loss": 1.4138, + "theoretical_loss": 3.7521666971223646, + "tokens_seen": 811925504 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380802407221665, + "loss": 1.4551, + "theoretical_loss": 3.752136983232073, + "tokens_seen": 811991040 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038079237713139417, + "loss": 1.5455, + "theoretical_loss": 3.7521072724113376, + "tokens_seen": 812056576 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038078234704112335, + "loss": 1.4911, + "theoretical_loss": 3.7520775646595954, + "tokens_seen": 812122112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003807723169508526, + "loss": 1.4025, + "theoretical_loss": 3.752047859976281, + "tokens_seen": 812187648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003807622868605817, + "loss": 1.4291, + "theoretical_loss": 3.75201815836083, + "tokens_seen": 812253184 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038075225677031095, + "loss": 1.4401, + "theoretical_loss": 3.751988459812678, + "tokens_seen": 812318720 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038074222668004013, + "loss": 1.5052, + "theoretical_loss": 3.7519587643312606, + "tokens_seen": 812384256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003807321965897693, + "loss": 1.5394, + "theoretical_loss": 3.751929071916015, + "tokens_seen": 812449792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003807221664994985, + "loss": 1.4674, + "theoretical_loss": 3.751899382566376, + "tokens_seen": 812515328 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003807121364092277, + "loss": 1.3944, + "theoretical_loss": 3.7518696962817795, + "tokens_seen": 812580864 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 300102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.428229570388794, + "objective/train/theoretical_loss": 3.7518400130616634, + "objective/train/tokens_used": 833106400, + "theoretical_loss": 3.7518400130616634, + "tokens_seen": 812646400 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038070210631895685, + "loss": 1.4731, + "theoretical_loss": 3.7518400130616634, + "tokens_seen": 812646400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806920762286861, + "loss": 1.5767, + "theoretical_loss": 3.751810332905463, + "tokens_seen": 812711936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806820461384152, + "loss": 1.4479, + "theoretical_loss": 3.751780655812616, + "tokens_seen": 812777472 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038067201604814445, + "loss": 1.4903, + "theoretical_loss": 3.7517509817825583, + "tokens_seen": 812843008 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806619859578736, + "loss": 1.4619, + "theoretical_loss": 3.7517213108147276, + "tokens_seen": 812908544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806519558676028, + "loss": 1.4758, + "theoretical_loss": 3.7516916429085603, + "tokens_seen": 812974080 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380641925777332, + "loss": 1.5157, + "theoretical_loss": 3.751661978063495, + "tokens_seen": 813039616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806318956870612, + "loss": 1.3697, + "theoretical_loss": 3.751632316278968, + "tokens_seen": 813105152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806218655967904, + "loss": 1.4867, + "theoretical_loss": 3.751602657554418, + "tokens_seen": 813170688 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038061183550651954, + "loss": 1.3225, + "theoretical_loss": 3.7515730018892812, + "tokens_seen": 813236224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003806018054162488, + "loss": 1.408, + "theoretical_loss": 3.7515433492829966, + "tokens_seen": 813301760 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038059177532597796, + "loss": 1.4859, + "theoretical_loss": 3.751513699735002, + "tokens_seen": 813367296 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038058174523570714, + "loss": 1.3603, + "theoretical_loss": 3.7514840532447353, + "tokens_seen": 813432832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003805717151454363, + "loss": 1.4477, + "theoretical_loss": 3.7514544098116356, + "tokens_seen": 813498368 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038056168505516555, + "loss": 1.545, + "theoretical_loss": 3.7514247694351406, + "tokens_seen": 813563904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003805516549648947, + "loss": 1.4861, + "theoretical_loss": 3.7513951321146894, + "tokens_seen": 813629440 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003805416248746239, + "loss": 1.4913, + "theoretical_loss": 3.7513654978497213, + "tokens_seen": 813694976 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038053159478435304, + "loss": 1.407, + "theoretical_loss": 3.751335866639674, + "tokens_seen": 813760512 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003805215646940823, + "loss": 1.4638, + "theoretical_loss": 3.751306238483987, + "tokens_seen": 813826048 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038051153460381146, + "loss": 1.5228, + "theoretical_loss": 3.7512766133820996, + "tokens_seen": 813891584 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038050150451354064, + "loss": 1.4307, + "theoretical_loss": 3.751246991333452, + "tokens_seen": 813957120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804914744232698, + "loss": 1.4564, + "theoretical_loss": 3.751217372337482, + "tokens_seen": 814022656 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380481444332999, + "loss": 1.5141, + "theoretical_loss": 3.751187756393631, + "tokens_seen": 814088192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804714142427282, + "loss": 1.4807, + "theoretical_loss": 3.751158143501338, + "tokens_seen": 814153728 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804613841524574, + "loss": 1.4629, + "theoretical_loss": 3.7511285336600433, + "tokens_seen": 814219264 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038045135406218655, + "loss": 1.4176, + "theoretical_loss": 3.7510989268691866, + "tokens_seen": 814284800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804413239719158, + "loss": 1.6029, + "theoretical_loss": 3.751069323128209, + "tokens_seen": 814350336 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038043129388164496, + "loss": 1.4972, + "theoretical_loss": 3.7510397224365493, + "tokens_seen": 814415872 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038042126379137414, + "loss": 1.5468, + "theoretical_loss": 3.7510101247936496, + "tokens_seen": 814481408 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804112337011033, + "loss": 1.5172, + "theoretical_loss": 3.75098053019895, + "tokens_seen": 814546944 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003804012036108325, + "loss": 1.4561, + "theoretical_loss": 3.7509509386518918, + "tokens_seen": 814612480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003803911735205617, + "loss": 1.3627, + "theoretical_loss": 3.7509213501519154, + "tokens_seen": 814678016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003803811434302909, + "loss": 1.4275, + "theoretical_loss": 3.750891764698462, + "tokens_seen": 814743552 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038037111334002005, + "loss": 1.4492, + "theoretical_loss": 3.750862182290973, + "tokens_seen": 814809088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003803610832497493, + "loss": 1.3523, + "theoretical_loss": 3.7508326029288908, + "tokens_seen": 814874624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003803510531594784, + "loss": 1.5285, + "theoretical_loss": 3.750803026611656, + "tokens_seen": 814940160 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038034102306920765, + "loss": 1.4532, + "theoretical_loss": 3.7507734533387103, + "tokens_seen": 815005696 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038033099297893683, + "loss": 1.5399, + "theoretical_loss": 3.7507438831094957, + "tokens_seen": 815071232 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380320962888666, + "loss": 1.5514, + "theoretical_loss": 3.750714315923454, + "tokens_seen": 815136768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003803109327983952, + "loss": 1.5267, + "theoretical_loss": 3.7506847517800286, + "tokens_seen": 815202304 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038030090270812437, + "loss": 1.4464, + "theoretical_loss": 3.7506551906786605, + "tokens_seen": 815267840 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038029087261785355, + "loss": 1.4062, + "theoretical_loss": 3.7506256326187923, + "tokens_seen": 815333376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802808425275828, + "loss": 1.5848, + "theoretical_loss": 3.7505960775998677, + "tokens_seen": 815398912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802708124373119, + "loss": 1.4026, + "theoretical_loss": 3.7505665256213288, + "tokens_seen": 815464448 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038026078234704115, + "loss": 1.5609, + "theoretical_loss": 3.7505369766826178, + "tokens_seen": 815529984 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038025075225677033, + "loss": 1.4507, + "theoretical_loss": 3.7505074307831787, + "tokens_seen": 815595520 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802407221664995, + "loss": 1.4047, + "theoretical_loss": 3.7504778879224547, + "tokens_seen": 815661056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802306920762287, + "loss": 1.3716, + "theoretical_loss": 3.750448348099889, + "tokens_seen": 815726592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802206619859579, + "loss": 1.4724, + "theoretical_loss": 3.7504188113149244, + "tokens_seen": 815792128 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038021063189568705, + "loss": 1.3328, + "theoretical_loss": 3.750389277567006, + "tokens_seen": 815857664 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 301260, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4506981372833252, + "objective/train/theoretical_loss": 3.7503597468555765, + "objective/train/tokens_used": 836383200, + "theoretical_loss": 3.7503597468555765, + "tokens_seen": 815923200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003802006018054163, + "loss": 1.4678, + "theoretical_loss": 3.7503597468555765, + "tokens_seen": 815923200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003801905717151454, + "loss": 1.3516, + "theoretical_loss": 3.7503302191800802, + "tokens_seen": 815988736 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038018054162487465, + "loss": 1.4871, + "theoretical_loss": 3.750300694539961, + "tokens_seen": 816054272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003801705115346038, + "loss": 1.3589, + "theoretical_loss": 3.750271172934663, + "tokens_seen": 816119808 + }, + { + "epoch": 0.25, + "learning_rate": 0.000380160481444333, + "loss": 1.4975, + "theoretical_loss": 3.750241654363631, + "tokens_seen": 816185344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003801504513540622, + "loss": 1.4537, + "theoretical_loss": 3.75021213882631, + "tokens_seen": 816250880 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003801404212637914, + "loss": 1.5677, + "theoretical_loss": 3.7501826263221436, + "tokens_seen": 816316416 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038013039117352056, + "loss": 1.5216, + "theoretical_loss": 3.7501531168505773, + "tokens_seen": 816381952 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038012036108324974, + "loss": 1.5007, + "theoretical_loss": 3.750123610411056, + "tokens_seen": 816447488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003801103309929789, + "loss": 1.4544, + "theoretical_loss": 3.750094107003024, + "tokens_seen": 816513024 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038010030090270816, + "loss": 1.3104, + "theoretical_loss": 3.7500646066259278, + "tokens_seen": 816578560 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800902708124373, + "loss": 1.5603, + "theoretical_loss": 3.750035109279212, + "tokens_seen": 816644096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800802407221665, + "loss": 1.4577, + "theoretical_loss": 3.7500056149623227, + "tokens_seen": 816709632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800702106318957, + "loss": 1.3018, + "theoretical_loss": 3.7499761236747053, + "tokens_seen": 816775168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800601805416249, + "loss": 1.4276, + "theoretical_loss": 3.7499466354158058, + "tokens_seen": 816840704 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038005015045135406, + "loss": 1.514, + "theoretical_loss": 3.7499171501850697, + "tokens_seen": 816906240 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038004012036108324, + "loss": 1.4894, + "theoretical_loss": 3.749887667981944, + "tokens_seen": 816971776 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800300902708124, + "loss": 1.4253, + "theoretical_loss": 3.7498581888058746, + "tokens_seen": 817037312 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038002006018054166, + "loss": 1.526, + "theoretical_loss": 3.7498287126563072, + "tokens_seen": 817102848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003800100300902708, + "loss": 1.4363, + "theoretical_loss": 3.7497992395326896, + "tokens_seen": 817168384 + }, + { + "epoch": 0.25, + "learning_rate": 0.00038, + "loss": 1.4827, + "theoretical_loss": 3.749769769434468, + "tokens_seen": 817233920 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037998996990972915, + "loss": 1.5254, + "theoretical_loss": 3.7497403023610882, + "tokens_seen": 817299456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003799799398194584, + "loss": 1.5221, + "theoretical_loss": 3.7497108383119997, + "tokens_seen": 817364992 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037996990972918756, + "loss": 1.4964, + "theoretical_loss": 3.749681377286647, + "tokens_seen": 817430528 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037995987963891675, + "loss": 1.4132, + "theoretical_loss": 3.749651919284479, + "tokens_seen": 817496064 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003799498495486459, + "loss": 1.4523, + "theoretical_loss": 3.7496224643049425, + "tokens_seen": 817561600 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037993981945837516, + "loss": 1.4091, + "theoretical_loss": 3.7495930123474857, + "tokens_seen": 817627136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003799297893681043, + "loss": 1.5403, + "theoretical_loss": 3.749563563411556, + "tokens_seen": 817692672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003799197592778335, + "loss": 1.526, + "theoretical_loss": 3.7495341174966015, + "tokens_seen": 817758208 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037990972918756265, + "loss": 1.4771, + "theoretical_loss": 3.7495046746020693, + "tokens_seen": 817823744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003798996990972919, + "loss": 1.455, + "theoretical_loss": 3.7494752347274085, + "tokens_seen": 817889280 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037988966900702107, + "loss": 1.4983, + "theoretical_loss": 3.749445797872067, + "tokens_seen": 817954816 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037987963891675025, + "loss": 1.5314, + "theoretical_loss": 3.7494163640354943, + "tokens_seen": 818020352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003798696088264795, + "loss": 1.4227, + "theoretical_loss": 3.7493869332171377, + "tokens_seen": 818085888 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003798595787362086, + "loss": 1.4584, + "theoretical_loss": 3.7493575054164463, + "tokens_seen": 818151424 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037984954864593785, + "loss": 1.4344, + "theoretical_loss": 3.749328080632869, + "tokens_seen": 818216960 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037983951855566703, + "loss": 1.4299, + "theoretical_loss": 3.749298658865855, + "tokens_seen": 818282496 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003798294884653962, + "loss": 1.4034, + "theoretical_loss": 3.7492692401148537, + "tokens_seen": 818348032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003798194583751254, + "loss": 1.4665, + "theoretical_loss": 3.7492398243793144, + "tokens_seen": 818413568 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037980942828485457, + "loss": 1.3354, + "theoretical_loss": 3.7492104116586855, + "tokens_seen": 818479104 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037979939819458375, + "loss": 1.5627, + "theoretical_loss": 3.749181001952418, + "tokens_seen": 818544640 + }, + { + "epoch": 0.25, + "learning_rate": 0.000379789368104313, + "loss": 1.4671, + "theoretical_loss": 3.7491515952599617, + "tokens_seen": 818610176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003797793380140421, + "loss": 1.5089, + "theoretical_loss": 3.7491221915807653, + "tokens_seen": 818675712 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037976930792377135, + "loss": 1.4709, + "theoretical_loss": 3.74909279091428, + "tokens_seen": 818741248 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037975927783350053, + "loss": 1.5099, + "theoretical_loss": 3.749063393259955, + "tokens_seen": 818806784 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003797492477432297, + "loss": 1.5089, + "theoretical_loss": 3.749033998617241, + "tokens_seen": 818872320 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003797392176529589, + "loss": 1.3695, + "theoretical_loss": 3.7490046069855896, + "tokens_seen": 818937856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003797291875626881, + "loss": 1.3004, + "theoretical_loss": 3.74897521836445, + "tokens_seen": 819003392 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037971915747241725, + "loss": 1.3138, + "theoretical_loss": 3.7489458327532734, + "tokens_seen": 819068928 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003797091273821465, + "loss": 1.4793, + "theoretical_loss": 3.748916450151511, + "tokens_seen": 819134464 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 302476, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5230739116668701, + "objective/train/theoretical_loss": 3.7488870705586135, + "objective/train/tokens_used": 839660000, + "theoretical_loss": 3.7488870705586135, + "tokens_seen": 819200000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003796990972918756, + "loss": 1.5213, + "theoretical_loss": 3.7488870705586135, + "tokens_seen": 819200000 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037968906720160485, + "loss": 1.33, + "theoretical_loss": 3.7488576939740326, + "tokens_seen": 819265536 + }, + { + "epoch": 0.25, + "learning_rate": 0.000379679037111334, + "loss": 1.4352, + "theoretical_loss": 3.7488283203972195, + "tokens_seen": 819331072 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003796690070210632, + "loss": 1.47, + "theoretical_loss": 3.748798949827625, + "tokens_seen": 819396608 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003796589769307924, + "loss": 1.4271, + "theoretical_loss": 3.7487695822647016, + "tokens_seen": 819462144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003796489468405216, + "loss": 1.5346, + "theoretical_loss": 3.7487402177079003, + "tokens_seen": 819527680 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037963891675025076, + "loss": 1.4404, + "theoretical_loss": 3.7487108561566735, + "tokens_seen": 819593216 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037962888665997994, + "loss": 1.4822, + "theoretical_loss": 3.748681497610474, + "tokens_seen": 819658752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003796188565697091, + "loss": 1.4962, + "theoretical_loss": 3.748652142068753, + "tokens_seen": 819724288 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037960882647943836, + "loss": 1.4391, + "theoretical_loss": 3.748622789530963, + "tokens_seen": 819789824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795987963891675, + "loss": 1.5685, + "theoretical_loss": 3.748593439996557, + "tokens_seen": 819855360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795887662988967, + "loss": 1.4987, + "theoretical_loss": 3.748564093464987, + "tokens_seen": 819920896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795787362086259, + "loss": 1.4311, + "theoretical_loss": 3.748534749935706, + "tokens_seen": 819986432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795687061183551, + "loss": 1.3867, + "theoretical_loss": 3.748505409408167, + "tokens_seen": 820051968 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037955867602808426, + "loss": 1.3638, + "theoretical_loss": 3.7484760718818233, + "tokens_seen": 820117504 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037954864593781344, + "loss": 1.4973, + "theoretical_loss": 3.748446737356128, + "tokens_seen": 820183040 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795386158475426, + "loss": 1.5065, + "theoretical_loss": 3.7484174058305344, + "tokens_seen": 820248576 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037952858575727186, + "loss": 1.4829, + "theoretical_loss": 3.748388077304496, + "tokens_seen": 820314112 + }, + { + "epoch": 0.25, + "learning_rate": 0.000379518555667001, + "loss": 1.4216, + "theoretical_loss": 3.7483587517774666, + "tokens_seen": 820379648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003795085255767302, + "loss": 1.406, + "theoretical_loss": 3.7483294292488996, + "tokens_seen": 820445184 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037949849548645935, + "loss": 1.4779, + "theoretical_loss": 3.748300109718249, + "tokens_seen": 820510720 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003794884653961886, + "loss": 1.4153, + "theoretical_loss": 3.74827079318497, + "tokens_seen": 820576256 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037947843530591776, + "loss": 1.6167, + "theoretical_loss": 3.748241479648515, + "tokens_seen": 820641792 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037946840521564695, + "loss": 1.3092, + "theoretical_loss": 3.7482121691083394, + "tokens_seen": 820707328 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003794583751253761, + "loss": 1.4895, + "theoretical_loss": 3.748182861563898, + "tokens_seen": 820772864 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037944834503510536, + "loss": 1.4344, + "theoretical_loss": 3.748153557014645, + "tokens_seen": 820838400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003794383149448345, + "loss": 1.4043, + "theoretical_loss": 3.748124255460035, + "tokens_seen": 820903936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003794282848545637, + "loss": 1.5839, + "theoretical_loss": 3.748094956899523, + "tokens_seen": 820969472 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037941825476429285, + "loss": 1.5487, + "theoretical_loss": 3.7480656613325642, + "tokens_seen": 821035008 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003794082246740221, + "loss": 1.4875, + "theoretical_loss": 3.7480363687586147, + "tokens_seen": 821100544 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037939819458375127, + "loss": 1.4125, + "theoretical_loss": 3.7480070791771283, + "tokens_seen": 821166080 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037938816449348045, + "loss": 1.5952, + "theoretical_loss": 3.7479777925875615, + "tokens_seen": 821231616 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037937813440320963, + "loss": 1.5244, + "theoretical_loss": 3.74794850898937, + "tokens_seen": 821297152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003793681043129388, + "loss": 1.5353, + "theoretical_loss": 3.7479192283820084, + "tokens_seen": 821362688 + }, + { + "epoch": 0.25, + "learning_rate": 0.000379358074222668, + "loss": 1.4472, + "theoretical_loss": 3.747889950764934, + "tokens_seen": 821428224 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037934804413239723, + "loss": 1.5885, + "theoretical_loss": 3.747860676137603, + "tokens_seen": 821493760 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037933801404212635, + "loss": 1.4696, + "theoretical_loss": 3.74783140449947, + "tokens_seen": 821559296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003793279839518556, + "loss": 1.4536, + "theoretical_loss": 3.747802135849993, + "tokens_seen": 821624832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003793179538615847, + "loss": 1.5081, + "theoretical_loss": 3.747772870188628, + "tokens_seen": 821690368 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037930792377131395, + "loss": 1.4745, + "theoretical_loss": 3.747743607514831, + "tokens_seen": 821755904 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037929789368104313, + "loss": 1.4628, + "theoretical_loss": 3.74771434782806, + "tokens_seen": 821821440 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003792878635907723, + "loss": 1.3581, + "theoretical_loss": 3.747685091127771, + "tokens_seen": 821886976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003792778335005015, + "loss": 1.3775, + "theoretical_loss": 3.7476558374134217, + "tokens_seen": 821952512 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037926780341023073, + "loss": 1.6289, + "theoretical_loss": 3.7476265866844685, + "tokens_seen": 822018048 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037925777331995986, + "loss": 1.4166, + "theoretical_loss": 3.7475973389403694, + "tokens_seen": 822083584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003792477432296891, + "loss": 1.3646, + "theoretical_loss": 3.747568094180582, + "tokens_seen": 822149120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003792377131394182, + "loss": 1.4653, + "theoretical_loss": 3.747538852404563, + "tokens_seen": 822214656 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037922768304914746, + "loss": 1.3806, + "theoretical_loss": 3.747509613611771, + "tokens_seen": 822280192 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037921765295887664, + "loss": 1.4365, + "theoretical_loss": 3.747480377801664, + "tokens_seen": 822345728 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003792076228686058, + "loss": 1.4553, + "theoretical_loss": 3.7474511449737, + "tokens_seen": 822411264 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 303734, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.492797613143921, + "objective/train/theoretical_loss": 3.7474219151273367, + "objective/train/tokens_used": 842936800, + "theoretical_loss": 3.7474219151273367, + "tokens_seen": 822476800 + }, + { + "epoch": 0.25, + "learning_rate": 0.000379197592778335, + "loss": 1.5452, + "theoretical_loss": 3.7474219151273367, + "tokens_seen": 822476800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791875626880642, + "loss": 1.5299, + "theoretical_loss": 3.7473926882620328, + "tokens_seen": 822542336 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037917753259779336, + "loss": 1.5565, + "theoretical_loss": 3.7473634643772473, + "tokens_seen": 822607872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791675025075226, + "loss": 1.5405, + "theoretical_loss": 3.747334243472438, + "tokens_seen": 822673408 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791574724172517, + "loss": 1.3502, + "theoretical_loss": 3.747305025547064, + "tokens_seen": 822738944 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037914744232698096, + "loss": 1.5628, + "theoretical_loss": 3.747275810600584, + "tokens_seen": 822804480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791374122367101, + "loss": 1.4527, + "theoretical_loss": 3.747246598632458, + "tokens_seen": 822870016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791273821464393, + "loss": 1.438, + "theoretical_loss": 3.7472173896421435, + "tokens_seen": 822935552 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037911735205616856, + "loss": 1.4729, + "theoretical_loss": 3.7471881836291017, + "tokens_seen": 823001088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003791073219658977, + "loss": 1.4329, + "theoretical_loss": 3.7471589805927907, + "tokens_seen": 823066624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790972918756269, + "loss": 1.4519, + "theoretical_loss": 3.7471297805326707, + "tokens_seen": 823132160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790872617853561, + "loss": 1.5586, + "theoretical_loss": 3.7471005834482014, + "tokens_seen": 823197696 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790772316950853, + "loss": 1.5472, + "theoretical_loss": 3.7470713893388425, + "tokens_seen": 823263232 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037906720160481446, + "loss": 1.3898, + "theoretical_loss": 3.7470421982040545, + "tokens_seen": 823328768 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037905717151454364, + "loss": 1.3926, + "theoretical_loss": 3.7470130100432972, + "tokens_seen": 823394304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790471414242728, + "loss": 1.4257, + "theoretical_loss": 3.746983824856031, + "tokens_seen": 823459840 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037903711133400206, + "loss": 1.4081, + "theoretical_loss": 3.7469546426417164, + "tokens_seen": 823525376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790270812437312, + "loss": 1.5316, + "theoretical_loss": 3.746925463399814, + "tokens_seen": 823590912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003790170511534604, + "loss": 1.4162, + "theoretical_loss": 3.746896287129785, + "tokens_seen": 823656448 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037900702106318955, + "loss": 1.5216, + "theoretical_loss": 3.7468671138310894, + "tokens_seen": 823721984 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003789969909729188, + "loss": 1.5021, + "theoretical_loss": 3.7468379435031887, + "tokens_seen": 823787520 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037898696088264796, + "loss": 1.4198, + "theoretical_loss": 3.7468087761455444, + "tokens_seen": 823853056 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037897693079237715, + "loss": 1.4292, + "theoretical_loss": 3.746779611757617, + "tokens_seen": 823918592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003789669007021063, + "loss": 1.3152, + "theoretical_loss": 3.746750450338869, + "tokens_seen": 823984128 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037895687061183556, + "loss": 1.488, + "theoretical_loss": 3.746721291888761, + "tokens_seen": 824049664 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003789468405215647, + "loss": 1.4621, + "theoretical_loss": 3.7466921364067556, + "tokens_seen": 824115200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003789368104312939, + "loss": 1.5227, + "theoretical_loss": 3.746662983892314, + "tokens_seen": 824180736 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037892678034102305, + "loss": 1.5217, + "theoretical_loss": 3.746633834344898, + "tokens_seen": 824246272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003789167502507523, + "loss": 1.4829, + "theoretical_loss": 3.7466046877639707, + "tokens_seen": 824311808 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037890672016048147, + "loss": 1.443, + "theoretical_loss": 3.7465755441489943, + "tokens_seen": 824377344 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037889669007021065, + "loss": 1.5312, + "theoretical_loss": 3.74654640349943, + "tokens_seen": 824442880 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037888665997993983, + "loss": 1.5703, + "theoretical_loss": 3.7465172658147416, + "tokens_seen": 824508416 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378876629889669, + "loss": 1.4344, + "theoretical_loss": 3.7464881310943916, + "tokens_seen": 824573952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003788665997993982, + "loss": 1.372, + "theoretical_loss": 3.7464589993378423, + "tokens_seen": 824639488 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037885656970912743, + "loss": 1.4494, + "theoretical_loss": 3.7464298705445573, + "tokens_seen": 824705024 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037884653961885655, + "loss": 1.5109, + "theoretical_loss": 3.7464007447139993, + "tokens_seen": 824770560 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003788365095285858, + "loss": 1.5259, + "theoretical_loss": 3.746371621845632, + "tokens_seen": 824836096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003788264794383149, + "loss": 1.3971, + "theoretical_loss": 3.7463425019389187, + "tokens_seen": 824901632 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037881644934804415, + "loss": 1.443, + "theoretical_loss": 3.746313384993323, + "tokens_seen": 824967168 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037880641925777333, + "loss": 1.4712, + "theoretical_loss": 3.746284271008308, + "tokens_seen": 825032704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003787963891675025, + "loss": 1.5219, + "theoretical_loss": 3.7462551599833382, + "tokens_seen": 825098240 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003787863590772317, + "loss": 1.4298, + "theoretical_loss": 3.746226051917877, + "tokens_seen": 825163776 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037877632898696093, + "loss": 1.3537, + "theoretical_loss": 3.7461969468113896, + "tokens_seen": 825229312 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037876629889669006, + "loss": 1.5708, + "theoretical_loss": 3.746167844663339, + "tokens_seen": 825294848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003787562688064193, + "loss": 1.4243, + "theoretical_loss": 3.7461387454731905, + "tokens_seen": 825360384 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003787462387161484, + "loss": 1.3586, + "theoretical_loss": 3.746109649240408, + "tokens_seen": 825425920 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037873620862587766, + "loss": 1.596, + "theoretical_loss": 3.7460805559644568, + "tokens_seen": 825491456 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037872617853560684, + "loss": 1.4383, + "theoretical_loss": 3.7460514656448005, + "tokens_seen": 825556992 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378716148445336, + "loss": 1.489, + "theoretical_loss": 3.7460223782809057, + "tokens_seen": 825622528 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003787061183550652, + "loss": 1.5258, + "theoretical_loss": 3.7459932938722362, + "tokens_seen": 825688064 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 304950, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.395533561706543, + "objective/train/theoretical_loss": 3.745964212418258, + "objective/train/tokens_used": 846213600, + "theoretical_loss": 3.745964212418258, + "tokens_seen": 825753600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786960882647944, + "loss": 1.4638, + "theoretical_loss": 3.745964212418258, + "tokens_seen": 825753600 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037868605817452356, + "loss": 1.5339, + "theoretical_loss": 3.7459351339184366, + "tokens_seen": 825819136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786760280842528, + "loss": 1.5239, + "theoretical_loss": 3.7459060583722366, + "tokens_seen": 825884672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786659979939819, + "loss": 1.513, + "theoretical_loss": 3.745876985779124, + "tokens_seen": 825950208 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037865596790371116, + "loss": 1.482, + "theoretical_loss": 3.745847916138565, + "tokens_seen": 826015744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786459378134403, + "loss": 1.4506, + "theoretical_loss": 3.745818849450025, + "tokens_seen": 826081280 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786359077231695, + "loss": 1.488, + "theoretical_loss": 3.7457897857129705, + "tokens_seen": 826146816 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786258776328987, + "loss": 1.482, + "theoretical_loss": 3.7457607249268676, + "tokens_seen": 826212352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003786158475426279, + "loss": 1.4927, + "theoretical_loss": 3.7457316670911824, + "tokens_seen": 826277888 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037860581745235706, + "loss": 1.392, + "theoretical_loss": 3.7457026122053816, + "tokens_seen": 826343424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003785957873620863, + "loss": 1.4598, + "theoretical_loss": 3.7456735602689317, + "tokens_seen": 826408960 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003785857572718154, + "loss": 1.4266, + "theoretical_loss": 3.7456445112812995, + "tokens_seen": 826474496 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037857572718154466, + "loss": 1.4902, + "theoretical_loss": 3.745615465241952, + "tokens_seen": 826540032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003785656970912738, + "loss": 1.3234, + "theoretical_loss": 3.745586422150356, + "tokens_seen": 826605568 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378555667001003, + "loss": 1.5226, + "theoretical_loss": 3.745557382005979, + "tokens_seen": 826671104 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003785456369107322, + "loss": 1.4918, + "theoretical_loss": 3.7455283448082883, + "tokens_seen": 826736640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003785356068204614, + "loss": 1.5506, + "theoretical_loss": 3.7454993105567507, + "tokens_seen": 826802176 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037852557673019057, + "loss": 1.4566, + "theoretical_loss": 3.745470279250835, + "tokens_seen": 826867712 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037851554663991975, + "loss": 1.4269, + "theoretical_loss": 3.7454412508900075, + "tokens_seen": 826933248 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037850551654964893, + "loss": 1.405, + "theoretical_loss": 3.7454122254737374, + "tokens_seen": 826998784 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037849548645937816, + "loss": 1.5673, + "theoretical_loss": 3.7453832030014915, + "tokens_seen": 827064320 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003784854563691073, + "loss": 1.5608, + "theoretical_loss": 3.745354183472739, + "tokens_seen": 827129856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003784754262788365, + "loss": 1.6036, + "theoretical_loss": 3.745325166886947, + "tokens_seen": 827195392 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037846539618856565, + "loss": 1.5346, + "theoretical_loss": 3.7452961532435856, + "tokens_seen": 827260928 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003784553660982949, + "loss": 1.5385, + "theoretical_loss": 3.7452671425421213, + "tokens_seen": 827326464 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037844533600802407, + "loss": 1.5392, + "theoretical_loss": 3.7452381347820247, + "tokens_seen": 827392000 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037843530591775325, + "loss": 1.6373, + "theoretical_loss": 3.7452091299627637, + "tokens_seen": 827457536 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037842527582748243, + "loss": 1.5187, + "theoretical_loss": 3.745180128083807, + "tokens_seen": 827523072 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037841524573721167, + "loss": 1.4724, + "theoretical_loss": 3.7451511291446242, + "tokens_seen": 827588608 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003784052156469408, + "loss": 1.4574, + "theoretical_loss": 3.745122133144685, + "tokens_seen": 827654144 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037839518555667003, + "loss": 1.516, + "theoretical_loss": 3.7450931400834575, + "tokens_seen": 827719680 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037838515546639916, + "loss": 1.5164, + "theoretical_loss": 3.7450641499604123, + "tokens_seen": 827785216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003783751253761284, + "loss": 1.5322, + "theoretical_loss": 3.7450351627750185, + "tokens_seen": 827850752 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037836509528585763, + "loss": 1.4794, + "theoretical_loss": 3.7450061785267463, + "tokens_seen": 827916288 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037835506519558675, + "loss": 1.5236, + "theoretical_loss": 3.744977197215065, + "tokens_seen": 827981824 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378345035105316, + "loss": 1.4888, + "theoretical_loss": 3.744948218839446, + "tokens_seen": 828047360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003783350050150451, + "loss": 1.5263, + "theoretical_loss": 3.744919243399358, + "tokens_seen": 828112896 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037832497492477435, + "loss": 1.4465, + "theoretical_loss": 3.744890270894272, + "tokens_seen": 828178432 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037831494483450353, + "loss": 1.5519, + "theoretical_loss": 3.7448613013236587, + "tokens_seen": 828243968 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003783049147442327, + "loss": 1.5845, + "theoretical_loss": 3.744832334686988, + "tokens_seen": 828309504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782948846539619, + "loss": 1.6293, + "theoretical_loss": 3.744803370983732, + "tokens_seen": 828375040 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037828485456369113, + "loss": 1.5691, + "theoretical_loss": 3.7447744102133598, + "tokens_seen": 828440576 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037827482447342026, + "loss": 1.5118, + "theoretical_loss": 3.744745452375344, + "tokens_seen": 828506112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782647943831495, + "loss": 1.3884, + "theoretical_loss": 3.7447164974691547, + "tokens_seen": 828571648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782547642928786, + "loss": 1.4538, + "theoretical_loss": 3.744687545494264, + "tokens_seen": 828637184 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037824473420260786, + "loss": 1.5078, + "theoretical_loss": 3.744658596450143, + "tokens_seen": 828702720 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037823470411233704, + "loss": 1.4861, + "theoretical_loss": 3.744629650336263, + "tokens_seen": 828768256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782246740220662, + "loss": 1.4545, + "theoretical_loss": 3.744600707152096, + "tokens_seen": 828833792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782146439317954, + "loss": 1.5512, + "theoretical_loss": 3.7445717668971144, + "tokens_seen": 828899328 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003782046138415246, + "loss": 1.4283, + "theoretical_loss": 3.7445428295707894, + "tokens_seen": 828964864 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 306096, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3760583400726318, + "objective/train/theoretical_loss": 3.7445138951725934, + "objective/train/tokens_used": 849490400, + "theoretical_loss": 3.7445138951725934, + "tokens_seen": 829030400 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037819458375125376, + "loss": 1.4779, + "theoretical_loss": 3.7445138951725934, + "tokens_seen": 829030400 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378184553660983, + "loss": 1.4219, + "theoretical_loss": 3.7444849637019986, + "tokens_seen": 829095936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781745235707121, + "loss": 1.5865, + "theoretical_loss": 3.7444560351584775, + "tokens_seen": 829161472 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037816449348044136, + "loss": 1.4569, + "theoretical_loss": 3.744427109541503, + "tokens_seen": 829227008 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781544633901705, + "loss": 1.3728, + "theoretical_loss": 3.744398186850547, + "tokens_seen": 829292544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781444332998997, + "loss": 1.4307, + "theoretical_loss": 3.7443692670850828, + "tokens_seen": 829358080 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781344032096289, + "loss": 1.4977, + "theoretical_loss": 3.744340350244583, + "tokens_seen": 829423616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781243731193581, + "loss": 1.5131, + "theoretical_loss": 3.7443114363285215, + "tokens_seen": 829489152 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037811434302908726, + "loss": 1.3691, + "theoretical_loss": 3.74428252533637, + "tokens_seen": 829554688 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003781043129388165, + "loss": 1.5102, + "theoretical_loss": 3.7442536172676038, + "tokens_seen": 829620224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003780942828485456, + "loss": 1.4773, + "theoretical_loss": 3.744224712121695, + "tokens_seen": 829685760 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037808425275827486, + "loss": 1.3364, + "theoretical_loss": 3.744195809898118, + "tokens_seen": 829751296 + }, + { + "epoch": 0.25, + "learning_rate": 0.000378074222668004, + "loss": 1.3887, + "theoretical_loss": 3.7441669105963458, + "tokens_seen": 829816832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003780641925777332, + "loss": 1.58, + "theoretical_loss": 3.744138014215853, + "tokens_seen": 829882368 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003780541624874624, + "loss": 1.4436, + "theoretical_loss": 3.7441091207561135, + "tokens_seen": 829947904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003780441323971916, + "loss": 1.4658, + "theoretical_loss": 3.7440802302166016, + "tokens_seen": 830013440 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037803410230692077, + "loss": 1.4801, + "theoretical_loss": 3.7440513425967907, + "tokens_seen": 830078976 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037802407221664995, + "loss": 1.5174, + "theoretical_loss": 3.7440224578961563, + "tokens_seen": 830144512 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037801404212637913, + "loss": 1.4811, + "theoretical_loss": 3.743993576114173, + "tokens_seen": 830210048 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037800401203610836, + "loss": 1.3713, + "theoretical_loss": 3.7439646972503144, + "tokens_seen": 830275584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003779939819458375, + "loss": 1.5017, + "theoretical_loss": 3.7439358213040568, + "tokens_seen": 830341120 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037798395185556673, + "loss": 1.4616, + "theoretical_loss": 3.743906948274874, + "tokens_seen": 830406656 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037797392176529585, + "loss": 1.547, + "theoretical_loss": 3.7438780781622425, + "tokens_seen": 830472192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003779638916750251, + "loss": 1.4385, + "theoretical_loss": 3.743849210965636, + "tokens_seen": 830537728 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037795386158475427, + "loss": 1.4496, + "theoretical_loss": 3.7438203466845303, + "tokens_seen": 830603264 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037794383149448345, + "loss": 1.4696, + "theoretical_loss": 3.7437914853184022, + "tokens_seen": 830668800 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037793380140421263, + "loss": 1.5391, + "theoretical_loss": 3.743762626866726, + "tokens_seen": 830734336 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037792377131394187, + "loss": 1.3893, + "theoretical_loss": 3.7437337713289778, + "tokens_seen": 830799872 + }, + { + "epoch": 0.25, + "learning_rate": 0.000377913741223671, + "loss": 1.4241, + "theoretical_loss": 3.743704918704634, + "tokens_seen": 830865408 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037790371113340023, + "loss": 1.4916, + "theoretical_loss": 3.74367606899317, + "tokens_seen": 830930944 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037789368104312936, + "loss": 1.454, + "theoretical_loss": 3.743647222194063, + "tokens_seen": 830996480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003778836509528586, + "loss": 1.5215, + "theoretical_loss": 3.7436183783067882, + "tokens_seen": 831062016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003778736208625878, + "loss": 1.4056, + "theoretical_loss": 3.7435895373308226, + "tokens_seen": 831127552 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037786359077231695, + "loss": 1.4034, + "theoretical_loss": 3.743560699265643, + "tokens_seen": 831193088 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037785356068204614, + "loss": 1.527, + "theoretical_loss": 3.743531864110726, + "tokens_seen": 831258624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003778435305917753, + "loss": 1.4251, + "theoretical_loss": 3.743503031865549, + "tokens_seen": 831324160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003778335005015045, + "loss": 1.5059, + "theoretical_loss": 3.743474202529588, + "tokens_seen": 831389696 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037782347041123373, + "loss": 1.4178, + "theoretical_loss": 3.743445376102321, + "tokens_seen": 831455232 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037781344032096286, + "loss": 1.4848, + "theoretical_loss": 3.7434165525832253, + "tokens_seen": 831520768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003778034102306921, + "loss": 1.4307, + "theoretical_loss": 3.7433877319717777, + "tokens_seen": 831586304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003777933801404213, + "loss": 1.5043, + "theoretical_loss": 3.743358914267456, + "tokens_seen": 831651840 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037778335005015046, + "loss": 1.6129, + "theoretical_loss": 3.743330099469738, + "tokens_seen": 831717376 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037777331995987964, + "loss": 1.3745, + "theoretical_loss": 3.7433012875781015, + "tokens_seen": 831782912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003777632898696088, + "loss": 1.3691, + "theoretical_loss": 3.743272478592025, + "tokens_seen": 831848448 + }, + { + "epoch": 0.25, + "learning_rate": 0.000377753259779338, + "loss": 1.6046, + "theoretical_loss": 3.743243672510986, + "tokens_seen": 831913984 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037774322968906724, + "loss": 1.5031, + "theoretical_loss": 3.7432148693344627, + "tokens_seen": 831979520 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037773319959879636, + "loss": 1.5603, + "theoretical_loss": 3.7431860690619336, + "tokens_seen": 832045056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003777231695085256, + "loss": 1.4595, + "theoretical_loss": 3.7431572716928776, + "tokens_seen": 832110592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003777131394182547, + "loss": 1.4305, + "theoretical_loss": 3.743128477226773, + "tokens_seen": 832176128 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037770310932798396, + "loss": 1.4945, + "theoretical_loss": 3.7430996856630987, + "tokens_seen": 832241664 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 307306, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4790654182434082, + "objective/train/theoretical_loss": 3.743070897001333, + "objective/train/tokens_used": 852767200, + "theoretical_loss": 3.743070897001333, + "tokens_seen": 832307200 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037769307923771314, + "loss": 1.4127, + "theoretical_loss": 3.743070897001333, + "tokens_seen": 832307200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776830491474423, + "loss": 1.4917, + "theoretical_loss": 3.743042111240956, + "tokens_seen": 832372736 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776730190571715, + "loss": 1.436, + "theoretical_loss": 3.7430133283814464, + "tokens_seen": 832438272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776629889669007, + "loss": 1.4083, + "theoretical_loss": 3.7429845484222835, + "tokens_seen": 832503808 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037765295887662987, + "loss": 1.5366, + "theoretical_loss": 3.742955771362947, + "tokens_seen": 832569344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776429287863591, + "loss": 1.4674, + "theoretical_loss": 3.7429269972029156, + "tokens_seen": 832634880 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037763289869608823, + "loss": 1.5662, + "theoretical_loss": 3.74289822594167, + "tokens_seen": 832700416 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037762286860581746, + "loss": 1.4472, + "theoretical_loss": 3.7428694575786903, + "tokens_seen": 832765952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776128385155467, + "loss": 1.4589, + "theoretical_loss": 3.742840692113455, + "tokens_seen": 832831488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003776028084252758, + "loss": 1.5815, + "theoretical_loss": 3.7428119295454456, + "tokens_seen": 832897024 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037759277833500506, + "loss": 1.4439, + "theoretical_loss": 3.7427831698741425, + "tokens_seen": 832962560 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003775827482447342, + "loss": 1.3736, + "theoretical_loss": 3.742754413099025, + "tokens_seen": 833028096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003775727181544634, + "loss": 1.5046, + "theoretical_loss": 3.742725659219574, + "tokens_seen": 833093632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003775626880641926, + "loss": 1.4587, + "theoretical_loss": 3.7426969082352706, + "tokens_seen": 833159168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003775526579739218, + "loss": 1.4591, + "theoretical_loss": 3.7426681601455947, + "tokens_seen": 833224704 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037754262788365097, + "loss": 1.4867, + "theoretical_loss": 3.7426394149500286, + "tokens_seen": 833290240 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037753259779338015, + "loss": 1.5028, + "theoretical_loss": 3.742610672648053, + "tokens_seen": 833355776 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037752256770310933, + "loss": 1.5201, + "theoretical_loss": 3.7425819332391477, + "tokens_seen": 833421312 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037751253761283857, + "loss": 1.4682, + "theoretical_loss": 3.7425531967227954, + "tokens_seen": 833486848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003775025075225677, + "loss": 1.4991, + "theoretical_loss": 3.7425244630984773, + "tokens_seen": 833552384 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037749247743229693, + "loss": 1.558, + "theoretical_loss": 3.742495732365675, + "tokens_seen": 833617920 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037748244734202605, + "loss": 1.4883, + "theoretical_loss": 3.74246700452387, + "tokens_seen": 833683456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003774724172517553, + "loss": 1.5273, + "theoretical_loss": 3.7424382795725446, + "tokens_seen": 833748992 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037746238716148447, + "loss": 1.4165, + "theoretical_loss": 3.7424095575111807, + "tokens_seen": 833814528 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037745235707121365, + "loss": 1.3689, + "theoretical_loss": 3.74238083833926, + "tokens_seen": 833880064 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037744232698094283, + "loss": 1.5189, + "theoretical_loss": 3.742352122056265, + "tokens_seen": 833945600 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037743229689067207, + "loss": 1.5201, + "theoretical_loss": 3.7423234086616777, + "tokens_seen": 834011136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003774222668004012, + "loss": 1.4992, + "theoretical_loss": 3.7422946981549816, + "tokens_seen": 834076672 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037741223671013043, + "loss": 1.5571, + "theoretical_loss": 3.7422659905356586, + "tokens_seen": 834142208 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037740220661985956, + "loss": 1.5722, + "theoretical_loss": 3.742237285803192, + "tokens_seen": 834207744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003773921765295888, + "loss": 1.5233, + "theoretical_loss": 3.7422085839570647, + "tokens_seen": 834273280 + }, + { + "epoch": 0.25, + "learning_rate": 0.000377382146439318, + "loss": 1.5991, + "theoretical_loss": 3.7421798849967596, + "tokens_seen": 834338816 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037737211634904715, + "loss": 1.5482, + "theoretical_loss": 3.742151188921759, + "tokens_seen": 834404352 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037736208625877634, + "loss": 1.5401, + "theoretical_loss": 3.742122495731548, + "tokens_seen": 834469888 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003773520561685055, + "loss": 1.5607, + "theoretical_loss": 3.742093805425609, + "tokens_seen": 834535424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003773420260782347, + "loss": 1.586, + "theoretical_loss": 3.7420651180034255, + "tokens_seen": 834600960 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037733199598796393, + "loss": 1.5525, + "theoretical_loss": 3.742036433464482, + "tokens_seen": 834666496 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037732196589769306, + "loss": 1.5014, + "theoretical_loss": 3.742007751808261, + "tokens_seen": 834732032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003773119358074223, + "loss": 1.4427, + "theoretical_loss": 3.741979073034248, + "tokens_seen": 834797568 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003773019057171515, + "loss": 1.565, + "theoretical_loss": 3.7419503971419266, + "tokens_seen": 834863104 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037729187562688066, + "loss": 1.3822, + "theoretical_loss": 3.7419217241307807, + "tokens_seen": 834928640 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037728184553660984, + "loss": 1.5025, + "theoretical_loss": 3.741893054000295, + "tokens_seen": 834994176 + }, + { + "epoch": 0.25, + "learning_rate": 0.000377271815446339, + "loss": 1.5682, + "theoretical_loss": 3.7418643867499544, + "tokens_seen": 835059712 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003772617853560682, + "loss": 1.4742, + "theoretical_loss": 3.7418357223792427, + "tokens_seen": 835125248 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037725175526579744, + "loss": 1.4846, + "theoretical_loss": 3.7418070608876453, + "tokens_seen": 835190784 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037724172517552656, + "loss": 1.4063, + "theoretical_loss": 3.741778402274647, + "tokens_seen": 835256320 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003772316950852558, + "loss": 1.4305, + "theoretical_loss": 3.7417497465397327, + "tokens_seen": 835321856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003772216649949849, + "loss": 1.4437, + "theoretical_loss": 3.741721093682388, + "tokens_seen": 835387392 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037721163490471416, + "loss": 1.4188, + "theoretical_loss": 3.7416924437020977, + "tokens_seen": 835452928 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037720160481444334, + "loss": 1.521, + "theoretical_loss": 3.741663796598348, + "tokens_seen": 835518464 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 308483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4324874877929688, + "objective/train/theoretical_loss": 3.741635152370624, + "objective/train/tokens_used": 856044000, + "theoretical_loss": 3.741635152370624, + "tokens_seen": 835584000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003771915747241725, + "loss": 1.4205, + "theoretical_loss": 3.741635152370624, + "tokens_seen": 835584000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003771815446339017, + "loss": 1.386, + "theoretical_loss": 3.7416065110184116, + "tokens_seen": 835649536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003771715145436309, + "loss": 1.5134, + "theoretical_loss": 3.741577872541196, + "tokens_seen": 835715072 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037716148445336007, + "loss": 1.5521, + "theoretical_loss": 3.741549236938464, + "tokens_seen": 835780608 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003771514543630893, + "loss": 1.4648, + "theoretical_loss": 3.7415206042097013, + "tokens_seen": 835846144 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037714142427281843, + "loss": 1.55, + "theoretical_loss": 3.7414919743543944, + "tokens_seen": 835911680 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037713139418254766, + "loss": 1.4179, + "theoretical_loss": 3.7414633473720293, + "tokens_seen": 835977216 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037712136409227685, + "loss": 1.4642, + "theoretical_loss": 3.741434723262093, + "tokens_seen": 836042752 + }, + { + "epoch": 0.25, + "learning_rate": 0.000377111334002006, + "loss": 1.5081, + "theoretical_loss": 3.7414061020240723, + "tokens_seen": 836108288 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003771013039117352, + "loss": 1.5285, + "theoretical_loss": 3.741377483657453, + "tokens_seen": 836173824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770912738214644, + "loss": 1.4155, + "theoretical_loss": 3.741348868161723, + "tokens_seen": 836239360 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037708124373119357, + "loss": 1.4966, + "theoretical_loss": 3.741320255536369, + "tokens_seen": 836304896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770712136409228, + "loss": 1.4533, + "theoretical_loss": 3.741291645780878, + "tokens_seen": 836370432 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037706118355065193, + "loss": 1.5325, + "theoretical_loss": 3.741263038894737, + "tokens_seen": 836435968 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037705115346038117, + "loss": 1.4803, + "theoretical_loss": 3.741234434877435, + "tokens_seen": 836501504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770411233701103, + "loss": 1.467, + "theoretical_loss": 3.7412058337284577, + "tokens_seen": 836567040 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037703109327983953, + "loss": 1.5222, + "theoretical_loss": 3.741177235447294, + "tokens_seen": 836632576 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770210631895687, + "loss": 1.5582, + "theoretical_loss": 3.7411486400334315, + "tokens_seen": 836698112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770110330992979, + "loss": 1.5059, + "theoretical_loss": 3.7411200474863575, + "tokens_seen": 836763648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003770010030090271, + "loss": 1.4681, + "theoretical_loss": 3.7410914578055605, + "tokens_seen": 836829184 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037699097291875625, + "loss": 1.6344, + "theoretical_loss": 3.741062870990529, + "tokens_seen": 836894720 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037698094282848544, + "loss": 1.3974, + "theoretical_loss": 3.741034287040751, + "tokens_seen": 836960256 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037697091273821467, + "loss": 1.5179, + "theoretical_loss": 3.7410057059557156, + "tokens_seen": 837025792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003769608826479438, + "loss": 1.4804, + "theoretical_loss": 3.7409771277349106, + "tokens_seen": 837091328 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037695085255767303, + "loss": 1.5397, + "theoretical_loss": 3.7409485523778256, + "tokens_seen": 837156864 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003769408224674022, + "loss": 1.4244, + "theoretical_loss": 3.740919979883949, + "tokens_seen": 837222400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003769307923771314, + "loss": 1.4048, + "theoretical_loss": 3.740891410252769, + "tokens_seen": 837287936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003769207622868606, + "loss": 1.4776, + "theoretical_loss": 3.740862843483777, + "tokens_seen": 837353472 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037691073219658976, + "loss": 1.4743, + "theoretical_loss": 3.7408342795764598, + "tokens_seen": 837419008 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037690070210631894, + "loss": 1.4682, + "theoretical_loss": 3.740805718530308, + "tokens_seen": 837484544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768906720160482, + "loss": 1.5537, + "theoretical_loss": 3.7407771603448117, + "tokens_seen": 837550080 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768806419257773, + "loss": 1.5887, + "theoretical_loss": 3.7407486050194594, + "tokens_seen": 837615616 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037687061183550654, + "loss": 1.4643, + "theoretical_loss": 3.7407200525537414, + "tokens_seen": 837681152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768605817452357, + "loss": 1.405, + "theoretical_loss": 3.740691502947148, + "tokens_seen": 837746688 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768505516549649, + "loss": 1.4633, + "theoretical_loss": 3.7406629561991687, + "tokens_seen": 837812224 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037684052156469413, + "loss": 1.5487, + "theoretical_loss": 3.740634412309294, + "tokens_seen": 837877760 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037683049147442326, + "loss": 1.4355, + "theoretical_loss": 3.740605871277014, + "tokens_seen": 837943296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768204613841525, + "loss": 1.5639, + "theoretical_loss": 3.7405773331018195, + "tokens_seen": 838008832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003768104312938817, + "loss": 1.3774, + "theoretical_loss": 3.740548797783201, + "tokens_seen": 838074368 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037680040120361086, + "loss": 1.5368, + "theoretical_loss": 3.740520265320649, + "tokens_seen": 838139904 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037679037111334004, + "loss": 1.4593, + "theoretical_loss": 3.740491735713654, + "tokens_seen": 838205440 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003767803410230692, + "loss": 1.4573, + "theoretical_loss": 3.740463208961708, + "tokens_seen": 838270976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003767703109327984, + "loss": 1.5938, + "theoretical_loss": 3.7404346850643013, + "tokens_seen": 838336512 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037676028084252764, + "loss": 1.4545, + "theoretical_loss": 3.740406164020926, + "tokens_seen": 838402048 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037675025075225676, + "loss": 1.4887, + "theoretical_loss": 3.740377645831072, + "tokens_seen": 838467584 + }, + { + "epoch": 0.25, + "learning_rate": 0.000376740220661986, + "loss": 1.4499, + "theoretical_loss": 3.7403491304942325, + "tokens_seen": 838533120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003767301905717151, + "loss": 1.4603, + "theoretical_loss": 3.740320618009898, + "tokens_seen": 838598656 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037672016048144436, + "loss": 1.4248, + "theoretical_loss": 3.7402921083775604, + "tokens_seen": 838664192 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037671013039117354, + "loss": 1.3918, + "theoretical_loss": 3.7402636015967117, + "tokens_seen": 838729728 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003767001003009027, + "loss": 1.4783, + "theoretical_loss": 3.7402350976668446, + "tokens_seen": 838795264 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 309674, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6205177307128906, + "objective/train/theoretical_loss": 3.7402065965874502, + "objective/train/tokens_used": 859320800, + "theoretical_loss": 3.7402065965874502, + "tokens_seen": 838860800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003766900702106319, + "loss": 1.5065, + "theoretical_loss": 3.7402065965874502, + "tokens_seen": 838860800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003766800401203611, + "loss": 1.4499, + "theoretical_loss": 3.7401780983580215, + "tokens_seen": 838926336 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037667001003009027, + "loss": 1.4439, + "theoretical_loss": 3.7401496029780503, + "tokens_seen": 838991872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003766599799398195, + "loss": 1.3849, + "theoretical_loss": 3.7401211104470296, + "tokens_seen": 839057408 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037664994984954863, + "loss": 1.4269, + "theoretical_loss": 3.740092620764452, + "tokens_seen": 839122944 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037663991975927786, + "loss": 1.5564, + "theoretical_loss": 3.74006413392981, + "tokens_seen": 839188480 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037662988966900705, + "loss": 1.4696, + "theoretical_loss": 3.7400356499425973, + "tokens_seen": 839254016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003766198595787362, + "loss": 1.5715, + "theoretical_loss": 3.7400071688023067, + "tokens_seen": 839319552 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003766098294884654, + "loss": 1.4198, + "theoretical_loss": 3.7399786905084307, + "tokens_seen": 839385088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003765997993981946, + "loss": 1.5019, + "theoretical_loss": 3.739950215060463, + "tokens_seen": 839450624 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037658976930792377, + "loss": 1.4755, + "theoretical_loss": 3.7399217424578977, + "tokens_seen": 839516160 + }, + { + "epoch": 0.25, + "learning_rate": 0.000376579739217653, + "loss": 1.4268, + "theoretical_loss": 3.739893272700227, + "tokens_seen": 839581696 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037656970912738213, + "loss": 1.5646, + "theoretical_loss": 3.7398648057869464, + "tokens_seen": 839647232 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037655967903711137, + "loss": 1.4711, + "theoretical_loss": 3.7398363417175484, + "tokens_seen": 839712768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003765496489468405, + "loss": 1.4665, + "theoretical_loss": 3.7398078804915276, + "tokens_seen": 839778304 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037653961885656973, + "loss": 1.4238, + "theoretical_loss": 3.739779422108377, + "tokens_seen": 839843840 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003765295887662989, + "loss": 1.4304, + "theoretical_loss": 3.7397509665675925, + "tokens_seen": 839909376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003765195586760281, + "loss": 1.4335, + "theoretical_loss": 3.7397225138686676, + "tokens_seen": 839974912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003765095285857573, + "loss": 1.4353, + "theoretical_loss": 3.739694064011097, + "tokens_seen": 840040448 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037649949849548645, + "loss": 1.5278, + "theoretical_loss": 3.7396656169943743, + "tokens_seen": 840105984 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037648946840521564, + "loss": 1.5076, + "theoretical_loss": 3.739637172817996, + "tokens_seen": 840171520 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037647943831494487, + "loss": 1.4329, + "theoretical_loss": 3.739608731481456, + "tokens_seen": 840237056 + }, + { + "epoch": 0.25, + "learning_rate": 0.000376469408224674, + "loss": 1.4759, + "theoretical_loss": 3.7395802929842494, + "tokens_seen": 840302592 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037645937813440323, + "loss": 1.5003, + "theoretical_loss": 3.739551857325871, + "tokens_seen": 840368128 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003764493480441324, + "loss": 1.4765, + "theoretical_loss": 3.7395234245058164, + "tokens_seen": 840433664 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003764393179538616, + "loss": 1.4612, + "theoretical_loss": 3.7394949945235814, + "tokens_seen": 840499200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003764292878635908, + "loss": 1.558, + "theoretical_loss": 3.7394665673786607, + "tokens_seen": 840564736 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037641925777331996, + "loss": 1.5617, + "theoretical_loss": 3.739438143070551, + "tokens_seen": 840630272 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037640922768304914, + "loss": 1.5273, + "theoretical_loss": 3.739409721598747, + "tokens_seen": 840695808 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003763991975927784, + "loss": 1.462, + "theoretical_loss": 3.739381302962745, + "tokens_seen": 840761344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003763891675025075, + "loss": 1.3882, + "theoretical_loss": 3.739352887162041, + "tokens_seen": 840826880 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037637913741223674, + "loss": 1.5101, + "theoretical_loss": 3.739324474196132, + "tokens_seen": 840892416 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037636910732196586, + "loss": 1.3432, + "theoretical_loss": 3.739296064064513, + "tokens_seen": 840957952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003763590772316951, + "loss": 1.5043, + "theoretical_loss": 3.739267656766681, + "tokens_seen": 841023488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003763490471414243, + "loss": 1.5412, + "theoretical_loss": 3.7392392523021325, + "tokens_seen": 841089024 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037633901705115346, + "loss": 1.5082, + "theoretical_loss": 3.7392108506703647, + "tokens_seen": 841154560 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037632898696088264, + "loss": 1.5134, + "theoretical_loss": 3.7391824518708736, + "tokens_seen": 841220096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0003763189568706119, + "loss": 1.4778, + "theoretical_loss": 3.739154055903157, + "tokens_seen": 841285632 + }, + { + "epoch": 0.25, + "learning_rate": 0.000376308926780341, + "loss": 1.4611, + "theoretical_loss": 3.7391256627667104, + "tokens_seen": 841351168 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037629889669007024, + "loss": 1.5058, + "theoretical_loss": 3.7390972724610334, + "tokens_seen": 841416704 + }, + { + "epoch": 0.25, + "learning_rate": 0.00037628886659979937, + "loss": 1.3955, + "theoretical_loss": 3.7390688849856213, + "tokens_seen": 841482240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003762788365095286, + "loss": 1.4274, + "theoretical_loss": 3.7390405003399723, + "tokens_seen": 841547776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003762688064192578, + "loss": 1.4696, + "theoretical_loss": 3.7390121185235836, + "tokens_seen": 841613312 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037625877632898696, + "loss": 1.4892, + "theoretical_loss": 3.7389837395359535, + "tokens_seen": 841678848 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037624874623871615, + "loss": 1.5135, + "theoretical_loss": 3.7389553633765797, + "tokens_seen": 841744384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003762387161484453, + "loss": 1.435, + "theoretical_loss": 3.73892699004496, + "tokens_seen": 841809920 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003762286860581745, + "loss": 1.4224, + "theoretical_loss": 3.7388986195405924, + "tokens_seen": 841875456 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037621865596790374, + "loss": 1.3808, + "theoretical_loss": 3.7388702518629753, + "tokens_seen": 841940992 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037620862587763287, + "loss": 1.4978, + "theoretical_loss": 3.738841887011607, + "tokens_seen": 842006528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003761985957873621, + "loss": 1.4933, + "theoretical_loss": 3.7388135249859866, + "tokens_seen": 842072064 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 310992, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.382948398590088, + "objective/train/theoretical_loss": 3.738785165785611, + "objective/train/tokens_used": 862597600, + "theoretical_loss": 3.738785165785611, + "tokens_seen": 842137600 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037618856569709123, + "loss": 1.4744, + "theoretical_loss": 3.738785165785611, + "tokens_seen": 842137600 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037617853560682047, + "loss": 1.4418, + "theoretical_loss": 3.7387568094099812, + "tokens_seen": 842203136 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037616850551654965, + "loss": 1.545, + "theoretical_loss": 3.738728455858594, + "tokens_seen": 842268672 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037615847542627883, + "loss": 1.5144, + "theoretical_loss": 3.73870010513095, + "tokens_seen": 842334208 + }, + { + "epoch": 0.26, + "learning_rate": 0.000376148445336008, + "loss": 1.5375, + "theoretical_loss": 3.7386717572265473, + "tokens_seen": 842399744 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037613841524573725, + "loss": 1.5562, + "theoretical_loss": 3.738643412144885, + "tokens_seen": 842465280 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037612838515546637, + "loss": 1.5368, + "theoretical_loss": 3.738615069885464, + "tokens_seen": 842530816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003761183550651956, + "loss": 1.4921, + "theoretical_loss": 3.738586730447782, + "tokens_seen": 842596352 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003761083249749248, + "loss": 1.4443, + "theoretical_loss": 3.7385583938313394, + "tokens_seen": 842661888 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037609829488465397, + "loss": 1.4599, + "theoretical_loss": 3.738530060035636, + "tokens_seen": 842727424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003760882647943832, + "loss": 1.5036, + "theoretical_loss": 3.7385017290601716, + "tokens_seen": 842792960 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037607823470411233, + "loss": 1.4486, + "theoretical_loss": 3.7384734009044465, + "tokens_seen": 842858496 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037606820461384157, + "loss": 1.5252, + "theoretical_loss": 3.7384450755679604, + "tokens_seen": 842924032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003760581745235707, + "loss": 1.4941, + "theoretical_loss": 3.7384167530502133, + "tokens_seen": 842989568 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037604814443329993, + "loss": 1.4776, + "theoretical_loss": 3.738388433350707, + "tokens_seen": 843055104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003760381143430291, + "loss": 1.3821, + "theoretical_loss": 3.7383601164689404, + "tokens_seen": 843120640 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003760280842527583, + "loss": 1.4779, + "theoretical_loss": 3.7383318024044145, + "tokens_seen": 843186176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003760180541624875, + "loss": 1.5293, + "theoretical_loss": 3.738303491156631, + "tokens_seen": 843251712 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037600802407221665, + "loss": 1.5051, + "theoretical_loss": 3.73827518272509, + "tokens_seen": 843317248 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037599799398194584, + "loss": 1.5426, + "theoretical_loss": 3.738246877109293, + "tokens_seen": 843382784 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037598796389167507, + "loss": 1.4385, + "theoretical_loss": 3.7382185743087404, + "tokens_seen": 843448320 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003759779338014042, + "loss": 1.5643, + "theoretical_loss": 3.7381902743229345, + "tokens_seen": 843513856 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037596790371113343, + "loss": 1.5587, + "theoretical_loss": 3.7381619771513757, + "tokens_seen": 843579392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003759578736208626, + "loss": 1.4512, + "theoretical_loss": 3.738133682793566, + "tokens_seen": 843644928 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003759478435305918, + "loss": 1.5451, + "theoretical_loss": 3.7381053912490074, + "tokens_seen": 843710464 + }, + { + "epoch": 0.26, + "learning_rate": 0.000375937813440321, + "loss": 1.5022, + "theoretical_loss": 3.7380771025172015, + "tokens_seen": 843776000 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037592778335005016, + "loss": 1.4347, + "theoretical_loss": 3.73804881659765, + "tokens_seen": 843841536 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037591775325977934, + "loss": 1.4756, + "theoretical_loss": 3.738020533489855, + "tokens_seen": 843907072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003759077231695086, + "loss": 1.4997, + "theoretical_loss": 3.737992253193319, + "tokens_seen": 843972608 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003758976930792377, + "loss": 1.4769, + "theoretical_loss": 3.737963975707544, + "tokens_seen": 844038144 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037588766298896694, + "loss": 1.4809, + "theoretical_loss": 3.7379357010320318, + "tokens_seen": 844103680 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037587763289869606, + "loss": 1.4047, + "theoretical_loss": 3.7379074291662864, + "tokens_seen": 844169216 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003758676028084253, + "loss": 1.426, + "theoretical_loss": 3.7378791601098094, + "tokens_seen": 844234752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003758575727181545, + "loss": 1.4386, + "theoretical_loss": 3.737850893862104, + "tokens_seen": 844300288 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037584754262788366, + "loss": 1.5223, + "theoretical_loss": 3.7378226304226727, + "tokens_seen": 844365824 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037583751253761284, + "loss": 1.39, + "theoretical_loss": 3.737794369791019, + "tokens_seen": 844431360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003758274824473421, + "loss": 1.5202, + "theoretical_loss": 3.7377661119666463, + "tokens_seen": 844496896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003758174523570712, + "loss": 1.6141, + "theoretical_loss": 3.7377378569490576, + "tokens_seen": 844562432 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037580742226680044, + "loss": 1.4737, + "theoretical_loss": 3.7377096047377565, + "tokens_seen": 844627968 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037579739217652957, + "loss": 1.4677, + "theoretical_loss": 3.7376813553322465, + "tokens_seen": 844693504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003757873620862588, + "loss": 1.5696, + "theoretical_loss": 3.737653108732031, + "tokens_seen": 844759040 + }, + { + "epoch": 0.26, + "learning_rate": 0.000375777331995988, + "loss": 1.5719, + "theoretical_loss": 3.7376248649366137, + "tokens_seen": 844824576 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037576730190571716, + "loss": 1.4085, + "theoretical_loss": 3.737596623945499, + "tokens_seen": 844890112 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037575727181544635, + "loss": 1.4421, + "theoretical_loss": 3.737568385758191, + "tokens_seen": 844955648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003757472417251755, + "loss": 1.4616, + "theoretical_loss": 3.737540150374194, + "tokens_seen": 845021184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003757372116349047, + "loss": 1.5162, + "theoretical_loss": 3.737511917793012, + "tokens_seen": 845086720 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037572718154463394, + "loss": 1.4445, + "theoretical_loss": 3.7374836880141493, + "tokens_seen": 845152256 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037571715145436307, + "loss": 1.5297, + "theoretical_loss": 3.737455461037111, + "tokens_seen": 845217792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003757071213640923, + "loss": 1.4731, + "theoretical_loss": 3.737427236861401, + "tokens_seen": 845283328 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037569709127382143, + "loss": 1.4606, + "theoretical_loss": 3.7373990154865253, + "tokens_seen": 845348864 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 312167, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4602057933807373, + "objective/train/theoretical_loss": 3.7373707969119883, + "objective/train/tokens_used": 865874400, + "theoretical_loss": 3.7373707969119883, + "tokens_seen": 845414400 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037568706118355067, + "loss": 1.4581, + "theoretical_loss": 3.7373707969119883, + "tokens_seen": 845414400 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037567703109327985, + "loss": 1.5532, + "theoretical_loss": 3.7373425811372942, + "tokens_seen": 845479936 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037566700100300903, + "loss": 1.5866, + "theoretical_loss": 3.7373143681619494, + "tokens_seen": 845545472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003756569709127382, + "loss": 1.4126, + "theoretical_loss": 3.737286157985459, + "tokens_seen": 845611008 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037564694082246745, + "loss": 1.5132, + "theoretical_loss": 3.737257950607328, + "tokens_seen": 845676544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003756369107321966, + "loss": 1.4864, + "theoretical_loss": 3.7372297460270625, + "tokens_seen": 845742080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003756268806419258, + "loss": 1.5227, + "theoretical_loss": 3.7372015442441677, + "tokens_seen": 845807616 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037561685055165494, + "loss": 1.423, + "theoretical_loss": 3.7371733452581504, + "tokens_seen": 845873152 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037560682046138417, + "loss": 1.3821, + "theoretical_loss": 3.737145149068515, + "tokens_seen": 845938688 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037559679037111335, + "loss": 1.4924, + "theoretical_loss": 3.737116955674769, + "tokens_seen": 846004224 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037558676028084253, + "loss": 1.4492, + "theoretical_loss": 3.7370887650764177, + "tokens_seen": 846069760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755767301905717, + "loss": 1.4815, + "theoretical_loss": 3.7370605772729677, + "tokens_seen": 846135296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755667001003009, + "loss": 1.4548, + "theoretical_loss": 3.7370323922639264, + "tokens_seen": 846200832 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755566700100301, + "loss": 1.3801, + "theoretical_loss": 3.7370042100487986, + "tokens_seen": 846266368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755466399197593, + "loss": 1.5033, + "theoretical_loss": 3.7369760306270923, + "tokens_seen": 846331904 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037553660982948844, + "loss": 1.3914, + "theoretical_loss": 3.7369478539983145, + "tokens_seen": 846397440 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755265797392177, + "loss": 1.4345, + "theoretical_loss": 3.736919680161971, + "tokens_seen": 846462976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003755165496489468, + "loss": 1.4919, + "theoretical_loss": 3.7368915091175703, + "tokens_seen": 846528512 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037550651955867604, + "loss": 1.501, + "theoretical_loss": 3.736863340864618, + "tokens_seen": 846594048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754964894684052, + "loss": 1.5816, + "theoretical_loss": 3.7368351754026228, + "tokens_seen": 846659584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754864593781344, + "loss": 1.4541, + "theoretical_loss": 3.7368070127310915, + "tokens_seen": 846725120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754764292878636, + "loss": 1.4545, + "theoretical_loss": 3.736778852849532, + "tokens_seen": 846790656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754663991975928, + "loss": 1.5363, + "theoretical_loss": 3.736750695757452, + "tokens_seen": 846856192 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037545636910732194, + "loss": 1.4549, + "theoretical_loss": 3.736722541454359, + "tokens_seen": 846921728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754463390170512, + "loss": 1.5233, + "theoretical_loss": 3.7366943899397613, + "tokens_seen": 846987264 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754363089267803, + "loss": 1.3905, + "theoretical_loss": 3.736666241213167, + "tokens_seen": 847052800 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037542627883650954, + "loss": 1.5106, + "theoretical_loss": 3.736638095274084, + "tokens_seen": 847118336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754162487462387, + "loss": 1.4823, + "theoretical_loss": 3.736609952122021, + "tokens_seen": 847183872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003754062186559679, + "loss": 1.3768, + "theoretical_loss": 3.736581811756486, + "tokens_seen": 847249408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003753961885656971, + "loss": 1.373, + "theoretical_loss": 3.736553674176988, + "tokens_seen": 847314944 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037538615847542626, + "loss": 1.461, + "theoretical_loss": 3.736525539383036, + "tokens_seen": 847380480 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037537612838515544, + "loss": 1.3629, + "theoretical_loss": 3.736497407374138, + "tokens_seen": 847446016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003753660982948847, + "loss": 1.4204, + "theoretical_loss": 3.7364692781498032, + "tokens_seen": 847511552 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037535606820461386, + "loss": 1.4332, + "theoretical_loss": 3.7364411517095415, + "tokens_seen": 847577088 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037534603811434304, + "loss": 1.5443, + "theoretical_loss": 3.736413028052861, + "tokens_seen": 847642624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003753360080240723, + "loss": 1.4944, + "theoretical_loss": 3.736384907179272, + "tokens_seen": 847708160 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003753259779338014, + "loss": 1.4878, + "theoretical_loss": 3.736356789088283, + "tokens_seen": 847773696 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037531594784353064, + "loss": 1.4185, + "theoretical_loss": 3.7363286737794046, + "tokens_seen": 847839232 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037530591775325977, + "loss": 1.5571, + "theoretical_loss": 3.7363005612521456, + "tokens_seen": 847904768 + }, + { + "epoch": 0.26, + "learning_rate": 0.000375295887662989, + "loss": 1.4298, + "theoretical_loss": 3.736272451506016, + "tokens_seen": 847970304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003752858575727182, + "loss": 1.4949, + "theoretical_loss": 3.7362443445405265, + "tokens_seen": 848035840 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037527582748244736, + "loss": 1.451, + "theoretical_loss": 3.736216240355186, + "tokens_seen": 848101376 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037526579739217655, + "loss": 1.5554, + "theoretical_loss": 3.7361881389495064, + "tokens_seen": 848166912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003752557673019057, + "loss": 1.5479, + "theoretical_loss": 3.736160040322996, + "tokens_seen": 848232448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003752457372116349, + "loss": 1.4214, + "theoretical_loss": 3.736131944475167, + "tokens_seen": 848297984 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037523570712136414, + "loss": 1.5465, + "theoretical_loss": 3.7361038514055287, + "tokens_seen": 848363520 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037522567703109327, + "loss": 1.4175, + "theoretical_loss": 3.7360757611135926, + "tokens_seen": 848429056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003752156469408225, + "loss": 1.3704, + "theoretical_loss": 3.7360476735988692, + "tokens_seen": 848494592 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037520561685055163, + "loss": 1.4438, + "theoretical_loss": 3.736019588860869, + "tokens_seen": 848560128 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037519558676028087, + "loss": 1.4823, + "theoretical_loss": 3.7359915068991043, + "tokens_seen": 848625664 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 313372, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.508387565612793, + "objective/train/theoretical_loss": 3.7359634277130853, + "objective/train/tokens_used": 869151200, + "theoretical_loss": 3.7359634277130853, + "tokens_seen": 848691200 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037518555667001005, + "loss": 1.4622, + "theoretical_loss": 3.7359634277130853, + "tokens_seen": 848691200 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037517552657973923, + "loss": 1.5832, + "theoretical_loss": 3.7359353513023237, + "tokens_seen": 848756736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003751654964894684, + "loss": 1.4243, + "theoretical_loss": 3.7359072776663305, + "tokens_seen": 848822272 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037515546639919765, + "loss": 1.4045, + "theoretical_loss": 3.735879206804618, + "tokens_seen": 848887808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003751454363089268, + "loss": 1.5059, + "theoretical_loss": 3.735851138716697, + "tokens_seen": 848953344 + }, + { + "epoch": 0.26, + "learning_rate": 0.000375135406218656, + "loss": 1.4822, + "theoretical_loss": 3.7358230734020803, + "tokens_seen": 849018880 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037512537612838514, + "loss": 1.3802, + "theoretical_loss": 3.7357950108602793, + "tokens_seen": 849084416 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037511534603811437, + "loss": 1.4406, + "theoretical_loss": 3.735766951090806, + "tokens_seen": 849149952 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037510531594784355, + "loss": 1.5004, + "theoretical_loss": 3.735738894093173, + "tokens_seen": 849215488 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037509528585757273, + "loss": 1.6133, + "theoretical_loss": 3.7357108398668917, + "tokens_seen": 849281024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750852557673019, + "loss": 1.5157, + "theoretical_loss": 3.7356827884114754, + "tokens_seen": 849346560 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750752256770311, + "loss": 1.6399, + "theoretical_loss": 3.735654739726437, + "tokens_seen": 849412096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750651955867603, + "loss": 1.4853, + "theoretical_loss": 3.735626693811288, + "tokens_seen": 849477632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750551654964895, + "loss": 1.5105, + "theoretical_loss": 3.7355986506655414, + "tokens_seen": 849543168 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037504513540621864, + "loss": 1.4074, + "theoretical_loss": 3.735570610288711, + "tokens_seen": 849608704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750351053159479, + "loss": 1.552, + "theoretical_loss": 3.73554257268031, + "tokens_seen": 849674240 + }, + { + "epoch": 0.26, + "learning_rate": 0.000375025075225677, + "loss": 1.4004, + "theoretical_loss": 3.73551453783985, + "tokens_seen": 849739776 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037501504513540624, + "loss": 1.3766, + "theoretical_loss": 3.7354865057668456, + "tokens_seen": 849805312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003750050150451354, + "loss": 1.4978, + "theoretical_loss": 3.73545847646081, + "tokens_seen": 849870848 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749949849548646, + "loss": 1.419, + "theoretical_loss": 3.735430449921256, + "tokens_seen": 849936384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749849548645938, + "loss": 1.3672, + "theoretical_loss": 3.735402426147698, + "tokens_seen": 850001920 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374974924774323, + "loss": 1.3761, + "theoretical_loss": 3.7353744051396505, + "tokens_seen": 850067456 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037496489468405214, + "loss": 1.3791, + "theoretical_loss": 3.7353463868966257, + "tokens_seen": 850132992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749548645937814, + "loss": 1.5112, + "theoretical_loss": 3.7353183714181393, + "tokens_seen": 850198528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749448345035105, + "loss": 1.5822, + "theoretical_loss": 3.735290358703704, + "tokens_seen": 850264064 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037493480441323974, + "loss": 1.4194, + "theoretical_loss": 3.735262348752835, + "tokens_seen": 850329600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749247743229689, + "loss": 1.5244, + "theoretical_loss": 3.7352343415650466, + "tokens_seen": 850395136 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749147442326981, + "loss": 1.5497, + "theoretical_loss": 3.7352063371398527, + "tokens_seen": 850460672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003749047141424273, + "loss": 1.4473, + "theoretical_loss": 3.7351783354767685, + "tokens_seen": 850526208 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037489468405215646, + "loss": 1.4569, + "theoretical_loss": 3.735150336575309, + "tokens_seen": 850591744 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037488465396188564, + "loss": 1.503, + "theoretical_loss": 3.7351223404349887, + "tokens_seen": 850657280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003748746238716149, + "loss": 1.554, + "theoretical_loss": 3.7350943470553224, + "tokens_seen": 850722816 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374864593781344, + "loss": 1.3705, + "theoretical_loss": 3.735066356435826, + "tokens_seen": 850788352 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037485456369107324, + "loss": 1.6047, + "theoretical_loss": 3.735038368576014, + "tokens_seen": 850853888 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037484453360080237, + "loss": 1.5747, + "theoretical_loss": 3.7350103834754025, + "tokens_seen": 850919424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003748345035105316, + "loss": 1.506, + "theoretical_loss": 3.734982401133506, + "tokens_seen": 850984960 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003748244734202608, + "loss": 1.4859, + "theoretical_loss": 3.734954421549841, + "tokens_seen": 851050496 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037481444332998997, + "loss": 1.3288, + "theoretical_loss": 3.734926444723923, + "tokens_seen": 851116032 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037480441323971915, + "loss": 1.4108, + "theoretical_loss": 3.734898470655268, + "tokens_seen": 851181568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003747943831494484, + "loss": 1.4042, + "theoretical_loss": 3.7348704993433914, + "tokens_seen": 851247104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003747843530591775, + "loss": 1.4951, + "theoretical_loss": 3.7348425307878097, + "tokens_seen": 851312640 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037477432296890675, + "loss": 1.4636, + "theoretical_loss": 3.7348145649880395, + "tokens_seen": 851378176 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037476429287863587, + "loss": 1.4608, + "theoretical_loss": 3.734786601943597, + "tokens_seen": 851443712 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003747542627883651, + "loss": 1.3187, + "theoretical_loss": 3.7347586416539986, + "tokens_seen": 851509248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003747442326980943, + "loss": 1.4541, + "theoretical_loss": 3.7347306841187606, + "tokens_seen": 851574784 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037473420260782347, + "loss": 1.5602, + "theoretical_loss": 3.7347027293373998, + "tokens_seen": 851640320 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037472417251755265, + "loss": 1.4379, + "theoretical_loss": 3.7346747773094338, + "tokens_seen": 851705856 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037471414242728183, + "loss": 1.4816, + "theoretical_loss": 3.7346468280343785, + "tokens_seen": 851771392 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374704112337011, + "loss": 1.4326, + "theoretical_loss": 3.734618881511752, + "tokens_seen": 851836928 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037469408224674025, + "loss": 1.4242, + "theoretical_loss": 3.7345909377410704, + "tokens_seen": 851902464 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 314587, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4998879432678223, + "objective/train/theoretical_loss": 3.7345629967218517, + "objective/train/tokens_used": 872428000, + "theoretical_loss": 3.7345629967218517, + "tokens_seen": 851968000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003746840521564694, + "loss": 1.4834, + "theoretical_loss": 3.7345629967218517, + "tokens_seen": 851968000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003746740220661986, + "loss": 1.4891, + "theoretical_loss": 3.7345350584536137, + "tokens_seen": 852033536 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037466399197592774, + "loss": 1.5874, + "theoretical_loss": 3.7345071229358733, + "tokens_seen": 852099072 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374653961885657, + "loss": 1.4983, + "theoretical_loss": 3.7344791901681487, + "tokens_seen": 852164608 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037464393179538615, + "loss": 1.4495, + "theoretical_loss": 3.734451260149957, + "tokens_seen": 852230144 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037463390170511534, + "loss": 1.4956, + "theoretical_loss": 3.7344233328808176, + "tokens_seen": 852295680 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037462387161484457, + "loss": 1.4838, + "theoretical_loss": 3.7343954083602466, + "tokens_seen": 852361216 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037461384152457375, + "loss": 1.5418, + "theoretical_loss": 3.734367486587764, + "tokens_seen": 852426752 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037460381143430293, + "loss": 1.4532, + "theoretical_loss": 3.7343395675628868, + "tokens_seen": 852492288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745937813440321, + "loss": 1.4182, + "theoretical_loss": 3.734311651285134, + "tokens_seen": 852557824 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745837512537613, + "loss": 1.4344, + "theoretical_loss": 3.7342837377540246, + "tokens_seen": 852623360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745737211634905, + "loss": 1.4809, + "theoretical_loss": 3.7342558269690764, + "tokens_seen": 852688896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745636910732197, + "loss": 1.4069, + "theoretical_loss": 3.7342279189298093, + "tokens_seen": 852754432 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037455366098294884, + "loss": 1.5904, + "theoretical_loss": 3.7342000136357405, + "tokens_seen": 852819968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745436308926781, + "loss": 1.4904, + "theoretical_loss": 3.734172111086391, + "tokens_seen": 852885504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745336008024072, + "loss": 1.5148, + "theoretical_loss": 3.7341442112812784, + "tokens_seen": 852951040 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037452357071213644, + "loss": 1.5024, + "theoretical_loss": 3.7341163142199223, + "tokens_seen": 853016576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745135406218656, + "loss": 1.4436, + "theoretical_loss": 3.734088419901843, + "tokens_seen": 853082112 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003745035105315948, + "loss": 1.5873, + "theoretical_loss": 3.734060528326559, + "tokens_seen": 853147648 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374493480441324, + "loss": 1.3891, + "theoretical_loss": 3.7340326394935905, + "tokens_seen": 853213184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744834503510532, + "loss": 1.4643, + "theoretical_loss": 3.734004753402457, + "tokens_seen": 853278720 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037447342026078234, + "loss": 1.443, + "theoretical_loss": 3.7339768700526785, + "tokens_seen": 853344256 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744633901705116, + "loss": 1.5088, + "theoretical_loss": 3.7339489894437756, + "tokens_seen": 853409792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744533600802407, + "loss": 1.4896, + "theoretical_loss": 3.733921111575267, + "tokens_seen": 853475328 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037444332998996994, + "loss": 1.5127, + "theoretical_loss": 3.7338932364466744, + "tokens_seen": 853540864 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744332998996991, + "loss": 1.4026, + "theoretical_loss": 3.7338653640575172, + "tokens_seen": 853606400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744232698094283, + "loss": 1.4463, + "theoretical_loss": 3.7338374944073163, + "tokens_seen": 853671936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003744132397191575, + "loss": 1.4099, + "theoretical_loss": 3.7338096274955923, + "tokens_seen": 853737472 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037440320962888666, + "loss": 1.5558, + "theoretical_loss": 3.733781763321866, + "tokens_seen": 853803008 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037439317953861585, + "loss": 1.5487, + "theoretical_loss": 3.7337539018856587, + "tokens_seen": 853868544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003743831494483451, + "loss": 1.524, + "theoretical_loss": 3.73372604318649, + "tokens_seen": 853934080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003743731193580742, + "loss": 1.5648, + "theoretical_loss": 3.733698187223882, + "tokens_seen": 853999616 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037436308926780344, + "loss": 1.4437, + "theoretical_loss": 3.733670333997356, + "tokens_seen": 854065152 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037435305917753257, + "loss": 1.5364, + "theoretical_loss": 3.733642483506433, + "tokens_seen": 854130688 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003743430290872618, + "loss": 1.3992, + "theoretical_loss": 3.7336146357506337, + "tokens_seen": 854196224 + }, + { + "epoch": 0.26, + "learning_rate": 0.000374332998996991, + "loss": 1.5234, + "theoretical_loss": 3.7335867907294817, + "tokens_seen": 854261760 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037432296890672017, + "loss": 1.4533, + "theoretical_loss": 3.733558948442497, + "tokens_seen": 854327296 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037431293881644935, + "loss": 1.4817, + "theoretical_loss": 3.733531108889202, + "tokens_seen": 854392832 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003743029087261786, + "loss": 1.3937, + "theoretical_loss": 3.733503272069118, + "tokens_seen": 854458368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003742928786359077, + "loss": 1.3624, + "theoretical_loss": 3.7334754379817685, + "tokens_seen": 854523904 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037428284854563695, + "loss": 1.4699, + "theoretical_loss": 3.733447606626674, + "tokens_seen": 854589440 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037427281845536607, + "loss": 1.4378, + "theoretical_loss": 3.7334197780033582, + "tokens_seen": 854654976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003742627883650953, + "loss": 1.3398, + "theoretical_loss": 3.733391952111343, + "tokens_seen": 854720512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003742527582748245, + "loss": 1.4919, + "theoretical_loss": 3.73336412895015, + "tokens_seen": 854786048 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037424272818455367, + "loss": 1.5483, + "theoretical_loss": 3.733336308519303, + "tokens_seen": 854851584 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037423269809428285, + "loss": 1.5048, + "theoretical_loss": 3.7333084908183247, + "tokens_seen": 854917120 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037422266800401203, + "loss": 1.5462, + "theoretical_loss": 3.7332806758467383, + "tokens_seen": 854982656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003742126379137412, + "loss": 1.4792, + "theoretical_loss": 3.7332528636040654, + "tokens_seen": 855048192 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037420260782347045, + "loss": 1.4598, + "theoretical_loss": 3.7332250540898304, + "tokens_seen": 855113728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741925777331996, + "loss": 1.4768, + "theoretical_loss": 3.733197247303556, + "tokens_seen": 855179264 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 315806, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3750395774841309, + "objective/train/theoretical_loss": 3.733169443244766, + "objective/train/tokens_used": 875704800, + "theoretical_loss": 3.733169443244766, + "tokens_seen": 855244800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741825476429288, + "loss": 1.4213, + "theoretical_loss": 3.733169443244766, + "tokens_seen": 855244800 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037417251755265794, + "loss": 1.4688, + "theoretical_loss": 3.7331416419129835, + "tokens_seen": 855310336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741624874623872, + "loss": 1.4838, + "theoretical_loss": 3.7331138433077324, + "tokens_seen": 855375872 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037415245737211635, + "loss": 1.5282, + "theoretical_loss": 3.733086047428536, + "tokens_seen": 855441408 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037414242728184554, + "loss": 1.3481, + "theoretical_loss": 3.733058254274919, + "tokens_seen": 855506944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741323971915747, + "loss": 1.3845, + "theoretical_loss": 3.7330304638464042, + "tokens_seen": 855572480 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037412236710130395, + "loss": 1.5371, + "theoretical_loss": 3.733002676142517, + "tokens_seen": 855638016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741123370110331, + "loss": 1.4247, + "theoretical_loss": 3.7329748911627805, + "tokens_seen": 855703552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003741023069207623, + "loss": 1.2932, + "theoretical_loss": 3.7329471089067194, + "tokens_seen": 855769088 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037409227683049144, + "loss": 1.4356, + "theoretical_loss": 3.7329193293738587, + "tokens_seen": 855834624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740822467402207, + "loss": 1.446, + "theoretical_loss": 3.7328915525637227, + "tokens_seen": 855900160 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037407221664994986, + "loss": 1.4682, + "theoretical_loss": 3.7328637784758354, + "tokens_seen": 855965696 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037406218655967904, + "loss": 1.5692, + "theoretical_loss": 3.7328360071097224, + "tokens_seen": 856031232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740521564694082, + "loss": 1.5254, + "theoretical_loss": 3.7328082384649086, + "tokens_seen": 856096768 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740421263791374, + "loss": 1.4634, + "theoretical_loss": 3.7327804725409184, + "tokens_seen": 856162304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740320962888666, + "loss": 1.4675, + "theoretical_loss": 3.7327527093372774, + "tokens_seen": 856227840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740220661985958, + "loss": 1.4498, + "theoretical_loss": 3.7327249488535106, + "tokens_seen": 856293376 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037401203610832494, + "loss": 1.3583, + "theoretical_loss": 3.7326971910891444, + "tokens_seen": 856358912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003740020060180542, + "loss": 1.412, + "theoretical_loss": 3.732669436043703, + "tokens_seen": 856424448 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037399197592778336, + "loss": 1.408, + "theoretical_loss": 3.7326416837167127, + "tokens_seen": 856489984 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037398194583751254, + "loss": 1.4103, + "theoretical_loss": 3.7326139341076994, + "tokens_seen": 856555520 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739719157472417, + "loss": 1.4444, + "theoretical_loss": 3.7325861872161887, + "tokens_seen": 856621056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739618856569709, + "loss": 1.3771, + "theoretical_loss": 3.732558443041707, + "tokens_seen": 856686592 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739518555667001, + "loss": 1.4766, + "theoretical_loss": 3.73253070158378, + "tokens_seen": 856752128 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739418254764293, + "loss": 1.4582, + "theoretical_loss": 3.732502962841934, + "tokens_seen": 856817664 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037393179538615845, + "loss": 1.3634, + "theoretical_loss": 3.732475226815695, + "tokens_seen": 856883200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739217652958877, + "loss": 1.469, + "theoretical_loss": 3.7324474935045906, + "tokens_seen": 856948736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003739117352056168, + "loss": 1.4593, + "theoretical_loss": 3.7324197629081466, + "tokens_seen": 857014272 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037390170511534605, + "loss": 1.4607, + "theoretical_loss": 3.7323920350258897, + "tokens_seen": 857079808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003738916750250752, + "loss": 1.4227, + "theoretical_loss": 3.732364309857347, + "tokens_seen": 857145344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003738816449348044, + "loss": 1.441, + "theoretical_loss": 3.7323365874020453, + "tokens_seen": 857210880 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037387161484453364, + "loss": 1.4423, + "theoretical_loss": 3.732308867659511, + "tokens_seen": 857276416 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037386158475426277, + "loss": 1.3895, + "theoretical_loss": 3.732281150629273, + "tokens_seen": 857341952 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373851554663992, + "loss": 1.5693, + "theoretical_loss": 3.732253436310857, + "tokens_seen": 857407488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003738415245737212, + "loss": 1.4614, + "theoretical_loss": 3.732225724703791, + "tokens_seen": 857473024 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037383149448345037, + "loss": 1.4617, + "theoretical_loss": 3.7321980158076027, + "tokens_seen": 857538560 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037382146439317955, + "loss": 1.4863, + "theoretical_loss": 3.7321703096218197, + "tokens_seen": 857604096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003738114343029088, + "loss": 1.4198, + "theoretical_loss": 3.7321426061459704, + "tokens_seen": 857669632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003738014042126379, + "loss": 1.5551, + "theoretical_loss": 3.7321149053795812, + "tokens_seen": 857735168 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037379137412236715, + "loss": 1.4487, + "theoretical_loss": 3.7320872073221807, + "tokens_seen": 857800704 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037378134403209627, + "loss": 1.4753, + "theoretical_loss": 3.732059511973298, + "tokens_seen": 857866240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003737713139418255, + "loss": 1.4062, + "theoretical_loss": 3.7320318193324598, + "tokens_seen": 857931776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003737612838515547, + "loss": 1.4489, + "theoretical_loss": 3.732004129399196, + "tokens_seen": 857997312 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037375125376128387, + "loss": 1.4784, + "theoretical_loss": 3.731976442173034, + "tokens_seen": 858062848 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037374122367101305, + "loss": 1.4535, + "theoretical_loss": 3.731948757653503, + "tokens_seen": 858128384 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037373119358074223, + "loss": 1.3594, + "theoretical_loss": 3.7319210758401318, + "tokens_seen": 858193920 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003737211634904714, + "loss": 1.3962, + "theoretical_loss": 3.7318933967324486, + "tokens_seen": 858259456 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037371113340020065, + "loss": 1.5237, + "theoretical_loss": 3.731865720329983, + "tokens_seen": 858324992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003737011033099298, + "loss": 1.5129, + "theoretical_loss": 3.731838046632264, + "tokens_seen": 858390528 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373691073219659, + "loss": 1.4551, + "theoretical_loss": 3.73181037563882, + "tokens_seen": 858456064 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 317142, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4829411506652832, + "objective/train/theoretical_loss": 3.7317827073491814, + "objective/train/tokens_used": 878981600, + "theoretical_loss": 3.7317827073491814, + "tokens_seen": 858521600 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037368104312938814, + "loss": 1.5026, + "theoretical_loss": 3.7317827073491814, + "tokens_seen": 858521600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003736710130391174, + "loss": 1.452, + "theoretical_loss": 3.7317550417628773, + "tokens_seen": 858587136 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037366098294884655, + "loss": 1.5678, + "theoretical_loss": 3.731727378879437, + "tokens_seen": 858652672 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037365095285857574, + "loss": 1.4015, + "theoretical_loss": 3.7316997186983905, + "tokens_seen": 858718208 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003736409227683049, + "loss": 1.3878, + "theoretical_loss": 3.7316720612192666, + "tokens_seen": 858783744 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037363089267803415, + "loss": 1.3947, + "theoretical_loss": 3.731644406441597, + "tokens_seen": 858849280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003736208625877633, + "loss": 1.6128, + "theoretical_loss": 3.7316167543649104, + "tokens_seen": 858914816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003736108324974925, + "loss": 1.4652, + "theoretical_loss": 3.7315891049887373, + "tokens_seen": 858980352 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037360080240722164, + "loss": 1.6068, + "theoretical_loss": 3.7315614583126084, + "tokens_seen": 859045888 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003735907723169509, + "loss": 1.5102, + "theoretical_loss": 3.731533814336053, + "tokens_seen": 859111424 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037358074222668006, + "loss": 1.4423, + "theoretical_loss": 3.7315061730586025, + "tokens_seen": 859176960 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037357071213640924, + "loss": 1.5639, + "theoretical_loss": 3.7314785344797867, + "tokens_seen": 859242496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003735606820461384, + "loss": 1.5302, + "theoretical_loss": 3.731450898599138, + "tokens_seen": 859308032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003735506519558676, + "loss": 1.4653, + "theoretical_loss": 3.731423265416185, + "tokens_seen": 859373568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003735406218655968, + "loss": 1.443, + "theoretical_loss": 3.7313956349304602, + "tokens_seen": 859439104 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373530591775326, + "loss": 1.4378, + "theoretical_loss": 3.7313680071414943, + "tokens_seen": 859504640 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037352056168505514, + "loss": 1.4143, + "theoretical_loss": 3.7313403820488187, + "tokens_seen": 859570176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003735105315947844, + "loss": 1.4146, + "theoretical_loss": 3.7313127596519644, + "tokens_seen": 859635712 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037350050150451356, + "loss": 1.4098, + "theoretical_loss": 3.7312851399504625, + "tokens_seen": 859701248 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037349047141424274, + "loss": 1.5943, + "theoretical_loss": 3.731257522943846, + "tokens_seen": 859766784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734804413239719, + "loss": 1.4642, + "theoretical_loss": 3.7312299086316445, + "tokens_seen": 859832320 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734704112337011, + "loss": 1.5301, + "theoretical_loss": 3.731202297013391, + "tokens_seen": 859897856 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734603811434303, + "loss": 1.4212, + "theoretical_loss": 3.7311746880886174, + "tokens_seen": 859963392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734503510531595, + "loss": 1.4416, + "theoretical_loss": 3.7311470818568555, + "tokens_seen": 860028928 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037344032096288865, + "loss": 1.4627, + "theoretical_loss": 3.7311194783176376, + "tokens_seen": 860094464 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734302908726179, + "loss": 1.5577, + "theoretical_loss": 3.7310918774704955, + "tokens_seen": 860160000 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373420260782347, + "loss": 1.5287, + "theoretical_loss": 3.731064279314962, + "tokens_seen": 860225536 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037341023069207625, + "loss": 1.3797, + "theoretical_loss": 3.7310366838505704, + "tokens_seen": 860291072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003734002006018054, + "loss": 1.4598, + "theoretical_loss": 3.7310090910768516, + "tokens_seen": 860356608 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003733901705115346, + "loss": 1.3794, + "theoretical_loss": 3.730981500993339, + "tokens_seen": 860422144 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003733801404212638, + "loss": 1.5352, + "theoretical_loss": 3.7309539135995657, + "tokens_seen": 860487680 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037337011033099297, + "loss": 1.3516, + "theoretical_loss": 3.7309263288950647, + "tokens_seen": 860553216 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037336008024072215, + "loss": 1.4359, + "theoretical_loss": 3.730898746879369, + "tokens_seen": 860618752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003733500501504514, + "loss": 1.6563, + "theoretical_loss": 3.730871167552011, + "tokens_seen": 860684288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003733400200601805, + "loss": 1.4708, + "theoretical_loss": 3.7308435909125253, + "tokens_seen": 860749824 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037332998996990975, + "loss": 1.5281, + "theoretical_loss": 3.7308160169604445, + "tokens_seen": 860815360 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037331995987963893, + "loss": 1.5936, + "theoretical_loss": 3.7307884456953024, + "tokens_seen": 860880896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003733099297893681, + "loss": 1.3945, + "theoretical_loss": 3.7307608771166327, + "tokens_seen": 860946432 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732998996990973, + "loss": 1.5068, + "theoretical_loss": 3.730733311223969, + "tokens_seen": 861011968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732898696088265, + "loss": 1.4729, + "theoretical_loss": 3.7307057480168453, + "tokens_seen": 861077504 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037327983951855565, + "loss": 1.4657, + "theoretical_loss": 3.730678187494796, + "tokens_seen": 861143040 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732698094282849, + "loss": 1.3734, + "theoretical_loss": 3.7306506296573536, + "tokens_seen": 861208576 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373259779338014, + "loss": 1.446, + "theoretical_loss": 3.7306230745040545, + "tokens_seen": 861274112 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037324974924774325, + "loss": 1.4965, + "theoretical_loss": 3.730595522034432, + "tokens_seen": 861339648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732397191574724, + "loss": 1.3628, + "theoretical_loss": 3.73056797224802, + "tokens_seen": 861405184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732296890672016, + "loss": 1.4784, + "theoretical_loss": 3.730540425144354, + "tokens_seen": 861470720 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003732196589769308, + "loss": 1.5023, + "theoretical_loss": 3.7305128807229684, + "tokens_seen": 861536256 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037320962888666, + "loss": 1.4771, + "theoretical_loss": 3.730485338983398, + "tokens_seen": 861601792 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037319959879638916, + "loss": 1.4155, + "theoretical_loss": 3.730457799925178, + "tokens_seen": 861667328 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037318956870611834, + "loss": 1.4153, + "theoretical_loss": 3.7304302635478432, + "tokens_seen": 861732864 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 318349, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3940141201019287, + "objective/train/theoretical_loss": 3.730402729850929, + "objective/train/tokens_used": 882258400, + "theoretical_loss": 3.730402729850929, + "tokens_seen": 861798400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731795386158475, + "loss": 1.4987, + "theoretical_loss": 3.730402729850929, + "tokens_seen": 861798400 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037316950852557675, + "loss": 1.6038, + "theoretical_loss": 3.7303751988339697, + "tokens_seen": 861863936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731594784353059, + "loss": 1.4058, + "theoretical_loss": 3.730347670496502, + "tokens_seen": 861929472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731494483450351, + "loss": 1.4752, + "theoretical_loss": 3.7303201448380605, + "tokens_seen": 861995008 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731394182547643, + "loss": 1.4804, + "theoretical_loss": 3.7302926218581813, + "tokens_seen": 862060544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731293881644935, + "loss": 1.3898, + "theoretical_loss": 3.730265101556401, + "tokens_seen": 862126080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003731193580742227, + "loss": 1.3839, + "theoretical_loss": 3.7302375839322535, + "tokens_seen": 862191616 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037310932798395184, + "loss": 1.4764, + "theoretical_loss": 3.7302100689852757, + "tokens_seen": 862257152 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003730992978936811, + "loss": 1.4257, + "theoretical_loss": 3.730182556715004, + "tokens_seen": 862322688 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037308926780341026, + "loss": 1.5427, + "theoretical_loss": 3.7301550471209746, + "tokens_seen": 862388224 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037307923771313944, + "loss": 1.5582, + "theoretical_loss": 3.730127540202723, + "tokens_seen": 862453760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003730692076228686, + "loss": 1.4555, + "theoretical_loss": 3.730100035959787, + "tokens_seen": 862519296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003730591775325978, + "loss": 1.5249, + "theoretical_loss": 3.730072534391702, + "tokens_seen": 862584832 + }, + { + "epoch": 0.26, + "learning_rate": 0.000373049147442327, + "loss": 1.4705, + "theoretical_loss": 3.730045035498005, + "tokens_seen": 862650368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003730391173520562, + "loss": 1.4833, + "theoretical_loss": 3.730017539278233, + "tokens_seen": 862715904 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037302908726178534, + "loss": 1.6243, + "theoretical_loss": 3.729990045731922, + "tokens_seen": 862781440 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003730190571715146, + "loss": 1.4511, + "theoretical_loss": 3.7299625548586106, + "tokens_seen": 862846976 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037300902708124376, + "loss": 1.4427, + "theoretical_loss": 3.729935066657835, + "tokens_seen": 862912512 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037299899699097294, + "loss": 1.5368, + "theoretical_loss": 3.7299075811291322, + "tokens_seen": 862978048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729889669007021, + "loss": 1.3992, + "theoretical_loss": 3.7298800982720404, + "tokens_seen": 863043584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729789368104313, + "loss": 1.5573, + "theoretical_loss": 3.7298526180860962, + "tokens_seen": 863109120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729689067201605, + "loss": 1.3559, + "theoretical_loss": 3.7298251405708376, + "tokens_seen": 863174656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729588766298897, + "loss": 1.4085, + "theoretical_loss": 3.729797665725802, + "tokens_seen": 863240192 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037294884653961885, + "loss": 1.4772, + "theoretical_loss": 3.729770193550528, + "tokens_seen": 863305728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729388164493481, + "loss": 1.365, + "theoretical_loss": 3.729742724044553, + "tokens_seen": 863371264 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729287863590772, + "loss": 1.444, + "theoretical_loss": 3.7297152572074146, + "tokens_seen": 863436800 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037291875626880645, + "loss": 1.4802, + "theoretical_loss": 3.729687793038652, + "tokens_seen": 863502336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003729087261785356, + "loss": 1.4923, + "theoretical_loss": 3.7296603315378025, + "tokens_seen": 863567872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003728986960882648, + "loss": 1.4527, + "theoretical_loss": 3.7296328727044052, + "tokens_seen": 863633408 + }, + { + "epoch": 0.26, + "learning_rate": 0.000372888665997994, + "loss": 1.4553, + "theoretical_loss": 3.729605416537998, + "tokens_seen": 863698944 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037287863590772317, + "loss": 1.4933, + "theoretical_loss": 3.7295779630381203, + "tokens_seen": 863764480 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037286860581745235, + "loss": 1.4739, + "theoretical_loss": 3.7295505122043098, + "tokens_seen": 863830016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003728585757271816, + "loss": 1.4166, + "theoretical_loss": 3.7295230640361066, + "tokens_seen": 863895552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003728485456369107, + "loss": 1.4818, + "theoretical_loss": 3.729495618533049, + "tokens_seen": 863961088 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037283851554663995, + "loss": 1.459, + "theoretical_loss": 3.7294681756946755, + "tokens_seen": 864026624 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037282848545636913, + "loss": 1.4394, + "theoretical_loss": 3.7294407355205257, + "tokens_seen": 864092160 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003728184553660983, + "loss": 1.556, + "theoretical_loss": 3.72941329801014, + "tokens_seen": 864157696 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003728084252758275, + "loss": 1.5332, + "theoretical_loss": 3.7293858631630563, + "tokens_seen": 864223232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727983951855567, + "loss": 1.4556, + "theoretical_loss": 3.7293584309788157, + "tokens_seen": 864288768 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037278836509528585, + "loss": 1.4866, + "theoretical_loss": 3.729331001456956, + "tokens_seen": 864354304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727783350050151, + "loss": 1.4755, + "theoretical_loss": 3.7293035745970182, + "tokens_seen": 864419840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727683049147442, + "loss": 1.4262, + "theoretical_loss": 3.729276150398542, + "tokens_seen": 864485376 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037275827482447345, + "loss": 1.4777, + "theoretical_loss": 3.7292487288610676, + "tokens_seen": 864550912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727482447342026, + "loss": 1.4266, + "theoretical_loss": 3.7292213099841343, + "tokens_seen": 864616448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727382146439318, + "loss": 1.4954, + "theoretical_loss": 3.7291938937672833, + "tokens_seen": 864681984 + }, + { + "epoch": 0.26, + "learning_rate": 0.000372728184553661, + "loss": 1.4233, + "theoretical_loss": 3.729166480210054, + "tokens_seen": 864747520 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003727181544633902, + "loss": 1.5819, + "theoretical_loss": 3.729139069311988, + "tokens_seen": 864813056 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037270812437311936, + "loss": 1.3286, + "theoretical_loss": 3.729111661072625, + "tokens_seen": 864878592 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037269809428284854, + "loss": 1.4582, + "theoretical_loss": 3.729084255491506, + "tokens_seen": 864944128 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726880641925777, + "loss": 1.4276, + "theoretical_loss": 3.7290568525681715, + "tokens_seen": 865009664 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 319421, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3921937942504883, + "objective/train/theoretical_loss": 3.7290294523021625, + "objective/train/tokens_used": 885535200, + "theoretical_loss": 3.7290294523021625, + "tokens_seen": 865075200 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037267803410230696, + "loss": 1.4347, + "theoretical_loss": 3.7290294523021625, + "tokens_seen": 865075200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726680040120361, + "loss": 1.4116, + "theoretical_loss": 3.729002054693021, + "tokens_seen": 865140736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726579739217653, + "loss": 1.5642, + "theoretical_loss": 3.7289746597402864, + "tokens_seen": 865206272 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726479438314945, + "loss": 1.5846, + "theoretical_loss": 3.7289472674435014, + "tokens_seen": 865271808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726379137412237, + "loss": 1.4771, + "theoretical_loss": 3.728919877802207, + "tokens_seen": 865337344 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037262788365095286, + "loss": 1.5109, + "theoretical_loss": 3.728892490815944, + "tokens_seen": 865402880 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037261785356068204, + "loss": 1.413, + "theoretical_loss": 3.7288651064842546, + "tokens_seen": 865468416 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003726078234704112, + "loss": 1.554, + "theoretical_loss": 3.728837724806681, + "tokens_seen": 865533952 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037259779338014046, + "loss": 1.4922, + "theoretical_loss": 3.7288103457827635, + "tokens_seen": 865599488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003725877632898696, + "loss": 1.5292, + "theoretical_loss": 3.7287829694120456, + "tokens_seen": 865665024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003725777331995988, + "loss": 1.3365, + "theoretical_loss": 3.728755595694069, + "tokens_seen": 865730560 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037256770310932795, + "loss": 1.5014, + "theoretical_loss": 3.7287282246283757, + "tokens_seen": 865796096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003725576730190572, + "loss": 1.5467, + "theoretical_loss": 3.728700856214507, + "tokens_seen": 865861632 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037254764292878636, + "loss": 1.4789, + "theoretical_loss": 3.7286734904520076, + "tokens_seen": 865927168 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037253761283851554, + "loss": 1.5251, + "theoretical_loss": 3.7286461273404177, + "tokens_seen": 865992704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003725275827482447, + "loss": 1.4341, + "theoretical_loss": 3.728618766879281, + "tokens_seen": 866058240 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037251755265797396, + "loss": 1.4261, + "theoretical_loss": 3.72859140906814, + "tokens_seen": 866123776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003725075225677031, + "loss": 1.3516, + "theoretical_loss": 3.7285640539065374, + "tokens_seen": 866189312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003724974924774323, + "loss": 1.4587, + "theoretical_loss": 3.7285367013940167, + "tokens_seen": 866254848 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037248746238716145, + "loss": 1.4914, + "theoretical_loss": 3.7285093515301213, + "tokens_seen": 866320384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003724774322968907, + "loss": 1.5136, + "theoretical_loss": 3.728482004314393, + "tokens_seen": 866385920 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037246740220661987, + "loss": 1.3173, + "theoretical_loss": 3.728454659746376, + "tokens_seen": 866451456 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037245737211634905, + "loss": 1.4534, + "theoretical_loss": 3.7284273178256133, + "tokens_seen": 866516992 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037244734202607823, + "loss": 1.4772, + "theoretical_loss": 3.728399978551649, + "tokens_seen": 866582528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003724373119358074, + "loss": 1.5103, + "theoretical_loss": 3.7283726419240266, + "tokens_seen": 866648064 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003724272818455366, + "loss": 1.5276, + "theoretical_loss": 3.7283453079422895, + "tokens_seen": 866713600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003724172517552658, + "loss": 1.4281, + "theoretical_loss": 3.7283179766059824, + "tokens_seen": 866779136 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037240722166499495, + "loss": 1.4869, + "theoretical_loss": 3.7282906479146476, + "tokens_seen": 866844672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723971915747242, + "loss": 1.4706, + "theoretical_loss": 3.728263321867831, + "tokens_seen": 866910208 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723871614844533, + "loss": 1.3574, + "theoretical_loss": 3.7282359984650757, + "tokens_seen": 866975744 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037237713139418255, + "loss": 1.5464, + "theoretical_loss": 3.7282086777059265, + "tokens_seen": 867041280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723671013039118, + "loss": 1.5424, + "theoretical_loss": 3.728181359589928, + "tokens_seen": 867106816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723570712136409, + "loss": 1.4533, + "theoretical_loss": 3.728154044116624, + "tokens_seen": 867172352 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037234704112337015, + "loss": 1.3955, + "theoretical_loss": 3.72812673128556, + "tokens_seen": 867237888 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037233701103309933, + "loss": 1.4285, + "theoretical_loss": 3.728099421096281, + "tokens_seen": 867303424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723269809428285, + "loss": 1.5222, + "theoretical_loss": 3.72807211354833, + "tokens_seen": 867368960 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723169508525577, + "loss": 1.413, + "theoretical_loss": 3.728044808641254, + "tokens_seen": 867434496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003723069207622869, + "loss": 1.6255, + "theoretical_loss": 3.7280175063745973, + "tokens_seen": 867500032 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037229689067201605, + "loss": 1.4462, + "theoretical_loss": 3.727990206747905, + "tokens_seen": 867565568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003722868605817453, + "loss": 1.6374, + "theoretical_loss": 3.7279629097607234, + "tokens_seen": 867631104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003722768304914744, + "loss": 1.4802, + "theoretical_loss": 3.7279356154125964, + "tokens_seen": 867696640 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037226680040120365, + "loss": 1.4971, + "theoretical_loss": 3.727908323703071, + "tokens_seen": 867762176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003722567703109328, + "loss": 1.5258, + "theoretical_loss": 3.7278810346316913, + "tokens_seen": 867827712 + }, + { + "epoch": 0.26, + "learning_rate": 0.000372246740220662, + "loss": 1.4123, + "theoretical_loss": 3.7278537481980045, + "tokens_seen": 867893248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003722367101303912, + "loss": 1.5987, + "theoretical_loss": 3.7278264644015566, + "tokens_seen": 867958784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003722266800401204, + "loss": 1.5464, + "theoretical_loss": 3.727799183241893, + "tokens_seen": 868024320 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037221664994984956, + "loss": 1.4654, + "theoretical_loss": 3.7277719047185593, + "tokens_seen": 868089856 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037220661985957874, + "loss": 1.4097, + "theoretical_loss": 3.7277446288311022, + "tokens_seen": 868155392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721965897693079, + "loss": 1.539, + "theoretical_loss": 3.7277173555790686, + "tokens_seen": 868220928 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037218655967903716, + "loss": 1.5208, + "theoretical_loss": 3.7276900849620045, + "tokens_seen": 868286464 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 320668, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4758570194244385, + "objective/train/theoretical_loss": 3.727662816979456, + "objective/train/tokens_used": 888812000, + "theoretical_loss": 3.727662816979456, + "tokens_seen": 868352000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721765295887663, + "loss": 1.4864, + "theoretical_loss": 3.727662816979456, + "tokens_seen": 868352000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721664994984955, + "loss": 1.4853, + "theoretical_loss": 3.727635551630971, + "tokens_seen": 868417536 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721564694082247, + "loss": 1.5263, + "theoretical_loss": 3.7276082889160955, + "tokens_seen": 868483072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721464393179539, + "loss": 1.4238, + "theoretical_loss": 3.727581028834376, + "tokens_seen": 868548608 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037213640922768306, + "loss": 1.5387, + "theoretical_loss": 3.7275537713853604, + "tokens_seen": 868614144 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037212637913741224, + "loss": 1.4231, + "theoretical_loss": 3.727526516568595, + "tokens_seen": 868679680 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003721163490471414, + "loss": 1.598, + "theoretical_loss": 3.727499264383628, + "tokens_seen": 868745216 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037210631895687066, + "loss": 1.4031, + "theoretical_loss": 3.7274720148300062, + "tokens_seen": 868810752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003720962888665998, + "loss": 1.449, + "theoretical_loss": 3.727444767907277, + "tokens_seen": 868876288 + }, + { + "epoch": 0.26, + "learning_rate": 0.000372086258776329, + "loss": 1.5045, + "theoretical_loss": 3.7274175236149887, + "tokens_seen": 868941824 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037207622868605815, + "loss": 1.4454, + "theoretical_loss": 3.7273902819526876, + "tokens_seen": 869007360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003720661985957874, + "loss": 1.5092, + "theoretical_loss": 3.7273630429199223, + "tokens_seen": 869072896 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037205616850551656, + "loss": 1.3473, + "theoretical_loss": 3.727335806516241, + "tokens_seen": 869138432 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037204613841524575, + "loss": 1.4592, + "theoretical_loss": 3.727308572741191, + "tokens_seen": 869203968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003720361083249749, + "loss": 1.4992, + "theoretical_loss": 3.7272813415943222, + "tokens_seen": 869269504 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037202607823470416, + "loss": 1.5211, + "theoretical_loss": 3.7272541130751806, + "tokens_seen": 869335040 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003720160481444333, + "loss": 1.4038, + "theoretical_loss": 3.727226887183315, + "tokens_seen": 869400576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003720060180541625, + "loss": 1.5215, + "theoretical_loss": 3.7271996639182756, + "tokens_seen": 869466112 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037199598796389165, + "loss": 1.4776, + "theoretical_loss": 3.727172443279609, + "tokens_seen": 869531648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003719859578736209, + "loss": 1.5777, + "theoretical_loss": 3.727145225266865, + "tokens_seen": 869597184 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037197592778335007, + "loss": 1.5772, + "theoretical_loss": 3.727118009879592, + "tokens_seen": 869662720 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037196589769307925, + "loss": 1.4377, + "theoretical_loss": 3.727090797117339, + "tokens_seen": 869728256 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037195586760280843, + "loss": 1.3492, + "theoretical_loss": 3.7270635869796553, + "tokens_seen": 869793792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003719458375125376, + "loss": 1.4957, + "theoretical_loss": 3.7270363794660892, + "tokens_seen": 869859328 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003719358074222668, + "loss": 1.5244, + "theoretical_loss": 3.727009174576191, + "tokens_seen": 869924864 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371925777331996, + "loss": 1.3437, + "theoretical_loss": 3.7269819723095097, + "tokens_seen": 869990400 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037191574724172515, + "loss": 1.4695, + "theoretical_loss": 3.7269547726655947, + "tokens_seen": 870055936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003719057171514544, + "loss": 1.5598, + "theoretical_loss": 3.7269275756439955, + "tokens_seen": 870121472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003718956870611835, + "loss": 1.5436, + "theoretical_loss": 3.7269003812442616, + "tokens_seen": 870187008 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037188565697091275, + "loss": 1.5149, + "theoretical_loss": 3.7268731894659437, + "tokens_seen": 870252544 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037187562688064193, + "loss": 1.6384, + "theoretical_loss": 3.726846000308591, + "tokens_seen": 870318080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003718655967903711, + "loss": 1.4928, + "theoretical_loss": 3.7268188137717537, + "tokens_seen": 870383616 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003718555667001003, + "loss": 1.4448, + "theoretical_loss": 3.7267916298549815, + "tokens_seen": 870449152 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037184553660982953, + "loss": 1.5533, + "theoretical_loss": 3.7267644485578257, + "tokens_seen": 870514688 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037183550651955866, + "loss": 1.4729, + "theoretical_loss": 3.726737269879836, + "tokens_seen": 870580224 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003718254764292879, + "loss": 1.4846, + "theoretical_loss": 3.7267100938205626, + "tokens_seen": 870645760 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371815446339017, + "loss": 1.6168, + "theoretical_loss": 3.726682920379556, + "tokens_seen": 870711296 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037180541624874625, + "loss": 1.3082, + "theoretical_loss": 3.726655749556368, + "tokens_seen": 870776832 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037179538615847544, + "loss": 1.4215, + "theoretical_loss": 3.726628581350549, + "tokens_seen": 870842368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717853560682046, + "loss": 1.5158, + "theoretical_loss": 3.7266014157616487, + "tokens_seen": 870907904 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717753259779338, + "loss": 1.4931, + "theoretical_loss": 3.72657425278922, + "tokens_seen": 870973440 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371765295887663, + "loss": 1.3603, + "theoretical_loss": 3.7265470924328126, + "tokens_seen": 871038976 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037175526579739216, + "loss": 1.436, + "theoretical_loss": 3.7265199346919786, + "tokens_seen": 871104512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717452357071214, + "loss": 1.5345, + "theoretical_loss": 3.7264927795662692, + "tokens_seen": 871170048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717352056168505, + "loss": 1.5787, + "theoretical_loss": 3.7264656270552354, + "tokens_seen": 871235584 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037172517552657976, + "loss": 1.4392, + "theoretical_loss": 3.72643847715843, + "tokens_seen": 871301120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717151454363089, + "loss": 1.4285, + "theoretical_loss": 3.726411329875403, + "tokens_seen": 871366656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003717051153460381, + "loss": 1.448, + "theoretical_loss": 3.726384185205707, + "tokens_seen": 871432192 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003716950852557673, + "loss": 1.4684, + "theoretical_loss": 3.7263570431488944, + "tokens_seen": 871497728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003716850551654965, + "loss": 1.5829, + "theoretical_loss": 3.7263299037045168, + "tokens_seen": 871563264 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 321822, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6351608037948608, + "objective/train/theoretical_loss": 3.7263027668721262, + "objective/train/tokens_used": 892088800, + "theoretical_loss": 3.7263027668721262, + "tokens_seen": 871628800 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037167502507522566, + "loss": 1.5616, + "theoretical_loss": 3.7263027668721262, + "tokens_seen": 871628800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003716649949849549, + "loss": 1.5311, + "theoretical_loss": 3.7262756326512747, + "tokens_seen": 871694336 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371654964894684, + "loss": 1.3816, + "theoretical_loss": 3.7262485010415154, + "tokens_seen": 871759872 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037164493480441326, + "loss": 1.4126, + "theoretical_loss": 3.7262213720424002, + "tokens_seen": 871825408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003716349047141424, + "loss": 1.4783, + "theoretical_loss": 3.726194245653482, + "tokens_seen": 871890944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003716248746238716, + "loss": 1.5964, + "theoretical_loss": 3.7261671218743135, + "tokens_seen": 871956480 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037161484453360086, + "loss": 1.4146, + "theoretical_loss": 3.7261400007044467, + "tokens_seen": 872022016 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037160481444333, + "loss": 1.4507, + "theoretical_loss": 3.7261128821434353, + "tokens_seen": 872087552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003715947843530592, + "loss": 1.4468, + "theoretical_loss": 3.7260857661908324, + "tokens_seen": 872153088 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037158475426278835, + "loss": 1.5494, + "theoretical_loss": 3.726058652846191, + "tokens_seen": 872218624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003715747241725176, + "loss": 1.4022, + "theoretical_loss": 3.726031542109064, + "tokens_seen": 872284160 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037156469408224676, + "loss": 1.4098, + "theoretical_loss": 3.7260044339790053, + "tokens_seen": 872349696 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037155466399197595, + "loss": 1.5066, + "theoretical_loss": 3.7259773284555675, + "tokens_seen": 872415232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003715446339017051, + "loss": 1.4461, + "theoretical_loss": 3.7259502255383055, + "tokens_seen": 872480768 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037153460381143436, + "loss": 1.4766, + "theoretical_loss": 3.7259231252267715, + "tokens_seen": 872546304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003715245737211635, + "loss": 1.4829, + "theoretical_loss": 3.72589602752052, + "tokens_seen": 872611840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003715145436308927, + "loss": 1.5228, + "theoretical_loss": 3.7258689324191057, + "tokens_seen": 872677376 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037150451354062185, + "loss": 1.496, + "theoretical_loss": 3.725841839922081, + "tokens_seen": 872742912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003714944834503511, + "loss": 1.5126, + "theoretical_loss": 3.7258147500290013, + "tokens_seen": 872808448 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037148445336008027, + "loss": 1.4908, + "theoretical_loss": 3.72578766273942, + "tokens_seen": 872873984 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037147442326980945, + "loss": 1.4501, + "theoretical_loss": 3.725760578052892, + "tokens_seen": 872939520 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037146439317953863, + "loss": 1.4434, + "theoretical_loss": 3.725733495968972, + "tokens_seen": 873005056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003714543630892678, + "loss": 1.5598, + "theoretical_loss": 3.7257064164872142, + "tokens_seen": 873070592 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371444332998997, + "loss": 1.4349, + "theoretical_loss": 3.7256793396071726, + "tokens_seen": 873136128 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037143430290872623, + "loss": 1.4217, + "theoretical_loss": 3.725652265328403, + "tokens_seen": 873201664 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037142427281845535, + "loss": 1.5159, + "theoretical_loss": 3.7256251936504596, + "tokens_seen": 873267200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003714142427281846, + "loss": 1.5829, + "theoretical_loss": 3.725598124572898, + "tokens_seen": 873332736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003714042126379137, + "loss": 1.4465, + "theoretical_loss": 3.725571058095273, + "tokens_seen": 873398272 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037139418254764295, + "loss": 1.4269, + "theoretical_loss": 3.7255439942171398, + "tokens_seen": 873463808 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037138415245737213, + "loss": 1.4394, + "theoretical_loss": 3.7255169329380537, + "tokens_seen": 873529344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003713741223671013, + "loss": 1.6021, + "theoretical_loss": 3.72548987425757, + "tokens_seen": 873594880 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003713640922768305, + "loss": 1.466, + "theoretical_loss": 3.725462818175245, + "tokens_seen": 873660416 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037135406218655973, + "loss": 1.3536, + "theoretical_loss": 3.7254357646906335, + "tokens_seen": 873725952 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037134403209628886, + "loss": 1.4477, + "theoretical_loss": 3.7254087138032914, + "tokens_seen": 873791488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003713340020060181, + "loss": 1.501, + "theoretical_loss": 3.725381665512775, + "tokens_seen": 873857024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003713239719157472, + "loss": 1.4334, + "theoretical_loss": 3.7253546198186402, + "tokens_seen": 873922560 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037131394182547645, + "loss": 1.5202, + "theoretical_loss": 3.7253275767204426, + "tokens_seen": 873988096 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037130391173520564, + "loss": 1.5477, + "theoretical_loss": 3.7253005362177385, + "tokens_seen": 874053632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003712938816449348, + "loss": 1.4864, + "theoretical_loss": 3.7252734983100853, + "tokens_seen": 874119168 + }, + { + "epoch": 0.26, + "learning_rate": 0.000371283851554664, + "loss": 1.477, + "theoretical_loss": 3.7252464629970383, + "tokens_seen": 874184704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003712738214643932, + "loss": 1.3603, + "theoretical_loss": 3.7252194302781536, + "tokens_seen": 874250240 + }, + { + "epoch": 0.26, + "learning_rate": 0.00037126379137412236, + "loss": 1.5193, + "theoretical_loss": 3.7251924001529892, + "tokens_seen": 874315776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003712537612838516, + "loss": 1.5791, + "theoretical_loss": 3.7251653726211016, + "tokens_seen": 874381312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0003712437311935807, + "loss": 1.4999, + "theoretical_loss": 3.7251383476820465, + "tokens_seen": 874446848 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037123370110330996, + "loss": 1.4953, + "theoretical_loss": 3.725111325335382, + "tokens_seen": 874512384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003712236710130391, + "loss": 1.4741, + "theoretical_loss": 3.7250843055806646, + "tokens_seen": 874577920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003712136409227683, + "loss": 1.3727, + "theoretical_loss": 3.7250572884174518, + "tokens_seen": 874643456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003712036108324975, + "loss": 1.5571, + "theoretical_loss": 3.7250302738453005, + "tokens_seen": 874708992 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711935807422267, + "loss": 1.6, + "theoretical_loss": 3.725003261863769, + "tokens_seen": 874774528 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037118355065195586, + "loss": 1.4283, + "theoretical_loss": 3.724976252472414, + "tokens_seen": 874840064 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 323008, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4682408571243286, + "objective/train/theoretical_loss": 3.7249492456707936, + "objective/train/tokens_used": 895365600, + "theoretical_loss": 3.7249492456707936, + "tokens_seen": 874905600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711735205616851, + "loss": 1.4081, + "theoretical_loss": 3.7249492456707936, + "tokens_seen": 874905600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711634904714142, + "loss": 1.6038, + "theoretical_loss": 3.724922241458464, + "tokens_seen": 874971136 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037115346038114346, + "loss": 1.3904, + "theoretical_loss": 3.7248952398349857, + "tokens_seen": 875036672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711434302908726, + "loss": 1.4478, + "theoretical_loss": 3.7248682407999145, + "tokens_seen": 875102208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711334002006018, + "loss": 1.4732, + "theoretical_loss": 3.7248412443528096, + "tokens_seen": 875167744 + }, + { + "epoch": 0.27, + "learning_rate": 0.000371123370110331, + "loss": 1.4165, + "theoretical_loss": 3.7248142504932282, + "tokens_seen": 875233280 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003711133400200602, + "loss": 1.5008, + "theoretical_loss": 3.7247872592207303, + "tokens_seen": 875298816 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037110330992978937, + "loss": 1.4963, + "theoretical_loss": 3.7247602705348717, + "tokens_seen": 875364352 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037109327983951855, + "loss": 1.4447, + "theoretical_loss": 3.7247332844352132, + "tokens_seen": 875429888 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037108324974924773, + "loss": 1.4293, + "theoretical_loss": 3.724706300921312, + "tokens_seen": 875495424 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037107321965897696, + "loss": 1.4814, + "theoretical_loss": 3.7246793199927275, + "tokens_seen": 875560960 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003710631895687061, + "loss": 1.6263, + "theoretical_loss": 3.7246523416490187, + "tokens_seen": 875626496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003710531594784353, + "loss": 1.467, + "theoretical_loss": 3.724625365889744, + "tokens_seen": 875692032 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037104312938816445, + "loss": 1.3809, + "theoretical_loss": 3.724598392714462, + "tokens_seen": 875757568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003710330992978937, + "loss": 1.4609, + "theoretical_loss": 3.724571422122733, + "tokens_seen": 875823104 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037102306920762287, + "loss": 1.4392, + "theoretical_loss": 3.724544454114115, + "tokens_seen": 875888640 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037101303911735205, + "loss": 1.387, + "theoretical_loss": 3.7245174886881687, + "tokens_seen": 875954176 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037100300902708123, + "loss": 1.5245, + "theoretical_loss": 3.7244905258444527, + "tokens_seen": 876019712 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037099297893681047, + "loss": 1.5356, + "theoretical_loss": 3.7244635655825267, + "tokens_seen": 876085248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003709829488465396, + "loss": 1.5649, + "theoretical_loss": 3.7244366079019504, + "tokens_seen": 876150784 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037097291875626883, + "loss": 1.467, + "theoretical_loss": 3.724409652802284, + "tokens_seen": 876216320 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037096288866599796, + "loss": 1.494, + "theoretical_loss": 3.7243827002830865, + "tokens_seen": 876281856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003709528585757272, + "loss": 1.5046, + "theoretical_loss": 3.7243557503439186, + "tokens_seen": 876347392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003709428284854564, + "loss": 1.445, + "theoretical_loss": 3.72432880298434, + "tokens_seen": 876412928 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037093279839518555, + "loss": 1.5712, + "theoretical_loss": 3.7243018582039116, + "tokens_seen": 876478464 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037092276830491474, + "loss": 1.4465, + "theoretical_loss": 3.724274916002193, + "tokens_seen": 876544000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003709127382146439, + "loss": 1.5129, + "theoretical_loss": 3.724247976378745, + "tokens_seen": 876609536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003709027081243731, + "loss": 1.6064, + "theoretical_loss": 3.724221039333128, + "tokens_seen": 876675072 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037089267803410233, + "loss": 1.4498, + "theoretical_loss": 3.724194104864903, + "tokens_seen": 876740608 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037088264794383146, + "loss": 1.5283, + "theoretical_loss": 3.72416717297363, + "tokens_seen": 876806144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003708726178535607, + "loss": 1.4309, + "theoretical_loss": 3.7241402436588706, + "tokens_seen": 876871680 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037086258776328993, + "loss": 1.4035, + "theoretical_loss": 3.724113316920185, + "tokens_seen": 876937216 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037085255767301906, + "loss": 1.5751, + "theoretical_loss": 3.724086392757136, + "tokens_seen": 877002752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003708425275827483, + "loss": 1.3914, + "theoretical_loss": 3.7240594711692827, + "tokens_seen": 877068288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003708324974924774, + "loss": 1.4748, + "theoretical_loss": 3.7240325521561872, + "tokens_seen": 877133824 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037082246740220665, + "loss": 1.4439, + "theoretical_loss": 3.724005635717411, + "tokens_seen": 877199360 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037081243731193584, + "loss": 1.4711, + "theoretical_loss": 3.7239787218525158, + "tokens_seen": 877264896 + }, + { + "epoch": 0.27, + "learning_rate": 0.000370802407221665, + "loss": 1.4431, + "theoretical_loss": 3.723951810561063, + "tokens_seen": 877330432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707923771313942, + "loss": 1.4607, + "theoretical_loss": 3.723924901842614, + "tokens_seen": 877395968 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707823470411234, + "loss": 1.4904, + "theoretical_loss": 3.7238979956967317, + "tokens_seen": 877461504 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037077231695085256, + "loss": 1.4818, + "theoretical_loss": 3.7238710921229767, + "tokens_seen": 877527040 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707622868605818, + "loss": 1.3345, + "theoretical_loss": 3.7238441911209117, + "tokens_seen": 877592576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707522567703109, + "loss": 1.4508, + "theoretical_loss": 3.723817292690099, + "tokens_seen": 877658112 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037074222668004016, + "loss": 1.5356, + "theoretical_loss": 3.723790396830101, + "tokens_seen": 877723648 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707321965897693, + "loss": 1.503, + "theoretical_loss": 3.7237635035404795, + "tokens_seen": 877789184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707221664994985, + "loss": 1.5166, + "theoretical_loss": 3.723736612820797, + "tokens_seen": 877854720 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707121364092277, + "loss": 1.5051, + "theoretical_loss": 3.7237097246706163, + "tokens_seen": 877920256 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003707021063189569, + "loss": 1.5142, + "theoretical_loss": 3.7236828390895003, + "tokens_seen": 877985792 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037069207622868606, + "loss": 1.5419, + "theoretical_loss": 3.7236559560770113, + "tokens_seen": 878051328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003706820461384153, + "loss": 1.4994, + "theoretical_loss": 3.7236290756327133, + "tokens_seen": 878116864 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 324331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3624552488327026, + "objective/train/theoretical_loss": 3.7236021977561675, + "objective/train/tokens_used": 898642400, + "theoretical_loss": 3.7236021977561675, + "tokens_seen": 878182400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003706720160481444, + "loss": 1.4803, + "theoretical_loss": 3.7236021977561675, + "tokens_seen": 878182400 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037066198595787366, + "loss": 1.4102, + "theoretical_loss": 3.723575322446939, + "tokens_seen": 878247936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003706519558676028, + "loss": 1.5104, + "theoretical_loss": 3.723548449704589, + "tokens_seen": 878313472 + }, + { + "epoch": 0.27, + "learning_rate": 0.000370641925777332, + "loss": 1.3667, + "theoretical_loss": 3.7235215795286827, + "tokens_seen": 878379008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003706318956870612, + "loss": 1.547, + "theoretical_loss": 3.7234947119187822, + "tokens_seen": 878444544 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003706218655967904, + "loss": 1.369, + "theoretical_loss": 3.723467846874452, + "tokens_seen": 878510080 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037061183550651957, + "loss": 1.6232, + "theoretical_loss": 3.723440984395255, + "tokens_seen": 878575616 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037060180541624875, + "loss": 1.4603, + "theoretical_loss": 3.7234141244807555, + "tokens_seen": 878641152 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037059177532597793, + "loss": 1.421, + "theoretical_loss": 3.7233872671305175, + "tokens_seen": 878706688 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037058174523570716, + "loss": 1.4063, + "theoretical_loss": 3.723360412344104, + "tokens_seen": 878772224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003705717151454363, + "loss": 1.432, + "theoretical_loss": 3.7233335601210804, + "tokens_seen": 878837760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003705616850551655, + "loss": 1.4999, + "theoretical_loss": 3.7233067104610096, + "tokens_seen": 878903296 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037055165496489465, + "loss": 1.5248, + "theoretical_loss": 3.723279863363457, + "tokens_seen": 878968832 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003705416248746239, + "loss": 1.5467, + "theoretical_loss": 3.723253018827986, + "tokens_seen": 879034368 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037053159478435307, + "loss": 1.3519, + "theoretical_loss": 3.723226176854162, + "tokens_seen": 879099904 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037052156469408225, + "loss": 1.4633, + "theoretical_loss": 3.723199337441549, + "tokens_seen": 879165440 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037051153460381143, + "loss": 1.4204, + "theoretical_loss": 3.723172500589712, + "tokens_seen": 879230976 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037050150451354067, + "loss": 1.5216, + "theoretical_loss": 3.7231456662982154, + "tokens_seen": 879296512 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003704914744232698, + "loss": 1.4172, + "theoretical_loss": 3.723118834566625, + "tokens_seen": 879362048 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037048144433299903, + "loss": 1.549, + "theoretical_loss": 3.7230920053945056, + "tokens_seen": 879427584 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037047141424272816, + "loss": 1.5158, + "theoretical_loss": 3.7230651787814217, + "tokens_seen": 879493120 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003704613841524574, + "loss": 1.3809, + "theoretical_loss": 3.7230383547269392, + "tokens_seen": 879558656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003704513540621866, + "loss": 1.5165, + "theoretical_loss": 3.723011533230623, + "tokens_seen": 879624192 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037044132397191575, + "loss": 1.5422, + "theoretical_loss": 3.7229847142920383, + "tokens_seen": 879689728 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037043129388164494, + "loss": 1.5332, + "theoretical_loss": 3.7229578979107516, + "tokens_seen": 879755264 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003704212637913741, + "loss": 1.4787, + "theoretical_loss": 3.722931084086328, + "tokens_seen": 879820800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003704112337011033, + "loss": 1.5755, + "theoretical_loss": 3.722904272818334, + "tokens_seen": 879886336 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037040120361083253, + "loss": 1.528, + "theoretical_loss": 3.722877464106334, + "tokens_seen": 879951872 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037039117352056166, + "loss": 1.5236, + "theoretical_loss": 3.722850657949895, + "tokens_seen": 880017408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003703811434302909, + "loss": 1.5349, + "theoretical_loss": 3.722823854348583, + "tokens_seen": 880082944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003703711133400201, + "loss": 1.5697, + "theoretical_loss": 3.7227970533019645, + "tokens_seen": 880148480 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037036108324974926, + "loss": 1.4303, + "theoretical_loss": 3.7227702548096056, + "tokens_seen": 880214016 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037035105315947844, + "loss": 1.4456, + "theoretical_loss": 3.722743458871072, + "tokens_seen": 880279552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003703410230692076, + "loss": 1.4606, + "theoretical_loss": 3.7227166654859314, + "tokens_seen": 880345088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003703309929789368, + "loss": 1.4858, + "theoretical_loss": 3.7226898746537493, + "tokens_seen": 880410624 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037032096288866604, + "loss": 1.4029, + "theoretical_loss": 3.7226630863740935, + "tokens_seen": 880476160 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037031093279839516, + "loss": 1.499, + "theoretical_loss": 3.7226363006465304, + "tokens_seen": 880541696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003703009027081244, + "loss": 1.5119, + "theoretical_loss": 3.7226095174706266, + "tokens_seen": 880607232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003702908726178535, + "loss": 1.3869, + "theoretical_loss": 3.7225827368459496, + "tokens_seen": 880672768 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037028084252758276, + "loss": 1.3366, + "theoretical_loss": 3.722555958772066, + "tokens_seen": 880738304 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037027081243731194, + "loss": 1.4051, + "theoretical_loss": 3.722529183248544, + "tokens_seen": 880803840 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003702607823470411, + "loss": 1.3859, + "theoretical_loss": 3.7225024102749504, + "tokens_seen": 880869376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003702507522567703, + "loss": 1.4291, + "theoretical_loss": 3.7224756398508525, + "tokens_seen": 880934912 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003702407221664995, + "loss": 1.6119, + "theoretical_loss": 3.7224488719758178, + "tokens_seen": 881000448 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037023069207622867, + "loss": 1.6267, + "theoretical_loss": 3.7224221066494145, + "tokens_seen": 881065984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003702206619859579, + "loss": 1.5463, + "theoretical_loss": 3.7223953438712103, + "tokens_seen": 881131520 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037021063189568703, + "loss": 1.5037, + "theoretical_loss": 3.7223685836407734, + "tokens_seen": 881197056 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037020060180541626, + "loss": 1.4978, + "theoretical_loss": 3.7223418259576704, + "tokens_seen": 881262592 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037019057171514544, + "loss": 1.4578, + "theoretical_loss": 3.7223150708214714, + "tokens_seen": 881328128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003701805416248746, + "loss": 1.4412, + "theoretical_loss": 3.722288318231743, + "tokens_seen": 881393664 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 325483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4086384773254395, + "objective/train/theoretical_loss": 3.722261568188054, + "objective/train/tokens_used": 901919200, + "theoretical_loss": 3.722261568188054, + "tokens_seen": 881459200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003701705115346038, + "loss": 1.4059, + "theoretical_loss": 3.722261568188054, + "tokens_seen": 881459200 + }, + { + "epoch": 0.27, + "learning_rate": 0.000370160481444333, + "loss": 1.5476, + "theoretical_loss": 3.722234820689973, + "tokens_seen": 881524736 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037015045135406217, + "loss": 1.4161, + "theoretical_loss": 3.722208075737069, + "tokens_seen": 881590272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003701404212637914, + "loss": 1.5133, + "theoretical_loss": 3.7221813333289093, + "tokens_seen": 881655808 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037013039117352053, + "loss": 1.4902, + "theoretical_loss": 3.722154593465064, + "tokens_seen": 881721344 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037012036108324977, + "loss": 1.5466, + "theoretical_loss": 3.722127856145101, + "tokens_seen": 881786880 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037011033099297895, + "loss": 1.4579, + "theoretical_loss": 3.72210112136859, + "tokens_seen": 881852416 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037010030090270813, + "loss": 1.4904, + "theoretical_loss": 3.7220743891350994, + "tokens_seen": 881917952 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037009027081243736, + "loss": 1.4528, + "theoretical_loss": 3.7220476594441987, + "tokens_seen": 881983488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003700802407221665, + "loss": 1.5158, + "theoretical_loss": 3.7220209322954574, + "tokens_seen": 882049024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003700702106318957, + "loss": 1.4736, + "theoretical_loss": 3.7219942076884447, + "tokens_seen": 882114560 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037006018054162485, + "loss": 1.4852, + "theoretical_loss": 3.72196748562273, + "tokens_seen": 882180096 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003700501504513541, + "loss": 1.5083, + "theoretical_loss": 3.721940766097882, + "tokens_seen": 882245632 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037004012036108327, + "loss": 1.5224, + "theoretical_loss": 3.721914049113472, + "tokens_seen": 882311168 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037003009027081245, + "loss": 1.4139, + "theoretical_loss": 3.721887334669069, + "tokens_seen": 882376704 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037002006018054163, + "loss": 1.5826, + "theoretical_loss": 3.7218606227642423, + "tokens_seen": 882442240 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037001003009027087, + "loss": 1.4093, + "theoretical_loss": 3.721833913398563, + "tokens_seen": 882507776 + }, + { + "epoch": 0.27, + "learning_rate": 0.00037, + "loss": 1.4748, + "theoretical_loss": 3.7218072065716004, + "tokens_seen": 882573312 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036998996990972923, + "loss": 1.559, + "theoretical_loss": 3.7217805022829253, + "tokens_seen": 882638848 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036997993981945836, + "loss": 1.5706, + "theoretical_loss": 3.721753800532108, + "tokens_seen": 882704384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003699699097291876, + "loss": 1.501, + "theoretical_loss": 3.7217271013187183, + "tokens_seen": 882769920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003699598796389168, + "loss": 1.4322, + "theoretical_loss": 3.721700404642327, + "tokens_seen": 882835456 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036994984954864595, + "loss": 1.4362, + "theoretical_loss": 3.721673710502505, + "tokens_seen": 882900992 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036993981945837514, + "loss": 1.422, + "theoretical_loss": 3.7216470188988224, + "tokens_seen": 882966528 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003699297893681043, + "loss": 1.5184, + "theoretical_loss": 3.721620329830851, + "tokens_seen": 883032064 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003699197592778335, + "loss": 1.5265, + "theoretical_loss": 3.721593643298161, + "tokens_seen": 883097600 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036990972918756273, + "loss": 1.4225, + "theoretical_loss": 3.7215669593003238, + "tokens_seen": 883163136 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036989969909729186, + "loss": 1.5854, + "theoretical_loss": 3.72154027783691, + "tokens_seen": 883228672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003698896690070211, + "loss": 1.427, + "theoretical_loss": 3.7215135989074914, + "tokens_seen": 883294208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003698796389167503, + "loss": 1.4318, + "theoretical_loss": 3.7214869225116396, + "tokens_seen": 883359744 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036986960882647946, + "loss": 1.412, + "theoretical_loss": 3.7214602486489254, + "tokens_seen": 883425280 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036985957873620864, + "loss": 1.4629, + "theoretical_loss": 3.721433577318921, + "tokens_seen": 883490816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003698495486459378, + "loss": 1.3443, + "theoretical_loss": 3.721406908521197, + "tokens_seen": 883556352 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369839518555667, + "loss": 1.4279, + "theoretical_loss": 3.7213802422553264, + "tokens_seen": 883621888 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036982948846539624, + "loss": 1.4237, + "theoretical_loss": 3.7213535785208807, + "tokens_seen": 883687424 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036981945837512536, + "loss": 1.5412, + "theoretical_loss": 3.7213269173174313, + "tokens_seen": 883752960 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003698094282848546, + "loss": 1.4375, + "theoretical_loss": 3.7213002586445514, + "tokens_seen": 883818496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003697993981945837, + "loss": 1.5033, + "theoretical_loss": 3.721273602501812, + "tokens_seen": 883884032 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036978936810431296, + "loss": 1.4332, + "theoretical_loss": 3.721246948888786, + "tokens_seen": 883949568 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036977933801404214, + "loss": 1.5919, + "theoretical_loss": 3.7212202978050457, + "tokens_seen": 884015104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003697693079237713, + "loss": 1.4492, + "theoretical_loss": 3.721193649250164, + "tokens_seen": 884080640 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003697592778335005, + "loss": 1.4992, + "theoretical_loss": 3.7211670032237123, + "tokens_seen": 884146176 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003697492477432297, + "loss": 1.4206, + "theoretical_loss": 3.721140359725265, + "tokens_seen": 884211712 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036973921765295887, + "loss": 1.4775, + "theoretical_loss": 3.7211137187543937, + "tokens_seen": 884277248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003697291875626881, + "loss": 1.5248, + "theoretical_loss": 3.7210870803106717, + "tokens_seen": 884342784 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036971915747241723, + "loss": 1.3022, + "theoretical_loss": 3.721060444393672, + "tokens_seen": 884408320 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036970912738214646, + "loss": 1.5189, + "theoretical_loss": 3.721033811002968, + "tokens_seen": 884473856 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036969909729187565, + "loss": 1.4622, + "theoretical_loss": 3.721007180138132, + "tokens_seen": 884539392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003696890672016048, + "loss": 1.3366, + "theoretical_loss": 3.720980551798739, + "tokens_seen": 884604928 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369679037111334, + "loss": 1.5363, + "theoretical_loss": 3.720953925984361, + "tokens_seen": 884670464 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 326745, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6544883251190186, + "objective/train/theoretical_loss": 3.720927302694572, + "objective/train/tokens_used": 905196000, + "theoretical_loss": 3.720927302694572, + "tokens_seen": 884736000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003696690070210632, + "loss": 1.5285, + "theoretical_loss": 3.720927302694572, + "tokens_seen": 884736000 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036965897693079237, + "loss": 1.4837, + "theoretical_loss": 3.7209006819289456, + "tokens_seen": 884801536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003696489468405216, + "loss": 1.486, + "theoretical_loss": 3.7208740636870554, + "tokens_seen": 884867072 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036963891675025073, + "loss": 1.4996, + "theoretical_loss": 3.7208474479684757, + "tokens_seen": 884932608 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036962888665997997, + "loss": 1.5466, + "theoretical_loss": 3.72082083477278, + "tokens_seen": 884998144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003696188565697091, + "loss": 1.487, + "theoretical_loss": 3.7207942240995426, + "tokens_seen": 885063680 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036960882647943833, + "loss": 1.5067, + "theoretical_loss": 3.7207676159483376, + "tokens_seen": 885129216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003695987963891675, + "loss": 1.3325, + "theoretical_loss": 3.720741010318739, + "tokens_seen": 885194752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003695887662988967, + "loss": 1.4185, + "theoretical_loss": 3.720714407210322, + "tokens_seen": 885260288 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036957873620862587, + "loss": 1.3753, + "theoretical_loss": 3.72068780662266, + "tokens_seen": 885325824 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036956870611835505, + "loss": 1.4417, + "theoretical_loss": 3.7206612085553283, + "tokens_seen": 885391360 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036955867602808424, + "loss": 1.4784, + "theoretical_loss": 3.7206346130079018, + "tokens_seen": 885456896 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036954864593781347, + "loss": 1.502, + "theoretical_loss": 3.720608019979954, + "tokens_seen": 885522432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003695386158475426, + "loss": 1.5251, + "theoretical_loss": 3.7205814294710615, + "tokens_seen": 885587968 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036952858575727183, + "loss": 1.4535, + "theoretical_loss": 3.720554841480798, + "tokens_seen": 885653504 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369518555667001, + "loss": 1.4767, + "theoretical_loss": 3.720528256008739, + "tokens_seen": 885719040 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003695085255767302, + "loss": 1.5493, + "theoretical_loss": 3.72050167305446, + "tokens_seen": 885784576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003694984954864594, + "loss": 1.426, + "theoretical_loss": 3.7204750926175354, + "tokens_seen": 885850112 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036948846539618856, + "loss": 1.467, + "theoretical_loss": 3.7204485146975417, + "tokens_seen": 885915648 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036947843530591774, + "loss": 1.4913, + "theoretical_loss": 3.720421939294054, + "tokens_seen": 885981184 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369468405215647, + "loss": 1.4386, + "theoretical_loss": 3.7203953664066476, + "tokens_seen": 886046720 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003694583751253761, + "loss": 1.4995, + "theoretical_loss": 3.7203687960348986, + "tokens_seen": 886112256 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036944834503510534, + "loss": 1.3947, + "theoretical_loss": 3.7203422281783824, + "tokens_seen": 886177792 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036943831494483446, + "loss": 1.5381, + "theoretical_loss": 3.7203156628366756, + "tokens_seen": 886243328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003694282848545637, + "loss": 1.5602, + "theoretical_loss": 3.720289100009353, + "tokens_seen": 886308864 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003694182547642929, + "loss": 1.3445, + "theoretical_loss": 3.7202625396959923, + "tokens_seen": 886374400 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036940822467402206, + "loss": 1.4637, + "theoretical_loss": 3.7202359818961686, + "tokens_seen": 886439936 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036939819458375124, + "loss": 1.5082, + "theoretical_loss": 3.7202094266094585, + "tokens_seen": 886505472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003693881644934805, + "loss": 1.474, + "theoretical_loss": 3.720182873835439, + "tokens_seen": 886571008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003693781344032096, + "loss": 1.3668, + "theoretical_loss": 3.7201563235736854, + "tokens_seen": 886636544 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036936810431293884, + "loss": 1.4261, + "theoretical_loss": 3.7201297758237755, + "tokens_seen": 886702080 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369358074222668, + "loss": 1.4801, + "theoretical_loss": 3.720103230585286, + "tokens_seen": 886767616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003693480441323972, + "loss": 1.4561, + "theoretical_loss": 3.7200766878577927, + "tokens_seen": 886833152 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036933801404212644, + "loss": 1.4744, + "theoretical_loss": 3.720050147640873, + "tokens_seen": 886898688 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036932798395185556, + "loss": 1.4882, + "theoretical_loss": 3.7200236099341044, + "tokens_seen": 886964224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003693179538615848, + "loss": 1.4742, + "theoretical_loss": 3.719997074737064, + "tokens_seen": 887029760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003693079237713139, + "loss": 1.5661, + "theoretical_loss": 3.7199705420493285, + "tokens_seen": 887095296 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036929789368104316, + "loss": 1.4976, + "theoretical_loss": 3.719944011870476, + "tokens_seen": 887160832 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036928786359077234, + "loss": 1.4573, + "theoretical_loss": 3.7199174842000833, + "tokens_seen": 887226368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003692778335005015, + "loss": 1.5111, + "theoretical_loss": 3.719890959037728, + "tokens_seen": 887291904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003692678034102307, + "loss": 1.4621, + "theoretical_loss": 3.719864436382988, + "tokens_seen": 887357440 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003692577733199599, + "loss": 1.4983, + "theoretical_loss": 3.7198379162354414, + "tokens_seen": 887422976 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036924774322968907, + "loss": 1.4527, + "theoretical_loss": 3.7198113985946657, + "tokens_seen": 887488512 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003692377131394183, + "loss": 1.5124, + "theoretical_loss": 3.719784883460238, + "tokens_seen": 887554048 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036922768304914743, + "loss": 1.3931, + "theoretical_loss": 3.719758370831738, + "tokens_seen": 887619584 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036921765295887666, + "loss": 1.4154, + "theoretical_loss": 3.719731860708743, + "tokens_seen": 887685120 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036920762286860585, + "loss": 1.4117, + "theoretical_loss": 3.7197053530908315, + "tokens_seen": 887750656 + }, + { + "epoch": 0.27, + "learning_rate": 0.000369197592778335, + "loss": 1.5024, + "theoretical_loss": 3.7196788479775815, + "tokens_seen": 887816192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003691875626880642, + "loss": 1.3843, + "theoretical_loss": 3.7196523453685715, + "tokens_seen": 887881728 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003691775325977934, + "loss": 1.5168, + "theoretical_loss": 3.719625845263381, + "tokens_seen": 887947264 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 327973, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.337606430053711, + "objective/train/theoretical_loss": 3.719599347661587, + "objective/train/tokens_used": 908472800, + "theoretical_loss": 3.719599347661587, + "tokens_seen": 888012800 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036916750250752257, + "loss": 1.4046, + "theoretical_loss": 3.719599347661587, + "tokens_seen": 888012800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003691574724172518, + "loss": 1.5362, + "theoretical_loss": 3.71957285256277, + "tokens_seen": 888078336 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036914744232698093, + "loss": 1.4457, + "theoretical_loss": 3.7195463599665075, + "tokens_seen": 888143872 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036913741223671017, + "loss": 1.4812, + "theoretical_loss": 3.7195198698723795, + "tokens_seen": 888209408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003691273821464393, + "loss": 1.438, + "theoretical_loss": 3.719493382279965, + "tokens_seen": 888274944 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036911735205616853, + "loss": 1.4377, + "theoretical_loss": 3.719466897188843, + "tokens_seen": 888340480 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003691073219658977, + "loss": 1.4934, + "theoretical_loss": 3.7194404145985924, + "tokens_seen": 888406016 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690972918756269, + "loss": 1.378, + "theoretical_loss": 3.719413934508793, + "tokens_seen": 888471552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690872617853561, + "loss": 1.4867, + "theoretical_loss": 3.719387456919024, + "tokens_seen": 888537088 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036907723169508525, + "loss": 1.4747, + "theoretical_loss": 3.7193609818288653, + "tokens_seen": 888602624 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036906720160481444, + "loss": 1.3755, + "theoretical_loss": 3.7193345092378967, + "tokens_seen": 888668160 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036905717151454367, + "loss": 1.4386, + "theoretical_loss": 3.719308039145698, + "tokens_seen": 888733696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690471414242728, + "loss": 1.4923, + "theoretical_loss": 3.7192815715518486, + "tokens_seen": 888799232 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036903711133400203, + "loss": 1.4473, + "theoretical_loss": 3.719255106455929, + "tokens_seen": 888864768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690270812437312, + "loss": 1.5958, + "theoretical_loss": 3.7192286438575186, + "tokens_seen": 888930304 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690170511534604, + "loss": 1.4351, + "theoretical_loss": 3.7192021837561984, + "tokens_seen": 888995840 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003690070210631896, + "loss": 1.4551, + "theoretical_loss": 3.7191757261515486, + "tokens_seen": 889061376 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036899699097291876, + "loss": 1.4446, + "theoretical_loss": 3.7191492710431495, + "tokens_seen": 889126912 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036898696088264794, + "loss": 1.4005, + "theoretical_loss": 3.7191228184305816, + "tokens_seen": 889192448 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003689769307923772, + "loss": 1.5022, + "theoretical_loss": 3.719096368313425, + "tokens_seen": 889257984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003689669007021063, + "loss": 1.5097, + "theoretical_loss": 3.719069920691261, + "tokens_seen": 889323520 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036895687061183554, + "loss": 1.5164, + "theoretical_loss": 3.7190434755636708, + "tokens_seen": 889389056 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036894684052156466, + "loss": 1.471, + "theoretical_loss": 3.7190170329302346, + "tokens_seen": 889454592 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003689368104312939, + "loss": 1.5371, + "theoretical_loss": 3.718990592790534, + "tokens_seen": 889520128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003689267803410231, + "loss": 1.4878, + "theoretical_loss": 3.718964155144149, + "tokens_seen": 889585664 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036891675025075226, + "loss": 1.3687, + "theoretical_loss": 3.718937719990662, + "tokens_seen": 889651200 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036890672016048144, + "loss": 1.4737, + "theoretical_loss": 3.7189112873296537, + "tokens_seen": 889716736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688966900702107, + "loss": 1.4689, + "theoretical_loss": 3.718884857160706, + "tokens_seen": 889782272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688866599799398, + "loss": 1.5338, + "theoretical_loss": 3.7188584294834004, + "tokens_seen": 889847808 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036887662988966904, + "loss": 1.485, + "theoretical_loss": 3.718832004297318, + "tokens_seen": 889913344 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036886659979939817, + "loss": 1.509, + "theoretical_loss": 3.718805581602041, + "tokens_seen": 889978880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688565697091274, + "loss": 1.5186, + "theoretical_loss": 3.7187791613971504, + "tokens_seen": 890044416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688465396188566, + "loss": 1.4878, + "theoretical_loss": 3.7187527436822294, + "tokens_seen": 890109952 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036883650952858576, + "loss": 1.4817, + "theoretical_loss": 3.7187263284568592, + "tokens_seen": 890175488 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036882647943831494, + "loss": 1.4224, + "theoretical_loss": 3.7186999157206224, + "tokens_seen": 890241024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688164493480441, + "loss": 1.3721, + "theoretical_loss": 3.7186735054731006, + "tokens_seen": 890306560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003688064192577733, + "loss": 1.4175, + "theoretical_loss": 3.718647097713877, + "tokens_seen": 890372096 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036879638916750254, + "loss": 1.4518, + "theoretical_loss": 3.718620692442533, + "tokens_seen": 890437632 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036878635907723167, + "loss": 1.4579, + "theoretical_loss": 3.718594289658652, + "tokens_seen": 890503168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003687763289869609, + "loss": 1.4694, + "theoretical_loss": 3.7185678893618164, + "tokens_seen": 890568704 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036876629889669003, + "loss": 1.4854, + "theoretical_loss": 3.7185414915516084, + "tokens_seen": 890634240 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036875626880641927, + "loss": 1.4586, + "theoretical_loss": 3.7185150962276117, + "tokens_seen": 890699776 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036874623871614845, + "loss": 1.4764, + "theoretical_loss": 3.7184887033894087, + "tokens_seen": 890765312 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036873620862587763, + "loss": 1.3815, + "theoretical_loss": 3.718462313036582, + "tokens_seen": 890830848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003687261785356068, + "loss": 1.4464, + "theoretical_loss": 3.7184359251687162, + "tokens_seen": 890896384 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036871614844533605, + "loss": 1.4654, + "theoretical_loss": 3.718409539785393, + "tokens_seen": 890961920 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036870611835506517, + "loss": 1.5017, + "theoretical_loss": 3.7183831568861967, + "tokens_seen": 891027456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686960882647944, + "loss": 1.428, + "theoretical_loss": 3.718356776470711, + "tokens_seen": 891092992 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036868605817452353, + "loss": 1.5863, + "theoretical_loss": 3.7183303985385177, + "tokens_seen": 891158528 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036867602808425277, + "loss": 1.5903, + "theoretical_loss": 3.7183040230892024, + "tokens_seen": 891224064 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 329220, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3951911926269531, + "objective/train/theoretical_loss": 3.7182776501223476, + "objective/train/tokens_used": 911749600, + "theoretical_loss": 3.7182776501223476, + "tokens_seen": 891289600 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036866599799398195, + "loss": 1.4264, + "theoretical_loss": 3.7182776501223476, + "tokens_seen": 891289600 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036865596790371113, + "loss": 1.3742, + "theoretical_loss": 3.7182512796375375, + "tokens_seen": 891355136 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686459378134403, + "loss": 1.4511, + "theoretical_loss": 3.7182249116343566, + "tokens_seen": 891420672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686359077231695, + "loss": 1.5499, + "theoretical_loss": 3.7181985461123883, + "tokens_seen": 891486208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686258776328987, + "loss": 1.4955, + "theoretical_loss": 3.718172183071217, + "tokens_seen": 891551744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686158475426279, + "loss": 1.546, + "theoretical_loss": 3.7181458225104267, + "tokens_seen": 891617280 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003686058174523571, + "loss": 1.3481, + "theoretical_loss": 3.718119464429602, + "tokens_seen": 891682816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003685957873620863, + "loss": 1.5093, + "theoretical_loss": 3.718093108828327, + "tokens_seen": 891748352 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036858575727181545, + "loss": 1.4382, + "theoretical_loss": 3.718066755706187, + "tokens_seen": 891813888 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036857572718154464, + "loss": 1.4242, + "theoretical_loss": 3.718040405062766, + "tokens_seen": 891879424 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036856569709127387, + "loss": 1.5006, + "theoretical_loss": 3.7180140568976485, + "tokens_seen": 891944960 + }, + { + "epoch": 0.27, + "learning_rate": 0.000368555667001003, + "loss": 1.4927, + "theoretical_loss": 3.7179877112104194, + "tokens_seen": 892010496 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036854563691073223, + "loss": 1.419, + "theoretical_loss": 3.7179613680006645, + "tokens_seen": 892076032 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003685356068204614, + "loss": 1.4962, + "theoretical_loss": 3.7179350272679685, + "tokens_seen": 892141568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003685255767301906, + "loss": 1.5432, + "theoretical_loss": 3.7179086890119164, + "tokens_seen": 892207104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003685155466399198, + "loss": 1.4787, + "theoretical_loss": 3.717882353232093, + "tokens_seen": 892272640 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036850551654964896, + "loss": 1.5264, + "theoretical_loss": 3.7178560199280843, + "tokens_seen": 892338176 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036849548645937814, + "loss": 1.4693, + "theoretical_loss": 3.7178296890994753, + "tokens_seen": 892403712 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003684854563691074, + "loss": 1.5663, + "theoretical_loss": 3.717803360745852, + "tokens_seen": 892469248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003684754262788365, + "loss": 1.5432, + "theoretical_loss": 3.717777034866799, + "tokens_seen": 892534784 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036846539618856574, + "loss": 1.4876, + "theoretical_loss": 3.7177507114619037, + "tokens_seen": 892600320 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036845536609829486, + "loss": 1.586, + "theoretical_loss": 3.7177243905307504, + "tokens_seen": 892665856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003684453360080241, + "loss": 1.4639, + "theoretical_loss": 3.7176980720729267, + "tokens_seen": 892731392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003684353059177533, + "loss": 1.5998, + "theoretical_loss": 3.717671756088017, + "tokens_seen": 892796928 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036842527582748246, + "loss": 1.425, + "theoretical_loss": 3.717645442575608, + "tokens_seen": 892862464 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036841524573721164, + "loss": 1.513, + "theoretical_loss": 3.717619131535286, + "tokens_seen": 892928000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003684052156469409, + "loss": 1.5782, + "theoretical_loss": 3.7175928229666377, + "tokens_seen": 892993536 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036839518555667, + "loss": 1.5393, + "theoretical_loss": 3.717566516869249, + "tokens_seen": 893059072 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036838515546639924, + "loss": 1.3139, + "theoretical_loss": 3.7175402132427067, + "tokens_seen": 893124608 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036837512537612837, + "loss": 1.5089, + "theoretical_loss": 3.7175139120865976, + "tokens_seen": 893190144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003683650952858576, + "loss": 1.415, + "theoretical_loss": 3.7174876134005075, + "tokens_seen": 893255680 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003683550651955868, + "loss": 1.3127, + "theoretical_loss": 3.717461317184025, + "tokens_seen": 893321216 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036834503510531596, + "loss": 1.486, + "theoretical_loss": 3.7174350234367353, + "tokens_seen": 893386752 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036833500501504514, + "loss": 1.4697, + "theoretical_loss": 3.7174087321582263, + "tokens_seen": 893452288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003683249749247743, + "loss": 1.451, + "theoretical_loss": 3.717382443348085, + "tokens_seen": 893517824 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003683149448345035, + "loss": 1.458, + "theoretical_loss": 3.717356157005899, + "tokens_seen": 893583360 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036830491474423274, + "loss": 1.5573, + "theoretical_loss": 3.717329873131255, + "tokens_seen": 893648896 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036829488465396187, + "loss": 1.5556, + "theoretical_loss": 3.7173035917237405, + "tokens_seen": 893714432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003682848545636911, + "loss": 1.4955, + "theoretical_loss": 3.7172773127829437, + "tokens_seen": 893779968 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036827482447342023, + "loss": 1.5092, + "theoretical_loss": 3.7172510363084514, + "tokens_seen": 893845504 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036826479438314947, + "loss": 1.3797, + "theoretical_loss": 3.717224762299852, + "tokens_seen": 893911040 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036825476429287865, + "loss": 1.4621, + "theoretical_loss": 3.717198490756733, + "tokens_seen": 893976576 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036824473420260783, + "loss": 1.4333, + "theoretical_loss": 3.717172221678682, + "tokens_seen": 894042112 + }, + { + "epoch": 0.27, + "learning_rate": 0.000368234704112337, + "loss": 1.5048, + "theoretical_loss": 3.7171459550652877, + "tokens_seen": 894107648 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036822467402206625, + "loss": 1.5445, + "theoretical_loss": 3.717119690916138, + "tokens_seen": 894173184 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036821464393179537, + "loss": 1.5393, + "theoretical_loss": 3.717093429230821, + "tokens_seen": 894238720 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003682046138415246, + "loss": 1.4577, + "theoretical_loss": 3.7170671700089244, + "tokens_seen": 894304256 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036819458375125373, + "loss": 1.4713, + "theoretical_loss": 3.717040913250038, + "tokens_seen": 894369792 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036818455366098297, + "loss": 1.3665, + "theoretical_loss": 3.7170146589537496, + "tokens_seen": 894435328 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036817452357071215, + "loss": 1.649, + "theoretical_loss": 3.716988407119648, + "tokens_seen": 894500864 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 330336, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3623132705688477, + "objective/train/theoretical_loss": 3.7169621577473215, + "objective/train/tokens_used": 915026400, + "theoretical_loss": 3.7169621577473215, + "tokens_seen": 894566400 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036816449348044133, + "loss": 1.435, + "theoretical_loss": 3.7169621577473215, + "tokens_seen": 894566400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003681544633901705, + "loss": 1.4599, + "theoretical_loss": 3.716935910836359, + "tokens_seen": 894631936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003681444332998997, + "loss": 1.4651, + "theoretical_loss": 3.71690966638635, + "tokens_seen": 894697472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003681344032096289, + "loss": 1.4192, + "theoretical_loss": 3.716883424396883, + "tokens_seen": 894763008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003681243731193581, + "loss": 1.5875, + "theoretical_loss": 3.716857184867547, + "tokens_seen": 894828544 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036811434302908724, + "loss": 1.4215, + "theoretical_loss": 3.716830947797932, + "tokens_seen": 894894080 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003681043129388165, + "loss": 1.4322, + "theoretical_loss": 3.7168047131876265, + "tokens_seen": 894959616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003680942828485456, + "loss": 1.4524, + "theoretical_loss": 3.7167784810362203, + "tokens_seen": 895025152 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036808425275827484, + "loss": 1.5246, + "theoretical_loss": 3.7167522513433022, + "tokens_seen": 895090688 + }, + { + "epoch": 0.27, + "learning_rate": 0.000368074222668004, + "loss": 1.4762, + "theoretical_loss": 3.716726024108463, + "tokens_seen": 895156224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003680641925777332, + "loss": 1.4828, + "theoretical_loss": 3.716699799331292, + "tokens_seen": 895221760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003680541624874624, + "loss": 1.4972, + "theoretical_loss": 3.716673577011379, + "tokens_seen": 895287296 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003680441323971916, + "loss": 1.5374, + "theoretical_loss": 3.716647357148313, + "tokens_seen": 895352832 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036803410230692074, + "loss": 1.5599, + "theoretical_loss": 3.7166211397416853, + "tokens_seen": 895418368 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036802407221665, + "loss": 1.4493, + "theoretical_loss": 3.716594924791085, + "tokens_seen": 895483904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003680140421263791, + "loss": 1.5337, + "theoretical_loss": 3.7165687122961035, + "tokens_seen": 895549440 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036800401203610834, + "loss": 1.343, + "theoretical_loss": 3.71654250225633, + "tokens_seen": 895614976 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679939819458375, + "loss": 1.3709, + "theoretical_loss": 3.7165162946713552, + "tokens_seen": 895680512 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679839518555667, + "loss": 1.4391, + "theoretical_loss": 3.7164900895407698, + "tokens_seen": 895746048 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679739217652959, + "loss": 1.5509, + "theoretical_loss": 3.7164638868641635, + "tokens_seen": 895811584 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036796389167502506, + "loss": 1.5358, + "theoretical_loss": 3.716437686641129, + "tokens_seen": 895877120 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036795386158475424, + "loss": 1.4814, + "theoretical_loss": 3.7164114888712545, + "tokens_seen": 895942656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679438314944835, + "loss": 1.4295, + "theoretical_loss": 3.7163852935541333, + "tokens_seen": 896008192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679338014042126, + "loss": 1.49, + "theoretical_loss": 3.716359100689355, + "tokens_seen": 896073728 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036792377131394184, + "loss": 1.4457, + "theoretical_loss": 3.7163329102765106, + "tokens_seen": 896139264 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036791374122367097, + "loss": 1.4103, + "theoretical_loss": 3.7163067223151924, + "tokens_seen": 896204800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003679037111334002, + "loss": 1.3679, + "theoretical_loss": 3.7162805368049905, + "tokens_seen": 896270336 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003678936810431294, + "loss": 1.5088, + "theoretical_loss": 3.716254353745497, + "tokens_seen": 896335872 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036788365095285857, + "loss": 1.466, + "theoretical_loss": 3.7162281731363027, + "tokens_seen": 896401408 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036787362086258775, + "loss": 1.4298, + "theoretical_loss": 3.7162019949769998, + "tokens_seen": 896466944 + }, + { + "epoch": 0.27, + "learning_rate": 0.000367863590772317, + "loss": 1.4892, + "theoretical_loss": 3.71617581926718, + "tokens_seen": 896532480 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036785356068204616, + "loss": 1.3755, + "theoretical_loss": 3.7161496460064347, + "tokens_seen": 896598016 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036784353059177535, + "loss": 1.5028, + "theoretical_loss": 3.716123475194356, + "tokens_seen": 896663552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003678335005015045, + "loss": 1.3702, + "theoretical_loss": 3.7160973068305356, + "tokens_seen": 896729088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003678234704112337, + "loss": 1.4716, + "theoretical_loss": 3.7160711409145657, + "tokens_seen": 896794624 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036781344032096294, + "loss": 1.4617, + "theoretical_loss": 3.716044977446039, + "tokens_seen": 896860160 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036780341023069207, + "loss": 1.4493, + "theoretical_loss": 3.7160188164245467, + "tokens_seen": 896925696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003677933801404213, + "loss": 1.4459, + "theoretical_loss": 3.7159926578496822, + "tokens_seen": 896991232 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036778335005015043, + "loss": 1.3788, + "theoretical_loss": 3.715966501721037, + "tokens_seen": 897056768 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036777331995987967, + "loss": 1.4917, + "theoretical_loss": 3.7159403480382043, + "tokens_seen": 897122304 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036776328986960885, + "loss": 1.6165, + "theoretical_loss": 3.715914196800777, + "tokens_seen": 897187840 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036775325977933803, + "loss": 1.4329, + "theoretical_loss": 3.715888048008347, + "tokens_seen": 897253376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003677432296890672, + "loss": 1.4813, + "theoretical_loss": 3.7158619016605074, + "tokens_seen": 897318912 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036773319959879645, + "loss": 1.5262, + "theoretical_loss": 3.715835757756852, + "tokens_seen": 897384448 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036772316950852557, + "loss": 1.4006, + "theoretical_loss": 3.7158096162969727, + "tokens_seen": 897449984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003677131394182548, + "loss": 1.3923, + "theoretical_loss": 3.715783477280463, + "tokens_seen": 897515520 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036770310932798393, + "loss": 1.4669, + "theoretical_loss": 3.715757340706917, + "tokens_seen": 897581056 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036769307923771317, + "loss": 1.3283, + "theoretical_loss": 3.715731206575927, + "tokens_seen": 897646592 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036768304914744235, + "loss": 1.3491, + "theoretical_loss": 3.7157050748870866, + "tokens_seen": 897712128 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036767301905717153, + "loss": 1.387, + "theoretical_loss": 3.715678945639989, + "tokens_seen": 897777664 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 331577, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4502708911895752, + "objective/train/theoretical_loss": 3.7156528188342284, + "objective/train/tokens_used": 918303200, + "theoretical_loss": 3.7156528188342284, + "tokens_seen": 897843200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676629889669007, + "loss": 1.4978, + "theoretical_loss": 3.7156528188342284, + "tokens_seen": 897843200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676529588766299, + "loss": 1.4633, + "theoretical_loss": 3.7156266944693987, + "tokens_seen": 897908736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676429287863591, + "loss": 1.5919, + "theoretical_loss": 3.7156005725450933, + "tokens_seen": 897974272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676328986960883, + "loss": 1.4128, + "theoretical_loss": 3.7155744530609063, + "tokens_seen": 898039808 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036762286860581744, + "loss": 1.4943, + "theoretical_loss": 3.7155483360164316, + "tokens_seen": 898105344 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676128385155467, + "loss": 1.362, + "theoretical_loss": 3.7155222214112635, + "tokens_seen": 898170880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003676028084252758, + "loss": 1.4261, + "theoretical_loss": 3.715496109244996, + "tokens_seen": 898236416 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036759277833500504, + "loss": 1.5836, + "theoretical_loss": 3.715469999517224, + "tokens_seen": 898301952 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675827482447342, + "loss": 1.4847, + "theoretical_loss": 3.7154438922275403, + "tokens_seen": 898367488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675727181544634, + "loss": 1.4816, + "theoretical_loss": 3.7154177873755416, + "tokens_seen": 898433024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675626880641926, + "loss": 1.4447, + "theoretical_loss": 3.715391684960821, + "tokens_seen": 898498560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675526579739218, + "loss": 1.4626, + "theoretical_loss": 3.715365584982973, + "tokens_seen": 898564096 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036754262788365094, + "loss": 1.473, + "theoretical_loss": 3.7153394874415935, + "tokens_seen": 898629632 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675325977933802, + "loss": 1.4345, + "theoretical_loss": 3.7153133923362764, + "tokens_seen": 898695168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675225677031093, + "loss": 1.4215, + "theoretical_loss": 3.7152872996666177, + "tokens_seen": 898760704 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036751253761283854, + "loss": 1.5043, + "theoretical_loss": 3.715261209432212, + "tokens_seen": 898826240 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003675025075225677, + "loss": 1.4975, + "theoretical_loss": 3.715235121632654, + "tokens_seen": 898891776 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674924774322969, + "loss": 1.4488, + "theoretical_loss": 3.7152090362675394, + "tokens_seen": 898957312 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674824473420261, + "loss": 1.4415, + "theoretical_loss": 3.7151829533364635, + "tokens_seen": 899022848 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036747241725175526, + "loss": 1.5538, + "theoretical_loss": 3.715156872839022, + "tokens_seen": 899088384 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036746238716148444, + "loss": 1.41, + "theoretical_loss": 3.7151307947748107, + "tokens_seen": 899153920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674523570712137, + "loss": 1.4749, + "theoretical_loss": 3.715104719143424, + "tokens_seen": 899219456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674423269809428, + "loss": 1.421, + "theoretical_loss": 3.715078645944459, + "tokens_seen": 899284992 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036743229689067204, + "loss": 1.4362, + "theoretical_loss": 3.715052575177511, + "tokens_seen": 899350528 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036742226680040117, + "loss": 1.438, + "theoretical_loss": 3.715026506842176, + "tokens_seen": 899416064 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674122367101304, + "loss": 1.3688, + "theoretical_loss": 3.7150004409380504, + "tokens_seen": 899481600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003674022066198596, + "loss": 1.5637, + "theoretical_loss": 3.7149743774647295, + "tokens_seen": 899547136 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036739217652958877, + "loss": 1.4786, + "theoretical_loss": 3.7149483164218102, + "tokens_seen": 899612672 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036738214643931795, + "loss": 1.4021, + "theoretical_loss": 3.714922257808889, + "tokens_seen": 899678208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003673721163490472, + "loss": 1.4302, + "theoretical_loss": 3.7148962016255616, + "tokens_seen": 899743744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003673620862587763, + "loss": 1.5384, + "theoretical_loss": 3.714870147871425, + "tokens_seen": 899809280 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036735205616850555, + "loss": 1.4253, + "theoretical_loss": 3.7148440965460763, + "tokens_seen": 899874816 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036734202607823467, + "loss": 1.3683, + "theoretical_loss": 3.7148180476491115, + "tokens_seen": 899940352 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003673319959879639, + "loss": 1.4957, + "theoretical_loss": 3.714792001180127, + "tokens_seen": 900005888 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003673219658976931, + "loss": 1.4667, + "theoretical_loss": 3.714765957138721, + "tokens_seen": 900071424 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036731193580742227, + "loss": 1.4104, + "theoretical_loss": 3.71473991552449, + "tokens_seen": 900136960 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036730190571715145, + "loss": 1.3587, + "theoretical_loss": 3.71471387633703, + "tokens_seen": 900202496 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036729187562688063, + "loss": 1.5303, + "theoretical_loss": 3.7146878395759404, + "tokens_seen": 900268032 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003672818455366098, + "loss": 1.5819, + "theoretical_loss": 3.7146618052408167, + "tokens_seen": 900333568 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036727181544633905, + "loss": 1.4776, + "theoretical_loss": 3.714635773331257, + "tokens_seen": 900399104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003672617853560682, + "loss": 1.5168, + "theoretical_loss": 3.7146097438468586, + "tokens_seen": 900464640 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003672517552657974, + "loss": 1.491, + "theoretical_loss": 3.714583716787219, + "tokens_seen": 900530176 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036724172517552654, + "loss": 1.4013, + "theoretical_loss": 3.7145576921519363, + "tokens_seen": 900595712 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036723169508525577, + "loss": 1.4436, + "theoretical_loss": 3.714531669940608, + "tokens_seen": 900661248 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036722166499498495, + "loss": 1.4908, + "theoretical_loss": 3.714505650152832, + "tokens_seen": 900726784 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036721163490471414, + "loss": 1.4154, + "theoretical_loss": 3.7144796327882066, + "tokens_seen": 900792320 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003672016048144433, + "loss": 1.4286, + "theoretical_loss": 3.714453617846329, + "tokens_seen": 900857856 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036719157472417255, + "loss": 1.4569, + "theoretical_loss": 3.714427605326798, + "tokens_seen": 900923392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003671815446339017, + "loss": 1.5572, + "theoretical_loss": 3.714401595229212, + "tokens_seen": 900988928 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003671715145436309, + "loss": 1.4284, + "theoretical_loss": 3.714375587553169, + "tokens_seen": 901054464 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 332819, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5798795223236084, + "objective/train/theoretical_loss": 3.714349582298268, + "objective/train/tokens_used": 921580000, + "theoretical_loss": 3.714349582298268, + "tokens_seen": 901120000 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036716148445336004, + "loss": 1.5218, + "theoretical_loss": 3.714349582298268, + "tokens_seen": 901120000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003671514543630893, + "loss": 1.3866, + "theoretical_loss": 3.7143235794641067, + "tokens_seen": 901185536 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036714142427281846, + "loss": 1.4684, + "theoretical_loss": 3.7142975790502843, + "tokens_seen": 901251072 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036713139418254764, + "loss": 1.5234, + "theoretical_loss": 3.7142715810563995, + "tokens_seen": 901316608 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003671213640922768, + "loss": 1.4972, + "theoretical_loss": 3.7142455854820513, + "tokens_seen": 901382144 + }, + { + "epoch": 0.27, + "learning_rate": 0.000367111334002006, + "loss": 1.4508, + "theoretical_loss": 3.714219592326838, + "tokens_seen": 901447680 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036710130391173524, + "loss": 1.4752, + "theoretical_loss": 3.714193601590359, + "tokens_seen": 901513216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670912738214644, + "loss": 1.529, + "theoretical_loss": 3.7141676132722137, + "tokens_seen": 901578752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670812437311936, + "loss": 1.4583, + "theoretical_loss": 3.714141627372001, + "tokens_seen": 901644288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670712136409228, + "loss": 1.455, + "theoretical_loss": 3.7141156438893206, + "tokens_seen": 901709824 + }, + { + "epoch": 0.27, + "learning_rate": 0.000367061183550652, + "loss": 1.5062, + "theoretical_loss": 3.7140896628237714, + "tokens_seen": 901775360 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036705115346038114, + "loss": 1.4204, + "theoretical_loss": 3.7140636841749526, + "tokens_seen": 901840896 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670411233701104, + "loss": 1.3972, + "theoretical_loss": 3.714037707942465, + "tokens_seen": 901906432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670310932798395, + "loss": 1.553, + "theoretical_loss": 3.714011734125907, + "tokens_seen": 901971968 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036702106318956874, + "loss": 1.4895, + "theoretical_loss": 3.7139857627248793, + "tokens_seen": 902037504 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670110330992979, + "loss": 1.5165, + "theoretical_loss": 3.7139597937389817, + "tokens_seen": 902103040 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003670010030090271, + "loss": 1.5359, + "theoretical_loss": 3.7139338271678133, + "tokens_seen": 902168576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003669909729187563, + "loss": 1.4377, + "theoretical_loss": 3.7139078630109754, + "tokens_seen": 902234112 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036698094282848546, + "loss": 1.3671, + "theoretical_loss": 3.713881901268068, + "tokens_seen": 902299648 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036697091273821464, + "loss": 1.4612, + "theoretical_loss": 3.7138559419386903, + "tokens_seen": 902365184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003669608826479439, + "loss": 1.4359, + "theoretical_loss": 3.713829985022443, + "tokens_seen": 902430720 + }, + { + "epoch": 0.27, + "learning_rate": 0.000366950852557673, + "loss": 1.4147, + "theoretical_loss": 3.7138040305189275, + "tokens_seen": 902496256 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036694082246740224, + "loss": 1.5695, + "theoretical_loss": 3.713778078427744, + "tokens_seen": 902561792 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036693079237713137, + "loss": 1.4592, + "theoretical_loss": 3.7137521287484923, + "tokens_seen": 902627328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003669207622868606, + "loss": 1.4393, + "theoretical_loss": 3.713726181480774, + "tokens_seen": 902692864 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003669107321965898, + "loss": 1.4344, + "theoretical_loss": 3.71370023662419, + "tokens_seen": 902758400 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036690070210631897, + "loss": 1.5302, + "theoretical_loss": 3.7136742941783405, + "tokens_seen": 902823936 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036689067201604815, + "loss": 1.4615, + "theoretical_loss": 3.713648354142827, + "tokens_seen": 902889472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003668806419257774, + "loss": 1.5839, + "theoretical_loss": 3.713622416517251, + "tokens_seen": 902955008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003668706118355065, + "loss": 1.4293, + "theoretical_loss": 3.7135964813012126, + "tokens_seen": 903020544 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036686058174523575, + "loss": 1.4073, + "theoretical_loss": 3.7135705484943142, + "tokens_seen": 903086080 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036685055165496487, + "loss": 1.5704, + "theoretical_loss": 3.713544618096157, + "tokens_seen": 903151616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003668405215646941, + "loss": 1.3086, + "theoretical_loss": 3.713518690106342, + "tokens_seen": 903217152 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003668304914744233, + "loss": 1.4883, + "theoretical_loss": 3.7134927645244713, + "tokens_seen": 903282688 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036682046138415247, + "loss": 1.3935, + "theoretical_loss": 3.7134668413501464, + "tokens_seen": 903348224 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036681043129388165, + "loss": 1.4542, + "theoretical_loss": 3.713440920582969, + "tokens_seen": 903413760 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036680040120361083, + "loss": 1.4002, + "theoretical_loss": 3.713415002222541, + "tokens_seen": 903479296 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036679037111334, + "loss": 1.3507, + "theoretical_loss": 3.7133890862684646, + "tokens_seen": 903544832 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036678034102306925, + "loss": 1.4485, + "theoretical_loss": 3.7133631727203418, + "tokens_seen": 903610368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003667703109327984, + "loss": 1.6004, + "theoretical_loss": 3.7133372615777747, + "tokens_seen": 903675904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003667602808425276, + "loss": 1.5642, + "theoretical_loss": 3.713311352840366, + "tokens_seen": 903741440 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036675025075225674, + "loss": 1.3765, + "theoretical_loss": 3.713285446507717, + "tokens_seen": 903806976 + }, + { + "epoch": 0.27, + "learning_rate": 0.000366740220661986, + "loss": 1.4933, + "theoretical_loss": 3.7132595425794306, + "tokens_seen": 903872512 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036673019057171515, + "loss": 1.5064, + "theoretical_loss": 3.7132336410551097, + "tokens_seen": 903938048 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036672016048144434, + "loss": 1.5319, + "theoretical_loss": 3.713207741934357, + "tokens_seen": 904003584 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003667101303911735, + "loss": 1.4461, + "theoretical_loss": 3.713181845216775, + "tokens_seen": 904069120 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036670010030090275, + "loss": 1.406, + "theoretical_loss": 3.7131559509019665, + "tokens_seen": 904134656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003666900702106319, + "loss": 1.3801, + "theoretical_loss": 3.713130058989534, + "tokens_seen": 904200192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003666800401203611, + "loss": 1.3573, + "theoretical_loss": 3.713104169479082, + "tokens_seen": 904265728 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036667001003009024, + "loss": 1.5021, + "theoretical_loss": 3.713078282370211, + "tokens_seen": 904331264 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 333943, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.398763656616211, + "objective/train/theoretical_loss": 3.713052397662527, + "objective/train/tokens_used": 924856800, + "theoretical_loss": 3.713052397662527, + "tokens_seen": 904396800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003666599799398195, + "loss": 1.4082, + "theoretical_loss": 3.713052397662527, + "tokens_seen": 904396800 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036664994984954866, + "loss": 1.4509, + "theoretical_loss": 3.713026515355632, + "tokens_seen": 904462336 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036663991975927784, + "loss": 1.4089, + "theoretical_loss": 3.713000635449129, + "tokens_seen": 904527872 + }, + { + "epoch": 0.27, + "learning_rate": 0.000366629889669007, + "loss": 1.3632, + "theoretical_loss": 3.7129747579426224, + "tokens_seen": 904593408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003666198595787362, + "loss": 1.4396, + "theoretical_loss": 3.712948882835715, + "tokens_seen": 904658944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003666098294884654, + "loss": 1.4971, + "theoretical_loss": 3.7129230101280113, + "tokens_seen": 904724480 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003665997993981946, + "loss": 1.4247, + "theoretical_loss": 3.7128971398191144, + "tokens_seen": 904790016 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036658976930792374, + "loss": 1.601, + "theoretical_loss": 3.712871271908628, + "tokens_seen": 904855552 + }, + { + "epoch": 0.27, + "learning_rate": 0.000366579739217653, + "loss": 1.4481, + "theoretical_loss": 3.7128454063961565, + "tokens_seen": 904921088 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036656970912738216, + "loss": 1.4628, + "theoretical_loss": 3.712819543281304, + "tokens_seen": 904986624 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036655967903711134, + "loss": 1.4763, + "theoretical_loss": 3.712793682563675, + "tokens_seen": 905052160 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003665496489468405, + "loss": 1.4512, + "theoretical_loss": 3.712767824242873, + "tokens_seen": 905117696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003665396188565697, + "loss": 1.4047, + "theoretical_loss": 3.712741968318502, + "tokens_seen": 905183232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003665295887662989, + "loss": 1.4886, + "theoretical_loss": 3.712716114790168, + "tokens_seen": 905248768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003665195586760281, + "loss": 1.4299, + "theoretical_loss": 3.712690263657474, + "tokens_seen": 905314304 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036650952858575725, + "loss": 1.4284, + "theoretical_loss": 3.7126644149200247, + "tokens_seen": 905379840 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003664994984954865, + "loss": 1.598, + "theoretical_loss": 3.712638568577426, + "tokens_seen": 905445376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003664894684052156, + "loss": 1.4286, + "theoretical_loss": 3.712612724629282, + "tokens_seen": 905510912 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036647943831494484, + "loss": 1.3818, + "theoretical_loss": 3.712586883075198, + "tokens_seen": 905576448 + }, + { + "epoch": 0.27, + "learning_rate": 0.000366469408224674, + "loss": 1.3763, + "theoretical_loss": 3.7125610439147776, + "tokens_seen": 905641984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003664593781344032, + "loss": 1.4937, + "theoretical_loss": 3.7125352071476274, + "tokens_seen": 905707520 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003664493480441324, + "loss": 1.3543, + "theoretical_loss": 3.7125093727733516, + "tokens_seen": 905773056 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036643931795386157, + "loss": 1.3333, + "theoretical_loss": 3.7124835407915566, + "tokens_seen": 905838592 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036642928786359075, + "loss": 1.5157, + "theoretical_loss": 3.712457711201847, + "tokens_seen": 905904128 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036641925777332, + "loss": 1.5542, + "theoretical_loss": 3.712431884003828, + "tokens_seen": 905969664 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003664092276830491, + "loss": 1.4263, + "theoretical_loss": 3.7124060591971055, + "tokens_seen": 906035200 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036639919759277835, + "loss": 1.4833, + "theoretical_loss": 3.7123802367812857, + "tokens_seen": 906100736 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036638916750250753, + "loss": 1.3587, + "theoretical_loss": 3.712354416755973, + "tokens_seen": 906166272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003663791374122367, + "loss": 1.5492, + "theoretical_loss": 3.7123285991207746, + "tokens_seen": 906231808 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003663691073219659, + "loss": 1.4686, + "theoretical_loss": 3.712302783875296, + "tokens_seen": 906297344 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036635907723169507, + "loss": 1.415, + "theoretical_loss": 3.7122769710191426, + "tokens_seen": 906362880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003663490471414243, + "loss": 1.5223, + "theoretical_loss": 3.7122511605519217, + "tokens_seen": 906428416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003663390170511535, + "loss": 1.2935, + "theoretical_loss": 3.7122253524732383, + "tokens_seen": 906493952 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036632898696088267, + "loss": 1.3694, + "theoretical_loss": 3.712199546782699, + "tokens_seen": 906559488 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036631895687061185, + "loss": 1.4253, + "theoretical_loss": 3.7121737434799114, + "tokens_seen": 906625024 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036630892678034103, + "loss": 1.4519, + "theoretical_loss": 3.71214794256448, + "tokens_seen": 906690560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003662988966900702, + "loss": 1.4308, + "theoretical_loss": 3.712122144036013, + "tokens_seen": 906756096 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036628886659979945, + "loss": 1.5343, + "theoretical_loss": 3.712096347894116, + "tokens_seen": 906821632 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003662788365095286, + "loss": 1.3877, + "theoretical_loss": 3.7120705541383967, + "tokens_seen": 906887168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003662688064192578, + "loss": 1.4256, + "theoretical_loss": 3.7120447627684614, + "tokens_seen": 906952704 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036625877632898694, + "loss": 1.5254, + "theoretical_loss": 3.712018973783917, + "tokens_seen": 907018240 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003662487462387162, + "loss": 1.4255, + "theoretical_loss": 3.7119931871843708, + "tokens_seen": 907083776 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036623871614844535, + "loss": 1.3436, + "theoretical_loss": 3.71196740296943, + "tokens_seen": 907149312 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036622868605817454, + "loss": 1.5039, + "theoretical_loss": 3.7119416211387013, + "tokens_seen": 907214848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003662186559679037, + "loss": 1.4693, + "theoretical_loss": 3.7119158416917926, + "tokens_seen": 907280384 + }, + { + "epoch": 0.27, + "learning_rate": 0.00036620862587763295, + "loss": 1.3892, + "theoretical_loss": 3.711890064628311, + "tokens_seen": 907345920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003661985957873621, + "loss": 1.4225, + "theoretical_loss": 3.711864289947864, + "tokens_seen": 907411456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0003661885656970913, + "loss": 1.4308, + "theoretical_loss": 3.71183851765006, + "tokens_seen": 907476992 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036617853560682044, + "loss": 1.4497, + "theoretical_loss": 3.7118127477345055, + "tokens_seen": 907542528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003661685055165497, + "loss": 1.3964, + "theoretical_loss": 3.7117869802008094, + "tokens_seen": 907608064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 335091, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3655242919921875, + "objective/train/theoretical_loss": 3.7117612150485786, + "objective/train/tokens_used": 928133600, + "theoretical_loss": 3.7117612150485786, + "tokens_seen": 907673600 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036615847542627886, + "loss": 1.3873, + "theoretical_loss": 3.7117612150485786, + "tokens_seen": 907673600 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036614844533600804, + "loss": 1.4594, + "theoretical_loss": 3.7117354522774217, + "tokens_seen": 907739136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003661384152457372, + "loss": 1.435, + "theoretical_loss": 3.7117096918869468, + "tokens_seen": 907804672 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003661283851554664, + "loss": 1.5111, + "theoretical_loss": 3.7116839338767615, + "tokens_seen": 907870208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003661183550651956, + "loss": 1.3948, + "theoretical_loss": 3.7116581782464744, + "tokens_seen": 907935744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003661083249749248, + "loss": 1.5285, + "theoretical_loss": 3.711632424995694, + "tokens_seen": 908001280 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036609829488465394, + "loss": 1.5422, + "theoretical_loss": 3.7116066741240283, + "tokens_seen": 908066816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660882647943832, + "loss": 1.38, + "theoretical_loss": 3.7115809256310865, + "tokens_seen": 908132352 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036607823470411236, + "loss": 1.4057, + "theoretical_loss": 3.711555179516477, + "tokens_seen": 908197888 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036606820461384154, + "loss": 1.3667, + "theoretical_loss": 3.7115294357798083, + "tokens_seen": 908263424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660581745235707, + "loss": 1.4622, + "theoretical_loss": 3.7115036944206885, + "tokens_seen": 908328960 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660481444332999, + "loss": 1.5321, + "theoretical_loss": 3.7114779554387285, + "tokens_seen": 908394496 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660381143430291, + "loss": 1.4939, + "theoretical_loss": 3.7114522188335357, + "tokens_seen": 908460032 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660280842527583, + "loss": 1.4094, + "theoretical_loss": 3.7114264846047194, + "tokens_seen": 908525568 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036601805416248745, + "loss": 1.4553, + "theoretical_loss": 3.7114007527518886, + "tokens_seen": 908591104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003660080240722167, + "loss": 1.4075, + "theoretical_loss": 3.7113750232746536, + "tokens_seen": 908656640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659979939819458, + "loss": 1.4728, + "theoretical_loss": 3.711349296172623, + "tokens_seen": 908722176 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036598796389167504, + "loss": 1.4414, + "theoretical_loss": 3.7113235714454063, + "tokens_seen": 908787712 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659779338014042, + "loss": 1.4981, + "theoretical_loss": 3.711297849092613, + "tokens_seen": 908853248 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659679037111334, + "loss": 1.4463, + "theoretical_loss": 3.711272129113853, + "tokens_seen": 908918784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659578736208626, + "loss": 1.3897, + "theoretical_loss": 3.7112464115087365, + "tokens_seen": 908984320 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036594784353059177, + "loss": 1.4311, + "theoretical_loss": 3.7112206962768717, + "tokens_seen": 909049856 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036593781344032095, + "loss": 1.3231, + "theoretical_loss": 3.7111949834178706, + "tokens_seen": 909115392 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659277833500502, + "loss": 1.3899, + "theoretical_loss": 3.7111692729313415, + "tokens_seen": 909180928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003659177532597793, + "loss": 1.5008, + "theoretical_loss": 3.7111435648168953, + "tokens_seen": 909246464 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036590772316950855, + "loss": 1.4804, + "theoretical_loss": 3.7111178590741423, + "tokens_seen": 909312000 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036589769307923773, + "loss": 1.3806, + "theoretical_loss": 3.7110921557026924, + "tokens_seen": 909377536 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658876629889669, + "loss": 1.5133, + "theoretical_loss": 3.7110664547021566, + "tokens_seen": 909443072 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658776328986961, + "loss": 1.3389, + "theoretical_loss": 3.711040756072144, + "tokens_seen": 909508608 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036586760280842527, + "loss": 1.6119, + "theoretical_loss": 3.711015059812267, + "tokens_seen": 909574144 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036585757271815445, + "loss": 1.385, + "theoretical_loss": 3.710989365922135, + "tokens_seen": 909639680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658475426278837, + "loss": 1.4871, + "theoretical_loss": 3.710963674401359, + "tokens_seen": 909705216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658375125376128, + "loss": 1.4545, + "theoretical_loss": 3.7109379852495503, + "tokens_seen": 909770752 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036582748244734205, + "loss": 1.4918, + "theoretical_loss": 3.7109122984663196, + "tokens_seen": 909836288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658174523570712, + "loss": 1.4456, + "theoretical_loss": 3.7108866140512777, + "tokens_seen": 909901824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003658074222668004, + "loss": 1.4082, + "theoretical_loss": 3.7108609320040356, + "tokens_seen": 909967360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657973921765296, + "loss": 1.4797, + "theoretical_loss": 3.7108352523242045, + "tokens_seen": 910032896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657873620862588, + "loss": 1.4638, + "theoretical_loss": 3.710809575011397, + "tokens_seen": 910098432 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036577733199598796, + "loss": 1.4548, + "theoretical_loss": 3.7107839000652225, + "tokens_seen": 910163968 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036576730190571714, + "loss": 1.5099, + "theoretical_loss": 3.710758227485294, + "tokens_seen": 910229504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657572718154463, + "loss": 1.3896, + "theoretical_loss": 3.710732557271222, + "tokens_seen": 910295040 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036574724172517555, + "loss": 1.396, + "theoretical_loss": 3.710706889422619, + "tokens_seen": 910360576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657372116349047, + "loss": 1.4743, + "theoretical_loss": 3.7106812239390967, + "tokens_seen": 910426112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657271815446339, + "loss": 1.3694, + "theoretical_loss": 3.7106555608202667, + "tokens_seen": 910491648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657171514543631, + "loss": 1.448, + "theoretical_loss": 3.71062990006574, + "tokens_seen": 910557184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003657071213640923, + "loss": 1.3791, + "theoretical_loss": 3.7106042416751306, + "tokens_seen": 910622720 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036569709127382146, + "loss": 1.4394, + "theoretical_loss": 3.7105785856480495, + "tokens_seen": 910688256 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036568706118355064, + "loss": 1.432, + "theoretical_loss": 3.710552931984109, + "tokens_seen": 910753792 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003656770310932798, + "loss": 1.3754, + "theoretical_loss": 3.7105272806829213, + "tokens_seen": 910819328 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036566700100300906, + "loss": 1.4125, + "theoretical_loss": 3.7105016317440986, + "tokens_seen": 910884864 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 336141, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.501374363899231, + "objective/train/theoretical_loss": 3.710475985167254, + "objective/train/tokens_used": 931410400, + "theoretical_loss": 3.710475985167254, + "tokens_seen": 910950400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003656569709127382, + "loss": 1.5129, + "theoretical_loss": 3.710475985167254, + "tokens_seen": 910950400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003656469408224674, + "loss": 1.4074, + "theoretical_loss": 3.710450340952, + "tokens_seen": 911015936 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036563691073219655, + "loss": 1.4182, + "theoretical_loss": 3.710424699097949, + "tokens_seen": 911081472 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003656268806419258, + "loss": 1.4272, + "theoretical_loss": 3.710399059604714, + "tokens_seen": 911147008 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036561685055165496, + "loss": 1.4608, + "theoretical_loss": 3.710373422471908, + "tokens_seen": 911212544 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036560682046138414, + "loss": 1.3367, + "theoretical_loss": 3.710347787699143, + "tokens_seen": 911278080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655967903711134, + "loss": 1.3577, + "theoretical_loss": 3.710322155286033, + "tokens_seen": 911343616 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036558676028084256, + "loss": 1.4131, + "theoretical_loss": 3.710296525232191, + "tokens_seen": 911409152 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036557673019057174, + "loss": 1.4019, + "theoretical_loss": 3.71027089753723, + "tokens_seen": 911474688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655667001003009, + "loss": 1.3854, + "theoretical_loss": 3.7102452722007637, + "tokens_seen": 911540224 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655566700100301, + "loss": 1.3038, + "theoretical_loss": 3.710219649222404, + "tokens_seen": 911605760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655466399197593, + "loss": 1.4234, + "theoretical_loss": 3.710194028601767, + "tokens_seen": 911671296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655366098294885, + "loss": 1.4755, + "theoretical_loss": 3.7101684103384645, + "tokens_seen": 911736832 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036552657973921765, + "loss": 1.3558, + "theoretical_loss": 3.7101427944321106, + "tokens_seen": 911802368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003655165496489469, + "loss": 1.576, + "theoretical_loss": 3.7101171808823192, + "tokens_seen": 911867904 + }, + { + "epoch": 0.28, + "learning_rate": 0.000365506519558676, + "loss": 1.4895, + "theoretical_loss": 3.7100915696887036, + "tokens_seen": 911933440 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036549648946840525, + "loss": 1.3619, + "theoretical_loss": 3.710065960850878, + "tokens_seen": 911998976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003654864593781344, + "loss": 1.4698, + "theoretical_loss": 3.710040354368457, + "tokens_seen": 912064512 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003654764292878636, + "loss": 1.5466, + "theoretical_loss": 3.7100147502410543, + "tokens_seen": 912130048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003654663991975928, + "loss": 1.4713, + "theoretical_loss": 3.709989148468284, + "tokens_seen": 912195584 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036545636910732197, + "loss": 1.4625, + "theoretical_loss": 3.7099635490497604, + "tokens_seen": 912261120 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036544633901705115, + "loss": 1.4498, + "theoretical_loss": 3.709937951985099, + "tokens_seen": 912326656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003654363089267804, + "loss": 1.5513, + "theoretical_loss": 3.7099123572739128, + "tokens_seen": 912392192 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003654262788365095, + "loss": 1.3789, + "theoretical_loss": 3.709886764915817, + "tokens_seen": 912457728 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036541624874623875, + "loss": 1.4336, + "theoretical_loss": 3.7098611749104258, + "tokens_seen": 912523264 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036540621865596793, + "loss": 1.4504, + "theoretical_loss": 3.709835587257355, + "tokens_seen": 912588800 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653961885656971, + "loss": 1.4288, + "theoretical_loss": 3.7098100019562192, + "tokens_seen": 912654336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653861584754263, + "loss": 1.3178, + "theoretical_loss": 3.7097844190066325, + "tokens_seen": 912719872 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036537612838515547, + "loss": 1.4365, + "theoretical_loss": 3.7097588384082107, + "tokens_seen": 912785408 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036536609829488465, + "loss": 1.4734, + "theoretical_loss": 3.7097332601605686, + "tokens_seen": 912850944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653560682046139, + "loss": 1.3808, + "theoretical_loss": 3.7097076842633214, + "tokens_seen": 912916480 + }, + { + "epoch": 0.28, + "learning_rate": 0.000365346038114343, + "loss": 1.3749, + "theoretical_loss": 3.709682110716085, + "tokens_seen": 912982016 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036533600802407225, + "loss": 1.4762, + "theoretical_loss": 3.709656539518474, + "tokens_seen": 913047552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653259779338014, + "loss": 1.4592, + "theoretical_loss": 3.709630970670105, + "tokens_seen": 913113088 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653159478435306, + "loss": 1.4941, + "theoretical_loss": 3.709605404170592, + "tokens_seen": 913178624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003653059177532598, + "loss": 1.4351, + "theoretical_loss": 3.7095798400195523, + "tokens_seen": 913244160 + }, + { + "epoch": 0.28, + "learning_rate": 0.000365295887662989, + "loss": 1.4436, + "theoretical_loss": 3.7095542782166, + "tokens_seen": 913309696 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036528585757271816, + "loss": 1.3864, + "theoretical_loss": 3.7095287187613524, + "tokens_seen": 913375232 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036527582748244734, + "loss": 1.3582, + "theoretical_loss": 3.709503161653425, + "tokens_seen": 913440768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003652657973921765, + "loss": 1.5292, + "theoretical_loss": 3.7094776068924338, + "tokens_seen": 913506304 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036525576730190575, + "loss": 1.4352, + "theoretical_loss": 3.7094520544779948, + "tokens_seen": 913571840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003652457372116349, + "loss": 1.4753, + "theoretical_loss": 3.7094265044097243, + "tokens_seen": 913637376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003652357071213641, + "loss": 1.5176, + "theoretical_loss": 3.7094009566872383, + "tokens_seen": 913702912 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003652256770310933, + "loss": 1.4013, + "theoretical_loss": 3.7093754113101536, + "tokens_seen": 913768448 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003652156469408225, + "loss": 1.3205, + "theoretical_loss": 3.709349868278087, + "tokens_seen": 913833984 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036520561685055166, + "loss": 1.3927, + "theoretical_loss": 3.7093243275906547, + "tokens_seen": 913899520 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036519558676028084, + "loss": 1.4796, + "theoretical_loss": 3.7092987892474727, + "tokens_seen": 913965056 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036518555667001, + "loss": 1.4754, + "theoretical_loss": 3.709273253248159, + "tokens_seen": 914030592 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036517552657973926, + "loss": 1.3845, + "theoretical_loss": 3.7092477195923292, + "tokens_seen": 914096128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003651654964894684, + "loss": 1.3709, + "theoretical_loss": 3.709222188279602, + "tokens_seen": 914161664 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 337261, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4515018463134766, + "objective/train/theoretical_loss": 3.709196659309592, + "objective/train/tokens_used": 934687200, + "theoretical_loss": 3.709196659309592, + "tokens_seen": 914227200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003651554663991976, + "loss": 1.5311, + "theoretical_loss": 3.709196659309592, + "tokens_seen": 914227200 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036514543630892675, + "loss": 1.395, + "theoretical_loss": 3.7091711326819192, + "tokens_seen": 914292736 + }, + { + "epoch": 0.28, + "learning_rate": 0.000365135406218656, + "loss": 1.4503, + "theoretical_loss": 3.7091456083961982, + "tokens_seen": 914358272 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036512537612838516, + "loss": 1.4028, + "theoretical_loss": 3.7091200864520477, + "tokens_seen": 914423808 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036511534603811434, + "loss": 1.4551, + "theoretical_loss": 3.709094566849085, + "tokens_seen": 914489344 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003651053159478435, + "loss": 1.4293, + "theoretical_loss": 3.709069049586927, + "tokens_seen": 914554880 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036509528585757276, + "loss": 1.4345, + "theoretical_loss": 3.7090435346651915, + "tokens_seen": 914620416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003650852557673019, + "loss": 1.4356, + "theoretical_loss": 3.7090180220834967, + "tokens_seen": 914685952 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003650752256770311, + "loss": 1.412, + "theoretical_loss": 3.7089925118414597, + "tokens_seen": 914751488 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036506519558676025, + "loss": 1.4241, + "theoretical_loss": 3.7089670039386986, + "tokens_seen": 914817024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003650551654964895, + "loss": 1.4142, + "theoretical_loss": 3.7089414983748314, + "tokens_seen": 914882560 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036504513540621867, + "loss": 1.3657, + "theoretical_loss": 3.708915995149476, + "tokens_seen": 914948096 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036503510531594785, + "loss": 1.4325, + "theoretical_loss": 3.70889049426225, + "tokens_seen": 915013632 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036502507522567703, + "loss": 1.4694, + "theoretical_loss": 3.708864995712773, + "tokens_seen": 915079168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003650150451354062, + "loss": 1.4585, + "theoretical_loss": 3.708839499500662, + "tokens_seen": 915144704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003650050150451354, + "loss": 1.4749, + "theoretical_loss": 3.708814005625536, + "tokens_seen": 915210240 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003649949849548646, + "loss": 1.4648, + "theoretical_loss": 3.708788514087013, + "tokens_seen": 915275776 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036498495486459375, + "loss": 1.4447, + "theoretical_loss": 3.7087630248847123, + "tokens_seen": 915341312 + }, + { + "epoch": 0.28, + "learning_rate": 0.000364974924774323, + "loss": 1.4197, + "theoretical_loss": 3.7087375380182515, + "tokens_seen": 915406848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003649648946840521, + "loss": 1.3688, + "theoretical_loss": 3.7087120534872504, + "tokens_seen": 915472384 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036495486459378135, + "loss": 1.4772, + "theoretical_loss": 3.708686571291327, + "tokens_seen": 915537920 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036494483450351053, + "loss": 1.4447, + "theoretical_loss": 3.7086610914301006, + "tokens_seen": 915603456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003649348044132397, + "loss": 1.3771, + "theoretical_loss": 3.70863561390319, + "tokens_seen": 915668992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003649247743229689, + "loss": 1.4489, + "theoretical_loss": 3.708610138710215, + "tokens_seen": 915734528 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036491474423269813, + "loss": 1.4457, + "theoretical_loss": 3.708584665850794, + "tokens_seen": 915800064 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036490471414242726, + "loss": 1.4268, + "theoretical_loss": 3.7085591953245465, + "tokens_seen": 915865600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648946840521565, + "loss": 1.4094, + "theoretical_loss": 3.708533727131092, + "tokens_seen": 915931136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648846539618856, + "loss": 1.4326, + "theoretical_loss": 3.7085082612700493, + "tokens_seen": 915996672 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036487462387161485, + "loss": 1.462, + "theoretical_loss": 3.7084827977410395, + "tokens_seen": 916062208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648645937813441, + "loss": 1.446, + "theoretical_loss": 3.70845733654368, + "tokens_seen": 916127744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648545636910732, + "loss": 1.5881, + "theoretical_loss": 3.708431877677593, + "tokens_seen": 916193280 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036484453360080245, + "loss": 1.4718, + "theoretical_loss": 3.708406421142396, + "tokens_seen": 916258816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648345035105316, + "loss": 1.4706, + "theoretical_loss": 3.7083809669377104, + "tokens_seen": 916324352 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648244734202608, + "loss": 1.497, + "theoretical_loss": 3.7083555150631557, + "tokens_seen": 916389888 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036481444332999, + "loss": 1.3581, + "theoretical_loss": 3.708330065518352, + "tokens_seen": 916455424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003648044132397192, + "loss": 1.5855, + "theoretical_loss": 3.70830461830292, + "tokens_seen": 916520960 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036479438314944836, + "loss": 1.2957, + "theoretical_loss": 3.7082791734164786, + "tokens_seen": 916586496 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036478435305917754, + "loss": 1.3752, + "theoretical_loss": 3.708253730858649, + "tokens_seen": 916652032 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003647743229689067, + "loss": 1.4381, + "theoretical_loss": 3.7082282906290525, + "tokens_seen": 916717568 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036476429287863595, + "loss": 1.4105, + "theoretical_loss": 3.7082028527273083, + "tokens_seen": 916783104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003647542627883651, + "loss": 1.3811, + "theoretical_loss": 3.708177417153037, + "tokens_seen": 916848640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003647442326980943, + "loss": 1.4458, + "theoretical_loss": 3.70815198390586, + "tokens_seen": 916914176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003647342026078235, + "loss": 1.4196, + "theoretical_loss": 3.708126552985398, + "tokens_seen": 916979712 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003647241725175527, + "loss": 1.2641, + "theoretical_loss": 3.7081011243912716, + "tokens_seen": 917045248 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036471414242728186, + "loss": 1.4052, + "theoretical_loss": 3.7080756981231016, + "tokens_seen": 917110784 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036470411233701104, + "loss": 1.3453, + "theoretical_loss": 3.7080502741805104, + "tokens_seen": 917176320 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003646940822467402, + "loss": 1.5436, + "theoretical_loss": 3.708024852563117, + "tokens_seen": 917241856 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036468405215646946, + "loss": 1.4373, + "theoretical_loss": 3.707999433270544, + "tokens_seen": 917307392 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003646740220661986, + "loss": 1.4162, + "theoretical_loss": 3.7079740163024124, + "tokens_seen": 917372928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003646639919759278, + "loss": 1.4616, + "theoretical_loss": 3.7079486016583436, + "tokens_seen": 917438464 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 338425, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5196173191070557, + "objective/train/theoretical_loss": 3.7079231893379596, + "objective/train/tokens_used": 937964000, + "theoretical_loss": 3.7079231893379596, + "tokens_seen": 917504000 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036465396188565695, + "loss": 1.5509, + "theoretical_loss": 3.7079231893379596, + "tokens_seen": 917504000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003646439317953862, + "loss": 1.3725, + "theoretical_loss": 3.707897779340881, + "tokens_seen": 917569536 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036463390170511536, + "loss": 1.4792, + "theoretical_loss": 3.7078723716667303, + "tokens_seen": 917635072 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036462387161484454, + "loss": 1.3786, + "theoretical_loss": 3.7078469663151292, + "tokens_seen": 917700608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003646138415245737, + "loss": 1.4084, + "theoretical_loss": 3.707821563285699, + "tokens_seen": 917766144 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036460381143430296, + "loss": 1.5067, + "theoretical_loss": 3.707796162578062, + "tokens_seen": 917831680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645937813440321, + "loss": 1.4471, + "theoretical_loss": 3.7077707641918405, + "tokens_seen": 917897216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645837512537613, + "loss": 1.5008, + "theoretical_loss": 3.7077453681266563, + "tokens_seen": 917962752 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036457372116349045, + "loss": 1.3783, + "theoretical_loss": 3.7077199743821314, + "tokens_seen": 918028288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645636910732197, + "loss": 1.4746, + "theoretical_loss": 3.7076945829578882, + "tokens_seen": 918093824 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036455366098294887, + "loss": 1.4829, + "theoretical_loss": 3.7076691938535493, + "tokens_seen": 918159360 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036454363089267805, + "loss": 1.3989, + "theoretical_loss": 3.707643807068737, + "tokens_seen": 918224896 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036453360080240723, + "loss": 1.4484, + "theoretical_loss": 3.7076184226030744, + "tokens_seen": 918290432 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645235707121364, + "loss": 1.444, + "theoretical_loss": 3.707593040456184, + "tokens_seen": 918355968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645135406218656, + "loss": 1.4738, + "theoretical_loss": 3.7075676606276877, + "tokens_seen": 918421504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003645035105315948, + "loss": 1.4633, + "theoretical_loss": 3.707542283117209, + "tokens_seen": 918487040 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036449348044132395, + "loss": 1.4082, + "theoretical_loss": 3.7075169079243704, + "tokens_seen": 918552576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003644834503510532, + "loss": 1.5308, + "theoretical_loss": 3.7074915350487956, + "tokens_seen": 918618112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003644734202607823, + "loss": 1.4281, + "theoretical_loss": 3.707466164490107, + "tokens_seen": 918683648 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036446339017051155, + "loss": 1.5725, + "theoretical_loss": 3.707440796247928, + "tokens_seen": 918749184 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036445336008024073, + "loss": 1.4203, + "theoretical_loss": 3.7074154303218823, + "tokens_seen": 918814720 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003644433299899699, + "loss": 1.5804, + "theoretical_loss": 3.707390066711593, + "tokens_seen": 918880256 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003644332998996991, + "loss": 1.475, + "theoretical_loss": 3.7073647054166834, + "tokens_seen": 918945792 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036442326980942833, + "loss": 1.3397, + "theoretical_loss": 3.7073393464367768, + "tokens_seen": 919011328 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036441323971915746, + "loss": 1.5553, + "theoretical_loss": 3.707313989771497, + "tokens_seen": 919076864 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003644032096288867, + "loss": 1.5353, + "theoretical_loss": 3.707288635420468, + "tokens_seen": 919142400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643931795386158, + "loss": 1.4019, + "theoretical_loss": 3.7072632833833135, + "tokens_seen": 919207936 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036438314944834505, + "loss": 1.4363, + "theoretical_loss": 3.707237933659657, + "tokens_seen": 919273472 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036437311935807424, + "loss": 1.4069, + "theoretical_loss": 3.7072125862491228, + "tokens_seen": 919339008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643630892678034, + "loss": 1.4108, + "theoretical_loss": 3.7071872411513347, + "tokens_seen": 919404544 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643530591775326, + "loss": 1.5028, + "theoretical_loss": 3.707161898365918, + "tokens_seen": 919470080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643430290872618, + "loss": 1.4687, + "theoretical_loss": 3.707136557892495, + "tokens_seen": 919535616 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036433299899699096, + "loss": 1.473, + "theoretical_loss": 3.7071112197306917, + "tokens_seen": 919601152 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643229689067202, + "loss": 1.3441, + "theoretical_loss": 3.7070858838801315, + "tokens_seen": 919666688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003643129388164493, + "loss": 1.4917, + "theoretical_loss": 3.7070605503404392, + "tokens_seen": 919732224 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036430290872617856, + "loss": 1.5189, + "theoretical_loss": 3.7070352191112397, + "tokens_seen": 919797760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642928786359077, + "loss": 1.4443, + "theoretical_loss": 3.707009890192157, + "tokens_seen": 919863296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642828485456369, + "loss": 1.5564, + "theoretical_loss": 3.7069845635828167, + "tokens_seen": 919928832 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642728184553661, + "loss": 1.4128, + "theoretical_loss": 3.7069592392828428, + "tokens_seen": 919994368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642627883650953, + "loss": 1.3768, + "theoretical_loss": 3.706933917291861, + "tokens_seen": 920059904 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036425275827482446, + "loss": 1.3031, + "theoretical_loss": 3.7069085976094955, + "tokens_seen": 920125440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642427281845537, + "loss": 1.3364, + "theoretical_loss": 3.7068832802353726, + "tokens_seen": 920190976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642326980942828, + "loss": 1.4743, + "theoretical_loss": 3.7068579651691165, + "tokens_seen": 920256512 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036422266800401206, + "loss": 1.3866, + "theoretical_loss": 3.7068326524103528, + "tokens_seen": 920322048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642126379137412, + "loss": 1.4688, + "theoretical_loss": 3.706807341958706, + "tokens_seen": 920387584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003642026078234704, + "loss": 1.5031, + "theoretical_loss": 3.706782033813804, + "tokens_seen": 920453120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003641925777331996, + "loss": 1.4346, + "theoretical_loss": 3.70675672797527, + "tokens_seen": 920518656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003641825476429288, + "loss": 1.4292, + "theoretical_loss": 3.7067314244427303, + "tokens_seen": 920584192 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036417251755265797, + "loss": 1.4828, + "theoretical_loss": 3.7067061232158105, + "tokens_seen": 920649728 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036416248746238715, + "loss": 1.6371, + "theoretical_loss": 3.7066808242941374, + "tokens_seen": 920715264 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 339576, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4179141521453857, + "objective/train/theoretical_loss": 3.706655527677335, + "objective/train/tokens_used": 941240800, + "theoretical_loss": 3.706655527677335, + "tokens_seen": 920780800 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036415245737211633, + "loss": 1.5519, + "theoretical_loss": 3.706655527677335, + "tokens_seen": 920780800 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036414242728184556, + "loss": 1.5868, + "theoretical_loss": 3.706630233365032, + "tokens_seen": 920846336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003641323971915747, + "loss": 1.5015, + "theoretical_loss": 3.7066049413568516, + "tokens_seen": 920911872 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003641223671013039, + "loss": 1.5725, + "theoretical_loss": 3.7065796516524214, + "tokens_seen": 920977408 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036411233701103316, + "loss": 1.4684, + "theoretical_loss": 3.7065543642513683, + "tokens_seen": 921042944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003641023069207623, + "loss": 1.4039, + "theoretical_loss": 3.7065290791533174, + "tokens_seen": 921108480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003640922768304915, + "loss": 1.463, + "theoretical_loss": 3.7065037963578957, + "tokens_seen": 921174016 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036408224674022065, + "loss": 1.3413, + "theoretical_loss": 3.7064785158647293, + "tokens_seen": 921239552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003640722166499499, + "loss": 1.3382, + "theoretical_loss": 3.7064532376734456, + "tokens_seen": 921305088 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036406218655967907, + "loss": 1.4725, + "theoretical_loss": 3.70642796178367, + "tokens_seen": 921370624 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036405215646940825, + "loss": 1.4221, + "theoretical_loss": 3.706402688195031, + "tokens_seen": 921436160 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036404212637913743, + "loss": 1.3933, + "theoretical_loss": 3.706377416907154, + "tokens_seen": 921501696 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003640320962888666, + "loss": 1.4917, + "theoretical_loss": 3.7063521479196666, + "tokens_seen": 921567232 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003640220661985958, + "loss": 1.5067, + "theoretical_loss": 3.7063268812321954, + "tokens_seen": 921632768 + }, + { + "epoch": 0.28, + "learning_rate": 0.000364012036108325, + "loss": 1.5075, + "theoretical_loss": 3.7063016168443683, + "tokens_seen": 921698304 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036400200601805415, + "loss": 1.4396, + "theoretical_loss": 3.706276354755812, + "tokens_seen": 921763840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003639919759277834, + "loss": 1.3173, + "theoretical_loss": 3.706251094966153, + "tokens_seen": 921829376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003639819458375125, + "loss": 1.5228, + "theoretical_loss": 3.7062258374750208, + "tokens_seen": 921894912 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036397191574724175, + "loss": 1.4449, + "theoretical_loss": 3.7062005822820403, + "tokens_seen": 921960448 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036396188565697093, + "loss": 1.459, + "theoretical_loss": 3.7061753293868414, + "tokens_seen": 922025984 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003639518555667001, + "loss": 1.5897, + "theoretical_loss": 3.70615007878905, + "tokens_seen": 922091520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003639418254764293, + "loss": 1.3208, + "theoretical_loss": 3.706124830488295, + "tokens_seen": 922157056 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036393179538615853, + "loss": 1.383, + "theoretical_loss": 3.7060995844842033, + "tokens_seen": 922222592 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036392176529588766, + "loss": 1.5914, + "theoretical_loss": 3.706074340776403, + "tokens_seen": 922288128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003639117352056169, + "loss": 1.4322, + "theoretical_loss": 3.7060490993645225, + "tokens_seen": 922353664 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363901705115346, + "loss": 1.4034, + "theoretical_loss": 3.70602386024819, + "tokens_seen": 922419200 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036389167502507525, + "loss": 1.3608, + "theoretical_loss": 3.7059986234270332, + "tokens_seen": 922484736 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036388164493480444, + "loss": 1.4374, + "theoretical_loss": 3.7059733889006807, + "tokens_seen": 922550272 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003638716148445336, + "loss": 1.3635, + "theoretical_loss": 3.70594815666876, + "tokens_seen": 922615808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003638615847542628, + "loss": 1.4218, + "theoretical_loss": 3.7059229267309006, + "tokens_seen": 922681344 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363851554663992, + "loss": 1.3956, + "theoretical_loss": 3.7058976990867305, + "tokens_seen": 922746880 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036384152457372116, + "loss": 1.395, + "theoretical_loss": 3.7058724737358784, + "tokens_seen": 922812416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003638314944834504, + "loss": 1.41, + "theoretical_loss": 3.705847250677973, + "tokens_seen": 922877952 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003638214643931795, + "loss": 1.4194, + "theoretical_loss": 3.7058220299126425, + "tokens_seen": 922943488 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036381143430290876, + "loss": 1.5362, + "theoretical_loss": 3.705796811439517, + "tokens_seen": 923009024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003638014042126379, + "loss": 1.452, + "theoretical_loss": 3.7057715952582244, + "tokens_seen": 923074560 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637913741223671, + "loss": 1.4934, + "theoretical_loss": 3.705746381368394, + "tokens_seen": 923140096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637813440320963, + "loss": 1.5428, + "theoretical_loss": 3.705721169769655, + "tokens_seen": 923205632 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637713139418255, + "loss": 1.4066, + "theoretical_loss": 3.705695960461636, + "tokens_seen": 923271168 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036376128385155466, + "loss": 1.4645, + "theoretical_loss": 3.7056707534439672, + "tokens_seen": 923336704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637512537612839, + "loss": 1.4378, + "theoretical_loss": 3.7056455487162783, + "tokens_seen": 923402240 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363741223671013, + "loss": 1.3588, + "theoretical_loss": 3.705620346278198, + "tokens_seen": 923467776 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036373119358074226, + "loss": 1.467, + "theoretical_loss": 3.7055951461293546, + "tokens_seen": 923533312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637211634904714, + "loss": 1.4107, + "theoretical_loss": 3.70556994826938, + "tokens_seen": 923598848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637111334002006, + "loss": 1.304, + "theoretical_loss": 3.705544752697903, + "tokens_seen": 923664384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003637011033099298, + "loss": 1.3723, + "theoretical_loss": 3.705519559414553, + "tokens_seen": 923729920 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363691073219659, + "loss": 1.3557, + "theoretical_loss": 3.7054943684189605, + "tokens_seen": 923795456 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036368104312938817, + "loss": 1.3813, + "theoretical_loss": 3.705469179710755, + "tokens_seen": 923860992 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036367101303911735, + "loss": 1.4845, + "theoretical_loss": 3.705443993289567, + "tokens_seen": 923926528 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036366098294884653, + "loss": 1.3394, + "theoretical_loss": 3.7054188091550264, + "tokens_seen": 923992064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 340768, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5665326118469238, + "objective/train/theoretical_loss": 3.705393627306764, + "objective/train/tokens_used": 944517600, + "theoretical_loss": 3.705393627306764, + "tokens_seen": 924057600 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036365095285857576, + "loss": 1.472, + "theoretical_loss": 3.705393627306764, + "tokens_seen": 924057600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003636409227683049, + "loss": 1.4357, + "theoretical_loss": 3.7053684477444087, + "tokens_seen": 924123136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003636308926780341, + "loss": 1.3885, + "theoretical_loss": 3.7053432704675924, + "tokens_seen": 924188672 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036362086258776325, + "loss": 1.3459, + "theoretical_loss": 3.7053180954759446, + "tokens_seen": 924254208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003636108324974925, + "loss": 1.3682, + "theoretical_loss": 3.7052929227690967, + "tokens_seen": 924319744 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036360080240722167, + "loss": 1.6173, + "theoretical_loss": 3.7052677523466784, + "tokens_seen": 924385280 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036359077231695085, + "loss": 1.3722, + "theoretical_loss": 3.7052425842083214, + "tokens_seen": 924450816 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036358074222668003, + "loss": 1.4034, + "theoretical_loss": 3.705217418353656, + "tokens_seen": 924516352 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036357071213640927, + "loss": 1.4112, + "theoretical_loss": 3.7051922547823137, + "tokens_seen": 924581888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003635606820461384, + "loss": 1.3293, + "theoretical_loss": 3.7051670934939245, + "tokens_seen": 924647424 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036355065195586763, + "loss": 1.405, + "theoretical_loss": 3.70514193448812, + "tokens_seen": 924712960 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036354062186559676, + "loss": 1.3536, + "theoretical_loss": 3.705116777764532, + "tokens_seen": 924778496 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363530591775326, + "loss": 1.4146, + "theoretical_loss": 3.7050916233227906, + "tokens_seen": 924844032 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036352056168505517, + "loss": 1.5648, + "theoretical_loss": 3.705066471162528, + "tokens_seen": 924909568 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036351053159478435, + "loss": 1.4913, + "theoretical_loss": 3.7050413212833755, + "tokens_seen": 924975104 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036350050150451353, + "loss": 1.3638, + "theoretical_loss": 3.7050161736849647, + "tokens_seen": 925040640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634904714142427, + "loss": 1.3569, + "theoretical_loss": 3.7049910283669263, + "tokens_seen": 925106176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634804413239719, + "loss": 1.4867, + "theoretical_loss": 3.7049658853288934, + "tokens_seen": 925171712 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036347041123370113, + "loss": 1.4272, + "theoretical_loss": 3.7049407445704965, + "tokens_seen": 925237248 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036346038114343026, + "loss": 1.3493, + "theoretical_loss": 3.704915606091369, + "tokens_seen": 925302784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634503510531595, + "loss": 1.4552, + "theoretical_loss": 3.704890469891141, + "tokens_seen": 925368320 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634403209628886, + "loss": 1.4064, + "theoretical_loss": 3.704865335969446, + "tokens_seen": 925433856 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036343029087261786, + "loss": 1.386, + "theoretical_loss": 3.7048402043259148, + "tokens_seen": 925499392 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036342026078234704, + "loss": 1.4285, + "theoretical_loss": 3.704815074960181, + "tokens_seen": 925564928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634102306920762, + "loss": 1.4696, + "theoretical_loss": 3.704789947871876, + "tokens_seen": 925630464 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003634002006018054, + "loss": 1.3801, + "theoretical_loss": 3.7047648230606325, + "tokens_seen": 925696000 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036339017051153464, + "loss": 1.4662, + "theoretical_loss": 3.7047397005260834, + "tokens_seen": 925761536 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036338014042126376, + "loss": 1.3844, + "theoretical_loss": 3.70471458026786, + "tokens_seen": 925827072 + }, + { + "epoch": 0.28, + "learning_rate": 0.000363370110330993, + "loss": 1.5756, + "theoretical_loss": 3.7046894622855957, + "tokens_seen": 925892608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003633600802407222, + "loss": 1.4691, + "theoretical_loss": 3.704664346578924, + "tokens_seen": 925958144 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036335005015045136, + "loss": 1.4129, + "theoretical_loss": 3.704639233147476, + "tokens_seen": 926023680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003633400200601806, + "loss": 1.4621, + "theoretical_loss": 3.704614121990886, + "tokens_seen": 926089216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003633299899699097, + "loss": 1.517, + "theoretical_loss": 3.7045890131087864, + "tokens_seen": 926154752 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036331995987963896, + "loss": 1.3943, + "theoretical_loss": 3.70456390650081, + "tokens_seen": 926220288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003633099297893681, + "loss": 1.3641, + "theoretical_loss": 3.7045388021665904, + "tokens_seen": 926285824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632998996990973, + "loss": 1.3689, + "theoretical_loss": 3.7045137001057613, + "tokens_seen": 926351360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632898696088265, + "loss": 1.4193, + "theoretical_loss": 3.704488600317955, + "tokens_seen": 926416896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632798395185557, + "loss": 1.4644, + "theoretical_loss": 3.7044635028028052, + "tokens_seen": 926482432 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036326980942828486, + "loss": 1.5068, + "theoretical_loss": 3.7044384075599455, + "tokens_seen": 926547968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632597793380141, + "loss": 1.4275, + "theoretical_loss": 3.7044133145890097, + "tokens_seen": 926613504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632497492477432, + "loss": 1.3047, + "theoretical_loss": 3.7043882238896306, + "tokens_seen": 926679040 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036323971915747246, + "loss": 1.3346, + "theoretical_loss": 3.704363135461443, + "tokens_seen": 926744576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632296890672016, + "loss": 1.3924, + "theoretical_loss": 3.7043380493040807, + "tokens_seen": 926810112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003632196589769308, + "loss": 1.4474, + "theoretical_loss": 3.7043129654171767, + "tokens_seen": 926875648 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036320962888666, + "loss": 1.3542, + "theoretical_loss": 3.704287883800365, + "tokens_seen": 926941184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003631995987963892, + "loss": 1.4829, + "theoretical_loss": 3.7042628044532813, + "tokens_seen": 927006720 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036318956870611837, + "loss": 1.3581, + "theoretical_loss": 3.704237727375558, + "tokens_seen": 927072256 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036317953861584755, + "loss": 1.3872, + "theoretical_loss": 3.70421265256683, + "tokens_seen": 927137792 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036316950852557673, + "loss": 1.5073, + "theoretical_loss": 3.7041875800267317, + "tokens_seen": 927203328 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036315947843530596, + "loss": 1.4087, + "theoretical_loss": 3.704162509754898, + "tokens_seen": 927268864 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 341815, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3433459997177124, + "objective/train/theoretical_loss": 3.7041374417509623, + "objective/train/tokens_used": 947794400, + "theoretical_loss": 3.7041374417509623, + "tokens_seen": 927334400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003631494483450351, + "loss": 1.3743, + "theoretical_loss": 3.7041374417509623, + "tokens_seen": 927334400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003631394182547643, + "loss": 1.3955, + "theoretical_loss": 3.7041123760145593, + "tokens_seen": 927399936 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036312938816449345, + "loss": 1.4603, + "theoretical_loss": 3.704087312545325, + "tokens_seen": 927465472 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003631193580742227, + "loss": 1.4934, + "theoretical_loss": 3.7040622513428927, + "tokens_seen": 927531008 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036310932798395187, + "loss": 1.2873, + "theoretical_loss": 3.704037192406898, + "tokens_seen": 927596544 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036309929789368105, + "loss": 1.4028, + "theoretical_loss": 3.704012135736975, + "tokens_seen": 927662080 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036308926780341023, + "loss": 1.4414, + "theoretical_loss": 3.7039870813327602, + "tokens_seen": 927727616 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036307923771313947, + "loss": 1.3244, + "theoretical_loss": 3.703962029193888, + "tokens_seen": 927793152 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003630692076228686, + "loss": 1.4552, + "theoretical_loss": 3.7039369793199928, + "tokens_seen": 927858688 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036305917753259783, + "loss": 1.3448, + "theoretical_loss": 3.7039119317107105, + "tokens_seen": 927924224 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036304914744232696, + "loss": 1.3284, + "theoretical_loss": 3.7038868863656766, + "tokens_seen": 927989760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003630391173520562, + "loss": 1.4917, + "theoretical_loss": 3.7038618432845265, + "tokens_seen": 928055296 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036302908726178537, + "loss": 1.3992, + "theoretical_loss": 3.7038368024668955, + "tokens_seen": 928120832 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036301905717151455, + "loss": 1.5022, + "theoretical_loss": 3.703811763912419, + "tokens_seen": 928186368 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036300902708124374, + "loss": 1.399, + "theoretical_loss": 3.7037867276207335, + "tokens_seen": 928251904 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629989969909729, + "loss": 1.5576, + "theoretical_loss": 3.703761693591474, + "tokens_seen": 928317440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629889669007021, + "loss": 1.4677, + "theoretical_loss": 3.7037366618242764, + "tokens_seen": 928382976 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036297893681043133, + "loss": 1.4281, + "theoretical_loss": 3.703711632318777, + "tokens_seen": 928448512 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036296890672016046, + "loss": 1.4604, + "theoretical_loss": 3.703686605074612, + "tokens_seen": 928514048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629588766298897, + "loss": 1.4019, + "theoretical_loss": 3.7036615800914165, + "tokens_seen": 928579584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629488465396188, + "loss": 1.4178, + "theoretical_loss": 3.703636557368828, + "tokens_seen": 928645120 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036293881644934806, + "loss": 1.5434, + "theoretical_loss": 3.7036115369064815, + "tokens_seen": 928710656 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036292878635907724, + "loss": 1.5095, + "theoretical_loss": 3.7035865187040145, + "tokens_seen": 928776192 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629187562688064, + "loss": 1.4366, + "theoretical_loss": 3.703561502761063, + "tokens_seen": 928841728 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003629087261785356, + "loss": 1.4261, + "theoretical_loss": 3.7035364890772637, + "tokens_seen": 928907264 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036289869608826484, + "loss": 1.4966, + "theoretical_loss": 3.7035114776522526, + "tokens_seen": 928972800 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036288866599799396, + "loss": 1.3189, + "theoretical_loss": 3.703486468485667, + "tokens_seen": 929038336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003628786359077232, + "loss": 1.5882, + "theoretical_loss": 3.7034614615771435, + "tokens_seen": 929103872 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003628686058174523, + "loss": 1.5804, + "theoretical_loss": 3.7034364569263185, + "tokens_seen": 929169408 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036285857572718156, + "loss": 1.5193, + "theoretical_loss": 3.70341145453283, + "tokens_seen": 929234944 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036284854563691074, + "loss": 1.5165, + "theoretical_loss": 3.703386454396314, + "tokens_seen": 929300480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003628385155466399, + "loss": 1.4672, + "theoretical_loss": 3.703361456516408, + "tokens_seen": 929366016 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003628284854563691, + "loss": 1.4339, + "theoretical_loss": 3.7033364608927495, + "tokens_seen": 929431552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003628184553660983, + "loss": 1.382, + "theoretical_loss": 3.7033114675249754, + "tokens_seen": 929497088 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036280842527582747, + "loss": 1.4072, + "theoretical_loss": 3.703286476412724, + "tokens_seen": 929562624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003627983951855567, + "loss": 1.3187, + "theoretical_loss": 3.703261487555631, + "tokens_seen": 929628160 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036278836509528583, + "loss": 1.4047, + "theoretical_loss": 3.703236500953335, + "tokens_seen": 929693696 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036277833500501506, + "loss": 1.4454, + "theoretical_loss": 3.7032115166054735, + "tokens_seen": 929759232 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036276830491474424, + "loss": 1.4893, + "theoretical_loss": 3.7031865345116843, + "tokens_seen": 929824768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003627582748244734, + "loss": 1.303, + "theoretical_loss": 3.7031615546716052, + "tokens_seen": 929890304 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003627482447342026, + "loss": 1.4377, + "theoretical_loss": 3.7031365770848743, + "tokens_seen": 929955840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003627382146439318, + "loss": 1.412, + "theoretical_loss": 3.7031116017511287, + "tokens_seen": 930021376 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036272818455366097, + "loss": 1.3557, + "theoretical_loss": 3.7030866286700075, + "tokens_seen": 930086912 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003627181544633902, + "loss": 1.3532, + "theoretical_loss": 3.703061657841148, + "tokens_seen": 930152448 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036270812437311933, + "loss": 1.4122, + "theoretical_loss": 3.703036689264189, + "tokens_seen": 930217984 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036269809428284857, + "loss": 1.4567, + "theoretical_loss": 3.703011722938768, + "tokens_seen": 930283520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003626880641925777, + "loss": 1.5699, + "theoretical_loss": 3.702986758864524, + "tokens_seen": 930349056 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036267803410230693, + "loss": 1.4646, + "theoretical_loss": 3.7029617970410955, + "tokens_seen": 930414592 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003626680040120361, + "loss": 1.433, + "theoretical_loss": 3.702936837468121, + "tokens_seen": 930480128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003626579739217653, + "loss": 1.3191, + "theoretical_loss": 3.7029118801452388, + "tokens_seen": 930545664 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 343003, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.54551100730896, + "objective/train/theoretical_loss": 3.7028869250720877, + "objective/train/tokens_used": 951071200, + "theoretical_loss": 3.7028869250720877, + "tokens_seen": 930611200 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036264794383149447, + "loss": 1.5461, + "theoretical_loss": 3.7028869250720877, + "tokens_seen": 930611200 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036263791374122365, + "loss": 1.3517, + "theoretical_loss": 3.702861972248307, + "tokens_seen": 930676736 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036262788365095283, + "loss": 1.4145, + "theoretical_loss": 3.702837021673535, + "tokens_seen": 930742272 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036261785356068207, + "loss": 1.4379, + "theoretical_loss": 3.7028120733474106, + "tokens_seen": 930807808 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036260782347041125, + "loss": 1.4284, + "theoretical_loss": 3.7027871272695734, + "tokens_seen": 930873344 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036259779338014043, + "loss": 1.4569, + "theoretical_loss": 3.7027621834396625, + "tokens_seen": 930938880 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036258776328986967, + "loss": 1.503, + "theoretical_loss": 3.7027372418573163, + "tokens_seen": 931004416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003625777331995988, + "loss": 1.5669, + "theoretical_loss": 3.7027123025221744, + "tokens_seen": 931069952 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036256770310932803, + "loss": 1.3513, + "theoretical_loss": 3.702687365433877, + "tokens_seen": 931135488 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036255767301905716, + "loss": 1.4509, + "theoretical_loss": 3.7026624305920626, + "tokens_seen": 931201024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003625476429287864, + "loss": 1.4503, + "theoretical_loss": 3.702637497996371, + "tokens_seen": 931266560 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003625376128385156, + "loss": 1.4384, + "theoretical_loss": 3.702612567646442, + "tokens_seen": 931332096 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036252758274824475, + "loss": 1.5397, + "theoretical_loss": 3.7025876395419153, + "tokens_seen": 931397632 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036251755265797394, + "loss": 1.5978, + "theoretical_loss": 3.702562713682431, + "tokens_seen": 931463168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003625075225677031, + "loss": 1.382, + "theoretical_loss": 3.7025377900676277, + "tokens_seen": 931528704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003624974924774323, + "loss": 1.388, + "theoretical_loss": 3.7025128686971467, + "tokens_seen": 931594240 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036248746238716153, + "loss": 1.3867, + "theoretical_loss": 3.7024879495706275, + "tokens_seen": 931659776 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036247743229689066, + "loss": 1.4395, + "theoretical_loss": 3.70246303268771, + "tokens_seen": 931725312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003624674022066199, + "loss": 1.4243, + "theoretical_loss": 3.7024381180480352, + "tokens_seen": 931790848 + }, + { + "epoch": 0.28, + "learning_rate": 0.000362457372116349, + "loss": 1.4179, + "theoretical_loss": 3.702413205651243, + "tokens_seen": 931856384 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036244734202607826, + "loss": 1.4249, + "theoretical_loss": 3.7023882954969736, + "tokens_seen": 931921920 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036243731193580744, + "loss": 1.3824, + "theoretical_loss": 3.7023633875848674, + "tokens_seen": 931987456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003624272818455366, + "loss": 1.53, + "theoretical_loss": 3.7023384819145644, + "tokens_seen": 932052992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003624172517552658, + "loss": 1.4668, + "theoretical_loss": 3.7023135784857066, + "tokens_seen": 932118528 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036240722166499504, + "loss": 1.4867, + "theoretical_loss": 3.7022886772979344, + "tokens_seen": 932184064 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036239719157472416, + "loss": 1.3893, + "theoretical_loss": 3.702263778350887, + "tokens_seen": 932249600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623871614844534, + "loss": 1.3552, + "theoretical_loss": 3.702238881644208, + "tokens_seen": 932315136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623771313941825, + "loss": 1.4408, + "theoretical_loss": 3.7022139871775357, + "tokens_seen": 932380672 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036236710130391176, + "loss": 1.4395, + "theoretical_loss": 3.702189094950512, + "tokens_seen": 932446208 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036235707121364094, + "loss": 1.395, + "theoretical_loss": 3.702164204962779, + "tokens_seen": 932511744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623470411233701, + "loss": 1.4117, + "theoretical_loss": 3.7021393172139767, + "tokens_seen": 932577280 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623370110330993, + "loss": 1.4825, + "theoretical_loss": 3.7021144317037473, + "tokens_seen": 932642816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623269809428285, + "loss": 1.4474, + "theoretical_loss": 3.702089548431731, + "tokens_seen": 932708352 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036231695085255767, + "loss": 1.4224, + "theoretical_loss": 3.7020646673975706, + "tokens_seen": 932773888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003623069207622869, + "loss": 1.4252, + "theoretical_loss": 3.7020397886009064, + "tokens_seen": 932839424 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036229689067201603, + "loss": 1.3246, + "theoretical_loss": 3.7020149120413812, + "tokens_seen": 932904960 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036228686058174526, + "loss": 1.3801, + "theoretical_loss": 3.7019900377186357, + "tokens_seen": 932970496 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036227683049147444, + "loss": 1.5542, + "theoretical_loss": 3.7019651656323127, + "tokens_seen": 933036032 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003622668004012036, + "loss": 1.6151, + "theoretical_loss": 3.7019402957820526, + "tokens_seen": 933101568 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003622567703109328, + "loss": 1.4107, + "theoretical_loss": 3.7019154281674984, + "tokens_seen": 933167104 + }, + { + "epoch": 0.28, + "learning_rate": 0.000362246740220662, + "loss": 1.5861, + "theoretical_loss": 3.701890562788292, + "tokens_seen": 933232640 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036223671013039117, + "loss": 1.3027, + "theoretical_loss": 3.7018656996440757, + "tokens_seen": 933298176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003622266800401204, + "loss": 1.4222, + "theoretical_loss": 3.701840838734491, + "tokens_seen": 933363712 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036221664994984953, + "loss": 1.4199, + "theoretical_loss": 3.701815980059181, + "tokens_seen": 933429248 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036220661985957877, + "loss": 1.3504, + "theoretical_loss": 3.701791123617787, + "tokens_seen": 933494784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003621965897693079, + "loss": 1.4539, + "theoretical_loss": 3.7017662694099527, + "tokens_seen": 933560320 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036218655967903713, + "loss": 1.4018, + "theoretical_loss": 3.70174141743532, + "tokens_seen": 933625856 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003621765295887663, + "loss": 1.4044, + "theoretical_loss": 3.7017165676935315, + "tokens_seen": 933691392 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003621664994984955, + "loss": 1.3571, + "theoretical_loss": 3.701691720184229, + "tokens_seen": 933756928 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036215646940822467, + "loss": 1.4715, + "theoretical_loss": 3.7016668749070574, + "tokens_seen": 933822464 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 344002, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5621144771575928, + "objective/train/theoretical_loss": 3.7016420318616574, + "objective/train/tokens_used": 954348000, + "theoretical_loss": 3.7016420318616574, + "tokens_seen": 933888000 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036214643931795385, + "loss": 1.5394, + "theoretical_loss": 3.7016420318616574, + "tokens_seen": 933888000 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036213640922768303, + "loss": 1.455, + "theoretical_loss": 3.7016171910476734, + "tokens_seen": 933953536 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036212637913741227, + "loss": 1.483, + "theoretical_loss": 3.701592352464748, + "tokens_seen": 934019072 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003621163490471414, + "loss": 1.4648, + "theoretical_loss": 3.701567516112523, + "tokens_seen": 934084608 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036210631895687063, + "loss": 1.4939, + "theoretical_loss": 3.701542681990644, + "tokens_seen": 934150144 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003620962888665998, + "loss": 1.5304, + "theoretical_loss": 3.7015178500987522, + "tokens_seen": 934215680 + }, + { + "epoch": 0.28, + "learning_rate": 0.000362086258776329, + "loss": 1.4388, + "theoretical_loss": 3.701493020436492, + "tokens_seen": 934281216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003620762286860582, + "loss": 1.5349, + "theoretical_loss": 3.7014681930035067, + "tokens_seen": 934346752 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036206619859578736, + "loss": 1.439, + "theoretical_loss": 3.7014433677994396, + "tokens_seen": 934412288 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036205616850551654, + "loss": 1.4283, + "theoretical_loss": 3.7014185448239343, + "tokens_seen": 934477824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003620461384152458, + "loss": 1.5103, + "theoretical_loss": 3.7013937240766346, + "tokens_seen": 934543360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003620361083249749, + "loss": 1.4783, + "theoretical_loss": 3.7013689055571843, + "tokens_seen": 934608896 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036202607823470414, + "loss": 1.4735, + "theoretical_loss": 3.701344089265227, + "tokens_seen": 934674432 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036201604814443326, + "loss": 1.3376, + "theoretical_loss": 3.7013192752004063, + "tokens_seen": 934739968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003620060180541625, + "loss": 1.4582, + "theoretical_loss": 3.701294463362367, + "tokens_seen": 934805504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003619959879638917, + "loss": 1.4256, + "theoretical_loss": 3.701269653750753, + "tokens_seen": 934871040 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036198595787362086, + "loss": 1.4852, + "theoretical_loss": 3.701244846365208, + "tokens_seen": 934936576 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036197592778335004, + "loss": 1.4783, + "theoretical_loss": 3.701220041205377, + "tokens_seen": 935002112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003619658976930792, + "loss": 1.4517, + "theoretical_loss": 3.701195238270903, + "tokens_seen": 935067648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003619558676028084, + "loss": 1.3472, + "theoretical_loss": 3.701170437561432, + "tokens_seen": 935133184 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036194583751253764, + "loss": 1.4644, + "theoretical_loss": 3.701145639076607, + "tokens_seen": 935198720 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036193580742226677, + "loss": 1.3336, + "theoretical_loss": 3.7011208428160742, + "tokens_seen": 935264256 + }, + { + "epoch": 0.28, + "learning_rate": 0.000361925777331996, + "loss": 1.4021, + "theoretical_loss": 3.701096048779477, + "tokens_seen": 935329792 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003619157472417252, + "loss": 1.4442, + "theoretical_loss": 3.7010712569664603, + "tokens_seen": 935395328 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036190571715145436, + "loss": 1.3836, + "theoretical_loss": 3.701046467376669, + "tokens_seen": 935460864 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036189568706118354, + "loss": 1.593, + "theoretical_loss": 3.701021680009749, + "tokens_seen": 935526400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618856569709127, + "loss": 1.4086, + "theoretical_loss": 3.7009968948653436, + "tokens_seen": 935591936 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618756268806419, + "loss": 1.3818, + "theoretical_loss": 3.7009721119430985, + "tokens_seen": 935657472 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036186559679037114, + "loss": 1.4053, + "theoretical_loss": 3.7009473312426593, + "tokens_seen": 935723008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618555667001003, + "loss": 1.4378, + "theoretical_loss": 3.700922552763671, + "tokens_seen": 935788544 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618455366098295, + "loss": 1.4013, + "theoretical_loss": 3.700897776505779, + "tokens_seen": 935854080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618355065195587, + "loss": 1.4237, + "theoretical_loss": 3.7008730024686285, + "tokens_seen": 935919616 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036182547642928787, + "loss": 1.4754, + "theoretical_loss": 3.7008482306518644, + "tokens_seen": 935985152 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003618154463390171, + "loss": 1.4259, + "theoretical_loss": 3.7008234610551334, + "tokens_seen": 936050688 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036180541624874623, + "loss": 1.5204, + "theoretical_loss": 3.70079869367808, + "tokens_seen": 936116224 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036179538615847546, + "loss": 1.4653, + "theoretical_loss": 3.700773928520351, + "tokens_seen": 936181760 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036178535606820464, + "loss": 1.3794, + "theoretical_loss": 3.7007491655815916, + "tokens_seen": 936247296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003617753259779338, + "loss": 1.3767, + "theoretical_loss": 3.7007244048614476, + "tokens_seen": 936312832 + }, + { + "epoch": 0.28, + "learning_rate": 0.000361765295887663, + "loss": 1.4676, + "theoretical_loss": 3.700699646359565, + "tokens_seen": 936378368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003617552657973922, + "loss": 1.4231, + "theoretical_loss": 3.7006748900755904, + "tokens_seen": 936443904 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036174523570712137, + "loss": 1.4158, + "theoretical_loss": 3.700650136009169, + "tokens_seen": 936509440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003617352056168506, + "loss": 1.4393, + "theoretical_loss": 3.700625384159948, + "tokens_seen": 936574976 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036172517552657973, + "loss": 1.4849, + "theoretical_loss": 3.7006006345275724, + "tokens_seen": 936640512 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036171514543630897, + "loss": 1.4198, + "theoretical_loss": 3.70057588711169, + "tokens_seen": 936706048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003617051153460381, + "loss": 1.464, + "theoretical_loss": 3.700551141911946, + "tokens_seen": 936771584 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036169508525576733, + "loss": 1.4279, + "theoretical_loss": 3.700526398927988, + "tokens_seen": 936837120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003616850551654965, + "loss": 1.3801, + "theoretical_loss": 3.700501658159462, + "tokens_seen": 936902656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003616750250752257, + "loss": 1.3663, + "theoretical_loss": 3.7004769196060145, + "tokens_seen": 936968192 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036166499498495487, + "loss": 1.3772, + "theoretical_loss": 3.7004521832672923, + "tokens_seen": 937033728 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036165496489468405, + "loss": 1.378, + "theoretical_loss": 3.700427449142943, + "tokens_seen": 937099264 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 344763, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.387723445892334, + "objective/train/theoretical_loss": 3.700402717232613, + "objective/train/tokens_used": 957624800, + "theoretical_loss": 3.700402717232613, + "tokens_seen": 937164800 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036164493480441323, + "loss": 1.4025, + "theoretical_loss": 3.700402717232613, + "tokens_seen": 937164800 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036163490471414247, + "loss": 1.2827, + "theoretical_loss": 3.700377987535949, + "tokens_seen": 937230336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003616248746238716, + "loss": 1.4982, + "theoretical_loss": 3.700353260052599, + "tokens_seen": 937295872 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036161484453360083, + "loss": 1.4682, + "theoretical_loss": 3.7003285347822086, + "tokens_seen": 937361408 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036160481444333, + "loss": 1.4, + "theoretical_loss": 3.7003038117244267, + "tokens_seen": 937426944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003615947843530592, + "loss": 1.4757, + "theoretical_loss": 3.7002790908789, + "tokens_seen": 937492480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003615847542627884, + "loss": 1.3553, + "theoretical_loss": 3.7002543722452756, + "tokens_seen": 937558016 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036157472417251756, + "loss": 1.4013, + "theoretical_loss": 3.7002296558232013, + "tokens_seen": 937623552 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036156469408224674, + "loss": 1.4596, + "theoretical_loss": 3.7002049416123253, + "tokens_seen": 937689088 + }, + { + "epoch": 0.28, + "learning_rate": 0.000361554663991976, + "loss": 1.5031, + "theoretical_loss": 3.700180229612294, + "tokens_seen": 937754624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003615446339017051, + "loss": 1.465, + "theoretical_loss": 3.700155519822756, + "tokens_seen": 937820160 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036153460381143434, + "loss": 1.4031, + "theoretical_loss": 3.700130812243359, + "tokens_seen": 937885696 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036152457372116346, + "loss": 1.4713, + "theoretical_loss": 3.7001061068737506, + "tokens_seen": 937951232 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003615145436308927, + "loss": 1.4816, + "theoretical_loss": 3.700081403713579, + "tokens_seen": 938016768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003615045135406219, + "loss": 1.4055, + "theoretical_loss": 3.700056702762492, + "tokens_seen": 938082304 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036149448345035106, + "loss": 1.409, + "theoretical_loss": 3.700032004020138, + "tokens_seen": 938147840 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036148445336008024, + "loss": 1.5329, + "theoretical_loss": 3.7000073074861657, + "tokens_seen": 938213376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003614744232698094, + "loss": 1.5052, + "theoretical_loss": 3.699982613160223, + "tokens_seen": 938278912 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003614643931795386, + "loss": 1.4298, + "theoretical_loss": 3.6999579210419578, + "tokens_seen": 938344448 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036145436308926784, + "loss": 1.4956, + "theoretical_loss": 3.699933231131019, + "tokens_seen": 938409984 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036144433299899697, + "loss": 1.5248, + "theoretical_loss": 3.6999085434270547, + "tokens_seen": 938475520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003614343029087262, + "loss": 1.3398, + "theoretical_loss": 3.6998838579297146, + "tokens_seen": 938541056 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003614242728184554, + "loss": 1.3599, + "theoretical_loss": 3.699859174638646, + "tokens_seen": 938606592 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036141424272818456, + "loss": 1.3842, + "theoretical_loss": 3.6998344935534995, + "tokens_seen": 938672128 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036140421263791374, + "loss": 1.4279, + "theoretical_loss": 3.699809814673922, + "tokens_seen": 938737664 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003613941825476429, + "loss": 1.4392, + "theoretical_loss": 3.699785137999563, + "tokens_seen": 938803200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003613841524573721, + "loss": 1.4489, + "theoretical_loss": 3.6997604635300725, + "tokens_seen": 938868736 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036137412236710134, + "loss": 1.4641, + "theoretical_loss": 3.699735791265099, + "tokens_seen": 938934272 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036136409227683047, + "loss": 1.3576, + "theoretical_loss": 3.6997111212042917, + "tokens_seen": 938999808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003613540621865597, + "loss": 1.5036, + "theoretical_loss": 3.6996864533473, + "tokens_seen": 939065344 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036134403209628883, + "loss": 1.5215, + "theoretical_loss": 3.6996617876937723, + "tokens_seen": 939130880 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036133400200601807, + "loss": 1.4674, + "theoretical_loss": 3.6996371242433597, + "tokens_seen": 939196416 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036132397191574725, + "loss": 1.4425, + "theoretical_loss": 3.69961246299571, + "tokens_seen": 939261952 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036131394182547643, + "loss": 1.417, + "theoretical_loss": 3.699587803950474, + "tokens_seen": 939327488 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003613039117352056, + "loss": 1.449, + "theoretical_loss": 3.699563147107301, + "tokens_seen": 939393024 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036129388164493484, + "loss": 1.2858, + "theoretical_loss": 3.6995384924658405, + "tokens_seen": 939458560 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036128385155466397, + "loss": 1.3059, + "theoretical_loss": 3.699513840025743, + "tokens_seen": 939524096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003612738214643932, + "loss": 1.4875, + "theoretical_loss": 3.6994891897866573, + "tokens_seen": 939589632 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036126379137412233, + "loss": 1.3408, + "theoretical_loss": 3.6994645417482346, + "tokens_seen": 939655168 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036125376128385157, + "loss": 1.4651, + "theoretical_loss": 3.6994398959101242, + "tokens_seen": 939720704 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036124373119358075, + "loss": 1.4372, + "theoretical_loss": 3.6994152522719768, + "tokens_seen": 939786240 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036123370110330993, + "loss": 1.4866, + "theoretical_loss": 3.699390610833442, + "tokens_seen": 939851776 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003612236710130391, + "loss": 1.4875, + "theoretical_loss": 3.6993659715941702, + "tokens_seen": 939917312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003612136409227683, + "loss": 1.5367, + "theoretical_loss": 3.699341334553812, + "tokens_seen": 939982848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003612036108324975, + "loss": 1.3538, + "theoretical_loss": 3.6993166997120186, + "tokens_seen": 940048384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003611935807422267, + "loss": 1.4229, + "theoretical_loss": 3.6992920670684395, + "tokens_seen": 940113920 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036118355065195584, + "loss": 1.4786, + "theoretical_loss": 3.6992674366227254, + "tokens_seen": 940179456 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036117352056168507, + "loss": 1.4486, + "theoretical_loss": 3.6992428083745272, + "tokens_seen": 940244992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003611634904714142, + "loss": 1.3446, + "theoretical_loss": 3.699218182323496, + "tokens_seen": 940310528 + }, + { + "epoch": 0.28, + "learning_rate": 0.00036115346038114343, + "loss": 1.4475, + "theoretical_loss": 3.6991935584692826, + "tokens_seen": 940376064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 345855, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3819808959960938, + "objective/train/theoretical_loss": 3.6991689368115375, + "objective/train/tokens_used": 960901600, + "theoretical_loss": 3.6991689368115375, + "tokens_seen": 940441600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0003611434302908726, + "loss": 1.4241, + "theoretical_loss": 3.6991689368115375, + "tokens_seen": 940441600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003611334002006018, + "loss": 1.5502, + "theoretical_loss": 3.699144317349912, + "tokens_seen": 940507136 + }, + { + "epoch": 0.29, + "learning_rate": 0.000361123370110331, + "loss": 1.3481, + "theoretical_loss": 3.6991197000840574, + "tokens_seen": 940572672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003611133400200602, + "loss": 1.45, + "theoretical_loss": 3.6990950850136244, + "tokens_seen": 940638208 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003611033099297894, + "loss": 1.4268, + "theoretical_loss": 3.6990704721382652, + "tokens_seen": 940703744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003610932798395186, + "loss": 1.2914, + "theoretical_loss": 3.69904586145763, + "tokens_seen": 940769280 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036108324974924776, + "loss": 1.4752, + "theoretical_loss": 3.699021252971371, + "tokens_seen": 940834816 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036107321965897694, + "loss": 1.413, + "theoretical_loss": 3.698996646679139, + "tokens_seen": 940900352 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003610631895687062, + "loss": 1.4724, + "theoretical_loss": 3.6989720425805865, + "tokens_seen": 940965888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003610531594784353, + "loss": 1.4977, + "theoretical_loss": 3.6989474406753646, + "tokens_seen": 941031424 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036104312938816454, + "loss": 1.5131, + "theoretical_loss": 3.6989228409631254, + "tokens_seen": 941096960 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036103309929789366, + "loss": 1.5275, + "theoretical_loss": 3.6988982434435203, + "tokens_seen": 941162496 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003610230692076229, + "loss": 1.4723, + "theoretical_loss": 3.6988736481162015, + "tokens_seen": 941228032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003610130391173521, + "loss": 1.4309, + "theoretical_loss": 3.698849054980821, + "tokens_seen": 941293568 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036100300902708126, + "loss": 1.4245, + "theoretical_loss": 3.6988244640370302, + "tokens_seen": 941359104 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036099297893681044, + "loss": 1.4703, + "theoretical_loss": 3.6987998752844815, + "tokens_seen": 941424640 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003609829488465396, + "loss": 1.4711, + "theoretical_loss": 3.6987752887228282, + "tokens_seen": 941490176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003609729187562688, + "loss": 1.4281, + "theoretical_loss": 3.698750704351722, + "tokens_seen": 941555712 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036096288866599804, + "loss": 1.5148, + "theoretical_loss": 3.6987261221708145, + "tokens_seen": 941621248 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036095285857572717, + "loss": 1.5168, + "theoretical_loss": 3.6987015421797587, + "tokens_seen": 941686784 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003609428284854564, + "loss": 1.5841, + "theoretical_loss": 3.698676964378208, + "tokens_seen": 941752320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003609327983951856, + "loss": 1.3132, + "theoretical_loss": 3.6986523887658134, + "tokens_seen": 941817856 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036092276830491476, + "loss": 1.353, + "theoretical_loss": 3.6986278153422285, + "tokens_seen": 941883392 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036091273821464394, + "loss": 1.4694, + "theoretical_loss": 3.6986032441071064, + "tokens_seen": 941948928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003609027081243731, + "loss": 1.4807, + "theoretical_loss": 3.6985786750600993, + "tokens_seen": 942014464 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003608926780341023, + "loss": 1.3992, + "theoretical_loss": 3.6985541082008604, + "tokens_seen": 942080000 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036088264794383154, + "loss": 1.4196, + "theoretical_loss": 3.698529543529043, + "tokens_seen": 942145536 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036087261785356067, + "loss": 1.3881, + "theoretical_loss": 3.6985049810443, + "tokens_seen": 942211072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003608625877632899, + "loss": 1.4428, + "theoretical_loss": 3.6984804207462836, + "tokens_seen": 942276608 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036085255767301903, + "loss": 1.468, + "theoretical_loss": 3.698455862634648, + "tokens_seen": 942342144 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036084252758274827, + "loss": 1.4492, + "theoretical_loss": 3.6984313067090477, + "tokens_seen": 942407680 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036083249749247745, + "loss": 1.5267, + "theoretical_loss": 3.698406752969134, + "tokens_seen": 942473216 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036082246740220663, + "loss": 1.4849, + "theoretical_loss": 3.698382201414561, + "tokens_seen": 942538752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003608124373119358, + "loss": 1.421, + "theoretical_loss": 3.6983576520449835, + "tokens_seen": 942604288 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036080240722166505, + "loss": 1.3264, + "theoretical_loss": 3.6983331048600534, + "tokens_seen": 942669824 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036079237713139417, + "loss": 1.4344, + "theoretical_loss": 3.6983085598594254, + "tokens_seen": 942735360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003607823470411234, + "loss": 1.5218, + "theoretical_loss": 3.698284017042753, + "tokens_seen": 942800896 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036077231695085253, + "loss": 1.5612, + "theoretical_loss": 3.6982594764096905, + "tokens_seen": 942866432 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036076228686058177, + "loss": 1.4627, + "theoretical_loss": 3.6982349379598913, + "tokens_seen": 942931968 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036075225677031095, + "loss": 1.3242, + "theoretical_loss": 3.69821040169301, + "tokens_seen": 942997504 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036074222668004013, + "loss": 1.4852, + "theoretical_loss": 3.6981858676087005, + "tokens_seen": 943063040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003607321965897693, + "loss": 1.4517, + "theoretical_loss": 3.698161335706617, + "tokens_seen": 943128576 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003607221664994985, + "loss": 1.425, + "theoretical_loss": 3.6981368059864135, + "tokens_seen": 943194112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003607121364092277, + "loss": 1.547, + "theoretical_loss": 3.698112278447744, + "tokens_seen": 943259648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003607021063189569, + "loss": 1.5663, + "theoretical_loss": 3.6980877530902645, + "tokens_seen": 943325184 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036069207622868604, + "loss": 1.4733, + "theoretical_loss": 3.698063229913628, + "tokens_seen": 943390720 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036068204613841527, + "loss": 1.4588, + "theoretical_loss": 3.6980387089174895, + "tokens_seen": 943456256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003606720160481444, + "loss": 1.4949, + "theoretical_loss": 3.6980141901015045, + "tokens_seen": 943521792 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036066198595787364, + "loss": 1.4853, + "theoretical_loss": 3.697989673465327, + "tokens_seen": 943587328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003606519558676028, + "loss": 1.5337, + "theoretical_loss": 3.6979651590086107, + "tokens_seen": 943652864 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 346959, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3550227880477905, + "objective/train/theoretical_loss": 3.6979406467310127, + "objective/train/tokens_used": 964178400, + "theoretical_loss": 3.6979406467310127, + "tokens_seen": 943718400 + }, + { + "epoch": 0.29, + "learning_rate": 0.000360641925777332, + "loss": 1.3472, + "theoretical_loss": 3.6979406467310127, + "tokens_seen": 943718400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003606318956870612, + "loss": 1.4759, + "theoretical_loss": 3.6979161366321867, + "tokens_seen": 943783936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003606218655967904, + "loss": 1.4373, + "theoretical_loss": 3.6978916287117882, + "tokens_seen": 943849472 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036061183550651954, + "loss": 1.4486, + "theoretical_loss": 3.6978671229694724, + "tokens_seen": 943915008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003606018054162488, + "loss": 1.5165, + "theoretical_loss": 3.697842619404894, + "tokens_seen": 943980544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605917753259779, + "loss": 1.4304, + "theoretical_loss": 3.6978181180177083, + "tokens_seen": 944046080 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036058174523570714, + "loss": 1.5228, + "theoretical_loss": 3.6977936188075717, + "tokens_seen": 944111616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605717151454363, + "loss": 1.4453, + "theoretical_loss": 3.6977691217741384, + "tokens_seen": 944177152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605616850551655, + "loss": 1.4417, + "theoretical_loss": 3.697744626917065, + "tokens_seen": 944242688 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605516549648947, + "loss": 1.4132, + "theoretical_loss": 3.6977201342360067, + "tokens_seen": 944308224 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036054162487462386, + "loss": 1.4648, + "theoretical_loss": 3.6976956437306185, + "tokens_seen": 944373760 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036053159478435304, + "loss": 1.4268, + "theoretical_loss": 3.697671155400557, + "tokens_seen": 944439296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605215646940823, + "loss": 1.4678, + "theoretical_loss": 3.6976466692454784, + "tokens_seen": 944504832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003605115346038114, + "loss": 1.4588, + "theoretical_loss": 3.697622185265037, + "tokens_seen": 944570368 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036050150451354064, + "loss": 1.4624, + "theoretical_loss": 3.6975977034588907, + "tokens_seen": 944635904 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036049147442326977, + "loss": 1.3775, + "theoretical_loss": 3.6975732238266943, + "tokens_seen": 944701440 + }, + { + "epoch": 0.29, + "learning_rate": 0.000360481444332999, + "loss": 1.4726, + "theoretical_loss": 3.697548746368105, + "tokens_seen": 944766976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003604714142427282, + "loss": 1.4263, + "theoretical_loss": 3.697524271082778, + "tokens_seen": 944832512 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036046138415245737, + "loss": 1.4588, + "theoretical_loss": 3.6974997979703703, + "tokens_seen": 944898048 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036045135406218655, + "loss": 1.479, + "theoretical_loss": 3.697475327030538, + "tokens_seen": 944963584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003604413239719158, + "loss": 1.5276, + "theoretical_loss": 3.697450858262938, + "tokens_seen": 945029120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003604312938816449, + "loss": 1.3439, + "theoretical_loss": 3.697426391667226, + "tokens_seen": 945094656 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036042126379137414, + "loss": 1.4378, + "theoretical_loss": 3.6974019272430594, + "tokens_seen": 945160192 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036041123370110327, + "loss": 1.4277, + "theoretical_loss": 3.697377464990095, + "tokens_seen": 945225728 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003604012036108325, + "loss": 1.4277, + "theoretical_loss": 3.697353004907989, + "tokens_seen": 945291264 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003603911735205617, + "loss": 1.5005, + "theoretical_loss": 3.697328546996398, + "tokens_seen": 945356800 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036038114343029087, + "loss": 1.3902, + "theoretical_loss": 3.69730409125498, + "tokens_seen": 945422336 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036037111334002005, + "loss": 1.3425, + "theoretical_loss": 3.6972796376833914, + "tokens_seen": 945487872 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036036108324974923, + "loss": 1.4504, + "theoretical_loss": 3.6972551862812892, + "tokens_seen": 945553408 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036035105315947847, + "loss": 1.4075, + "theoretical_loss": 3.697230737048331, + "tokens_seen": 945618944 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036034102306920765, + "loss": 1.4053, + "theoretical_loss": 3.697206289984174, + "tokens_seen": 945684480 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036033099297893683, + "loss": 1.3319, + "theoretical_loss": 3.6971818450884752, + "tokens_seen": 945750016 + }, + { + "epoch": 0.29, + "learning_rate": 0.000360320962888666, + "loss": 1.4752, + "theoretical_loss": 3.6971574023608924, + "tokens_seen": 945815552 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036031093279839525, + "loss": 1.4053, + "theoretical_loss": 3.6971329618010826, + "tokens_seen": 945881088 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036030090270812437, + "loss": 1.3807, + "theoretical_loss": 3.697108523408704, + "tokens_seen": 945946624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003602908726178536, + "loss": 1.4068, + "theoretical_loss": 3.6970840871834136, + "tokens_seen": 946012160 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036028084252758273, + "loss": 1.3001, + "theoretical_loss": 3.697059653124869, + "tokens_seen": 946077696 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036027081243731197, + "loss": 1.5158, + "theoretical_loss": 3.697035221232729, + "tokens_seen": 946143232 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036026078234704115, + "loss": 1.512, + "theoretical_loss": 3.697010791506651, + "tokens_seen": 946208768 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036025075225677033, + "loss": 1.427, + "theoretical_loss": 3.6969863639462925, + "tokens_seen": 946274304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003602407221664995, + "loss": 1.5172, + "theoretical_loss": 3.696961938551312, + "tokens_seen": 946339840 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003602306920762287, + "loss": 1.3464, + "theoretical_loss": 3.696937515321368, + "tokens_seen": 946405376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003602206619859579, + "loss": 1.4797, + "theoretical_loss": 3.696913094256117, + "tokens_seen": 946470912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003602106318956871, + "loss": 1.4598, + "theoretical_loss": 3.6968886753552193, + "tokens_seen": 946536448 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036020060180541624, + "loss": 1.5312, + "theoretical_loss": 3.6968642586183327, + "tokens_seen": 946601984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601905717151455, + "loss": 1.3941, + "theoretical_loss": 3.6968398440451145, + "tokens_seen": 946667520 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601805416248746, + "loss": 1.4233, + "theoretical_loss": 3.696815431635225, + "tokens_seen": 946733056 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036017051153460384, + "loss": 1.506, + "theoretical_loss": 3.696791021388321, + "tokens_seen": 946798592 + }, + { + "epoch": 0.29, + "learning_rate": 0.000360160481444333, + "loss": 1.4395, + "theoretical_loss": 3.6967666133040624, + "tokens_seen": 946864128 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601504513540622, + "loss": 1.5681, + "theoretical_loss": 3.6967422073821075, + "tokens_seen": 946929664 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 348186, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4438738822937012, + "objective/train/theoretical_loss": 3.696717803622115, + "objective/train/tokens_used": 967455200, + "theoretical_loss": 3.696717803622115, + "tokens_seen": 946995200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601404212637914, + "loss": 1.4383, + "theoretical_loss": 3.696717803622115, + "tokens_seen": 946995200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601303911735206, + "loss": 1.343, + "theoretical_loss": 3.696693402023744, + "tokens_seen": 947060736 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036012036108324974, + "loss": 1.3634, + "theoretical_loss": 3.696669002586653, + "tokens_seen": 947126272 + }, + { + "epoch": 0.29, + "learning_rate": 0.000360110330992979, + "loss": 1.3904, + "theoretical_loss": 3.696644605310502, + "tokens_seen": 947191808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003601003009027081, + "loss": 1.4508, + "theoretical_loss": 3.696620210194949, + "tokens_seen": 947257344 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036009027081243734, + "loss": 1.5186, + "theoretical_loss": 3.696595817239654, + "tokens_seen": 947322880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003600802407221665, + "loss": 1.531, + "theoretical_loss": 3.6965714264442764, + "tokens_seen": 947388416 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003600702106318957, + "loss": 1.4756, + "theoretical_loss": 3.6965470378084744, + "tokens_seen": 947453952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003600601805416249, + "loss": 1.5037, + "theoretical_loss": 3.696522651331909, + "tokens_seen": 947519488 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036005015045135406, + "loss": 1.3805, + "theoretical_loss": 3.6964982670142383, + "tokens_seen": 947585024 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036004012036108324, + "loss": 1.5629, + "theoretical_loss": 3.6964738848551226, + "tokens_seen": 947650560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003600300902708125, + "loss": 1.379, + "theoretical_loss": 3.696449504854222, + "tokens_seen": 947716096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003600200601805416, + "loss": 1.4836, + "theoretical_loss": 3.696425127011195, + "tokens_seen": 947781632 + }, + { + "epoch": 0.29, + "learning_rate": 0.00036001003009027084, + "loss": 1.3762, + "theoretical_loss": 3.6964007513257022, + "tokens_seen": 947847168 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035999999999999997, + "loss": 1.5203, + "theoretical_loss": 3.696376377797404, + "tokens_seen": 947912704 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003599899699097292, + "loss": 1.3309, + "theoretical_loss": 3.696352006425959, + "tokens_seen": 947978240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003599799398194584, + "loss": 1.4679, + "theoretical_loss": 3.696327637211028, + "tokens_seen": 948043776 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035996990972918757, + "loss": 1.2921, + "theoretical_loss": 3.6963032701522716, + "tokens_seen": 948109312 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035995987963891675, + "loss": 1.4361, + "theoretical_loss": 3.6962789052493497, + "tokens_seen": 948174848 + }, + { + "epoch": 0.29, + "learning_rate": 0.000359949849548646, + "loss": 1.4009, + "theoretical_loss": 3.696254542501922, + "tokens_seen": 948240384 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003599398194583751, + "loss": 1.4055, + "theoretical_loss": 3.6962301819096495, + "tokens_seen": 948305920 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035992978936810434, + "loss": 1.5317, + "theoretical_loss": 3.696205823472192, + "tokens_seen": 948371456 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035991975927783347, + "loss": 1.4371, + "theoretical_loss": 3.696181467189211, + "tokens_seen": 948436992 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003599097291875627, + "loss": 1.4321, + "theoretical_loss": 3.6961571130603663, + "tokens_seen": 948502528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003598996990972919, + "loss": 1.6373, + "theoretical_loss": 3.6961327610853187, + "tokens_seen": 948568064 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035988966900702107, + "loss": 1.3699, + "theoretical_loss": 3.6961084112637295, + "tokens_seen": 948633600 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035987963891675025, + "loss": 1.5035, + "theoretical_loss": 3.6960840635952583, + "tokens_seen": 948699136 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035986960882647943, + "loss": 1.4429, + "theoretical_loss": 3.696059718079567, + "tokens_seen": 948764672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003598595787362086, + "loss": 1.5377, + "theoretical_loss": 3.696035374716316, + "tokens_seen": 948830208 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035984954864593785, + "loss": 1.452, + "theoretical_loss": 3.696011033505167, + "tokens_seen": 948895744 + }, + { + "epoch": 0.29, + "learning_rate": 0.000359839518555667, + "loss": 1.5171, + "theoretical_loss": 3.6959866944457813, + "tokens_seen": 948961280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003598294884653962, + "loss": 1.4549, + "theoretical_loss": 3.6959623575378187, + "tokens_seen": 949026816 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035981945837512534, + "loss": 1.4746, + "theoretical_loss": 3.6959380227809415, + "tokens_seen": 949092352 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035980942828485457, + "loss": 1.4863, + "theoretical_loss": 3.6959136901748115, + "tokens_seen": 949157888 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035979939819458375, + "loss": 1.4392, + "theoretical_loss": 3.695889359719089, + "tokens_seen": 949223424 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035978936810431293, + "loss": 1.4796, + "theoretical_loss": 3.695865031413436, + "tokens_seen": 949288960 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003597793380140421, + "loss": 1.4857, + "theoretical_loss": 3.695840705257514, + "tokens_seen": 949354496 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035976930792377135, + "loss": 1.4819, + "theoretical_loss": 3.6958163812509857, + "tokens_seen": 949420032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003597592778335005, + "loss": 1.5657, + "theoretical_loss": 3.695792059393511, + "tokens_seen": 949485568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003597492477432297, + "loss": 1.5442, + "theoretical_loss": 3.695767739684753, + "tokens_seen": 949551104 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035973921765295884, + "loss": 1.4537, + "theoretical_loss": 3.695743422124373, + "tokens_seen": 949616640 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003597291875626881, + "loss": 1.4569, + "theoretical_loss": 3.6957191067120334, + "tokens_seen": 949682176 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035971915747241726, + "loss": 1.4545, + "theoretical_loss": 3.6956947934473963, + "tokens_seen": 949747712 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035970912738214644, + "loss": 1.5682, + "theoretical_loss": 3.695670482330123, + "tokens_seen": 949813248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596990972918756, + "loss": 1.4079, + "theoretical_loss": 3.6956461733598767, + "tokens_seen": 949878784 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596890672016048, + "loss": 1.4499, + "theoretical_loss": 3.695621866536319, + "tokens_seen": 949944320 + }, + { + "epoch": 0.29, + "learning_rate": 0.000359679037111334, + "loss": 1.4184, + "theoretical_loss": 3.6955975618591133, + "tokens_seen": 950009856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596690070210632, + "loss": 1.487, + "theoretical_loss": 3.6955732593279205, + "tokens_seen": 950075392 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035965897693079234, + "loss": 1.5191, + "theoretical_loss": 3.6955489589424033, + "tokens_seen": 950140928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596489468405216, + "loss": 1.4321, + "theoretical_loss": 3.6955246607022256, + "tokens_seen": 950206464 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 349336, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3909180164337158, + "objective/train/theoretical_loss": 3.695500364607049, + "objective/train/tokens_used": 970732000, + "theoretical_loss": 3.695500364607049, + "tokens_seen": 950272000 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035963891675025076, + "loss": 1.4056, + "theoretical_loss": 3.695500364607049, + "tokens_seen": 950272000 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035962888665997994, + "loss": 1.3801, + "theoretical_loss": 3.695476070656537, + "tokens_seen": 950337536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596188565697091, + "loss": 1.5926, + "theoretical_loss": 3.6954517788503507, + "tokens_seen": 950403072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003596088264794383, + "loss": 1.3201, + "theoretical_loss": 3.695427489188155, + "tokens_seen": 950468608 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035959879638916754, + "loss": 1.3906, + "theoretical_loss": 3.6954032016696123, + "tokens_seen": 950534144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003595887662988967, + "loss": 1.4404, + "theoretical_loss": 3.6953789162943846, + "tokens_seen": 950599680 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003595787362086259, + "loss": 1.5915, + "theoretical_loss": 3.695354633062136, + "tokens_seen": 950665216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003595687061183551, + "loss": 1.3957, + "theoretical_loss": 3.6953303519725296, + "tokens_seen": 950730752 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035955867602808426, + "loss": 1.4086, + "theoretical_loss": 3.695306073025229, + "tokens_seen": 950796288 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035954864593781344, + "loss": 1.505, + "theoretical_loss": 3.6952817962198967, + "tokens_seen": 950861824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003595386158475427, + "loss": 1.4976, + "theoretical_loss": 3.6952575215561962, + "tokens_seen": 950927360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003595285857572718, + "loss": 1.3916, + "theoretical_loss": 3.695233249033792, + "tokens_seen": 950992896 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035951855566700104, + "loss": 1.4215, + "theoretical_loss": 3.6952089786523468, + "tokens_seen": 951058432 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035950852557673017, + "loss": 1.4036, + "theoretical_loss": 3.6951847104115236, + "tokens_seen": 951123968 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594984954864594, + "loss": 1.4862, + "theoretical_loss": 3.6951604443109876, + "tokens_seen": 951189504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594884653961886, + "loss": 1.4966, + "theoretical_loss": 3.695136180350402, + "tokens_seen": 951255040 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035947843530591777, + "loss": 1.4821, + "theoretical_loss": 3.69511191852943, + "tokens_seen": 951320576 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035946840521564695, + "loss": 1.5214, + "theoretical_loss": 3.695087658847737, + "tokens_seen": 951386112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594583751253762, + "loss": 1.4275, + "theoretical_loss": 3.695063401304986, + "tokens_seen": 951451648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594483450351053, + "loss": 1.4894, + "theoretical_loss": 3.695039145900841, + "tokens_seen": 951517184 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035943831494483454, + "loss": 1.3479, + "theoretical_loss": 3.6950148926349664, + "tokens_seen": 951582720 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035942828485456367, + "loss": 1.5236, + "theoretical_loss": 3.694990641507027, + "tokens_seen": 951648256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594182547642929, + "loss": 1.4666, + "theoretical_loss": 3.694966392516686, + "tokens_seen": 951713792 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003594082246740221, + "loss": 1.4248, + "theoretical_loss": 3.6949421456636085, + "tokens_seen": 951779328 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035939819458375127, + "loss": 1.4652, + "theoretical_loss": 3.694917900947459, + "tokens_seen": 951844864 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035938816449348045, + "loss": 1.4037, + "theoretical_loss": 3.6948936583679024, + "tokens_seen": 951910400 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035937813440320963, + "loss": 1.3066, + "theoretical_loss": 3.694869417924603, + "tokens_seen": 951975936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003593681043129388, + "loss": 1.3707, + "theoretical_loss": 3.694845179617225, + "tokens_seen": 952041472 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035935807422266805, + "loss": 1.5471, + "theoretical_loss": 3.6948209434454338, + "tokens_seen": 952107008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003593480441323972, + "loss": 1.3775, + "theoretical_loss": 3.6947967094088936, + "tokens_seen": 952172544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003593380140421264, + "loss": 1.5439, + "theoretical_loss": 3.6947724775072706, + "tokens_seen": 952238080 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035932798395185554, + "loss": 1.4193, + "theoretical_loss": 3.6947482477402285, + "tokens_seen": 952303616 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035931795386158477, + "loss": 1.3907, + "theoretical_loss": 3.694724020107433, + "tokens_seen": 952369152 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035930792377131395, + "loss": 1.5219, + "theoretical_loss": 3.6946997946085487, + "tokens_seen": 952434688 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035929789368104313, + "loss": 1.4926, + "theoretical_loss": 3.694675571243242, + "tokens_seen": 952500224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003592878635907723, + "loss": 1.3946, + "theoretical_loss": 3.6946513500111777, + "tokens_seen": 952565760 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035927783350050155, + "loss": 1.437, + "theoretical_loss": 3.6946271309120204, + "tokens_seen": 952631296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003592678034102307, + "loss": 1.4501, + "theoretical_loss": 3.6946029139454364, + "tokens_seen": 952696832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003592577733199599, + "loss": 1.3523, + "theoretical_loss": 3.6945786991110907, + "tokens_seen": 952762368 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035924774322968904, + "loss": 1.4409, + "theoretical_loss": 3.6945544864086495, + "tokens_seen": 952827904 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003592377131394183, + "loss": 1.4484, + "theoretical_loss": 3.6945302758377783, + "tokens_seen": 952893440 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035922768304914746, + "loss": 1.4414, + "theoretical_loss": 3.6945060673981427, + "tokens_seen": 952958976 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035921765295887664, + "loss": 1.3909, + "theoretical_loss": 3.6944818610894083, + "tokens_seen": 953024512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003592076228686058, + "loss": 1.3846, + "theoretical_loss": 3.6944576569112417, + "tokens_seen": 953090048 + }, + { + "epoch": 0.29, + "learning_rate": 0.000359197592778335, + "loss": 1.4342, + "theoretical_loss": 3.694433454863309, + "tokens_seen": 953155584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591875626880642, + "loss": 1.51, + "theoretical_loss": 3.694409254945275, + "tokens_seen": 953221120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591775325977934, + "loss": 1.5344, + "theoretical_loss": 3.694385057156807, + "tokens_seen": 953286656 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035916750250752254, + "loss": 1.4525, + "theoretical_loss": 3.6943608614975707, + "tokens_seen": 953352192 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591574724172518, + "loss": 1.5473, + "theoretical_loss": 3.6943366679672325, + "tokens_seen": 953417728 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035914744232698096, + "loss": 1.2595, + "theoretical_loss": 3.6943124765654587, + "tokens_seen": 953483264 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 350619, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3169611692428589, + "objective/train/theoretical_loss": 3.694288287291916, + "objective/train/tokens_used": 974008800, + "theoretical_loss": 3.694288287291916, + "tokens_seen": 953548800 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035913741223671014, + "loss": 1.3777, + "theoretical_loss": 3.694288287291916, + "tokens_seen": 953548800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591273821464393, + "loss": 1.39, + "theoretical_loss": 3.6942641001462713, + "tokens_seen": 953614336 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591173520561685, + "loss": 1.4515, + "theoretical_loss": 3.6942399151281897, + "tokens_seen": 953679872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003591073219658977, + "loss": 1.3755, + "theoretical_loss": 3.6942157322373395, + "tokens_seen": 953745408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003590972918756269, + "loss": 1.5936, + "theoretical_loss": 3.694191551473387, + "tokens_seen": 953810944 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035908726178535605, + "loss": 1.4753, + "theoretical_loss": 3.694167372835998, + "tokens_seen": 953876480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003590772316950853, + "loss": 1.4605, + "theoretical_loss": 3.6941431963248403, + "tokens_seen": 953942016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003590672016048144, + "loss": 1.4166, + "theoretical_loss": 3.6941190219395814, + "tokens_seen": 954007552 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035905717151454364, + "loss": 1.4631, + "theoretical_loss": 3.6940948496798875, + "tokens_seen": 954073088 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003590471414242728, + "loss": 1.415, + "theoretical_loss": 3.694070679545426, + "tokens_seen": 954138624 + }, + { + "epoch": 0.29, + "learning_rate": 0.000359037111334002, + "loss": 1.3265, + "theoretical_loss": 3.6940465115358636, + "tokens_seen": 954204160 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003590270812437312, + "loss": 1.4339, + "theoretical_loss": 3.6940223456508683, + "tokens_seen": 954269696 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035901705115346037, + "loss": 1.4992, + "theoretical_loss": 3.693998181890107, + "tokens_seen": 954335232 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035900702106318955, + "loss": 1.3957, + "theoretical_loss": 3.693974020253248, + "tokens_seen": 954400768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589969909729188, + "loss": 1.3663, + "theoretical_loss": 3.6939498607399566, + "tokens_seen": 954466304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589869608826479, + "loss": 1.3908, + "theoretical_loss": 3.693925703349903, + "tokens_seen": 954531840 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035897693079237715, + "loss": 1.4347, + "theoretical_loss": 3.6939015480827533, + "tokens_seen": 954597376 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035896690070210633, + "loss": 1.427, + "theoretical_loss": 3.6938773949381756, + "tokens_seen": 954662912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589568706118355, + "loss": 1.4702, + "theoretical_loss": 3.6938532439158376, + "tokens_seen": 954728448 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589468405215647, + "loss": 1.4016, + "theoretical_loss": 3.693829095015407, + "tokens_seen": 954793984 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035893681043129387, + "loss": 1.4686, + "theoretical_loss": 3.6938049482365525, + "tokens_seen": 954859520 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035892678034102305, + "loss": 1.465, + "theoretical_loss": 3.6937808035789415, + "tokens_seen": 954925056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589167502507523, + "loss": 1.3695, + "theoretical_loss": 3.6937566610422423, + "tokens_seen": 954990592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003589067201604814, + "loss": 1.4825, + "theoretical_loss": 3.6937325206261225, + "tokens_seen": 955056128 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035889669007021065, + "loss": 1.311, + "theoretical_loss": 3.693708382330251, + "tokens_seen": 955121664 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003588866599799398, + "loss": 1.5084, + "theoretical_loss": 3.6936842461542962, + "tokens_seen": 955187200 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358876629889669, + "loss": 1.5102, + "theoretical_loss": 3.693660112097926, + "tokens_seen": 955252736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003588665997993982, + "loss": 1.3828, + "theoretical_loss": 3.693635980160809, + "tokens_seen": 955318272 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003588565697091274, + "loss": 1.435, + "theoretical_loss": 3.693611850342614, + "tokens_seen": 955383808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003588465396188566, + "loss": 1.4925, + "theoretical_loss": 3.6935877226430094, + "tokens_seen": 955449344 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035883650952858574, + "loss": 1.4025, + "theoretical_loss": 3.6935635970616643, + "tokens_seen": 955514880 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035882647943831497, + "loss": 1.3891, + "theoretical_loss": 3.6935394735982463, + "tokens_seen": 955580416 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035881644934804415, + "loss": 1.4302, + "theoretical_loss": 3.6935153522524256, + "tokens_seen": 955645952 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035880641925777333, + "loss": 1.4439, + "theoretical_loss": 3.6934912330238703, + "tokens_seen": 955711488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003587963891675025, + "loss": 1.485, + "theoretical_loss": 3.6934671159122496, + "tokens_seen": 955777024 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035878635907723175, + "loss": 1.4861, + "theoretical_loss": 3.6934430009172328, + "tokens_seen": 955842560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003587763289869609, + "loss": 1.4692, + "theoretical_loss": 3.693418888038489, + "tokens_seen": 955908096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003587662988966901, + "loss": 1.3297, + "theoretical_loss": 3.693394777275687, + "tokens_seen": 955973632 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035875626880641924, + "loss": 1.3573, + "theoretical_loss": 3.693370668628496, + "tokens_seen": 956039168 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003587462387161485, + "loss": 1.4958, + "theoretical_loss": 3.693346562096586, + "tokens_seen": 956104704 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035873620862587766, + "loss": 1.4643, + "theoretical_loss": 3.693322457679626, + "tokens_seen": 956170240 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035872617853560684, + "loss": 1.5551, + "theoretical_loss": 3.693298355377286, + "tokens_seen": 956235776 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358716148445336, + "loss": 1.3855, + "theoretical_loss": 3.693274255189235, + "tokens_seen": 956301312 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003587061183550652, + "loss": 1.4296, + "theoretical_loss": 3.693250157115142, + "tokens_seen": 956366848 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586960882647944, + "loss": 1.4878, + "theoretical_loss": 3.6932260611546788, + "tokens_seen": 956432384 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586860581745236, + "loss": 1.4881, + "theoretical_loss": 3.693201967307514, + "tokens_seen": 956497920 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035867602808425274, + "loss": 1.6164, + "theoretical_loss": 3.6931778755733164, + "tokens_seen": 956563456 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358665997993982, + "loss": 1.4639, + "theoretical_loss": 3.6931537859517585, + "tokens_seen": 956628992 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035865596790371116, + "loss": 1.4957, + "theoretical_loss": 3.6931296984425077, + "tokens_seen": 956694528 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035864593781344034, + "loss": 1.3994, + "theoretical_loss": 3.693105613045236, + "tokens_seen": 956760064 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 351696, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4505877494812012, + "objective/train/theoretical_loss": 3.693081529759612, + "objective/train/tokens_used": 977285600, + "theoretical_loss": 3.693081529759612, + "tokens_seen": 956825600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586359077231695, + "loss": 1.425, + "theoretical_loss": 3.693081529759612, + "tokens_seen": 956825600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586258776328987, + "loss": 1.4943, + "theoretical_loss": 3.6930574485853076, + "tokens_seen": 956891136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586158475426279, + "loss": 1.5484, + "theoretical_loss": 3.6930333695219915, + "tokens_seen": 956956672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003586058174523571, + "loss": 1.4308, + "theoretical_loss": 3.693009292569336, + "tokens_seen": 957022208 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035859578736208625, + "loss": 1.4428, + "theoretical_loss": 3.69298521772701, + "tokens_seen": 957087744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003585857572718155, + "loss": 1.4682, + "theoretical_loss": 3.692961144994684, + "tokens_seen": 957153280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003585757271815446, + "loss": 1.5572, + "theoretical_loss": 3.6929370743720296, + "tokens_seen": 957218816 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035856569709127384, + "loss": 1.55, + "theoretical_loss": 3.692913005858717, + "tokens_seen": 957284352 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358555667001003, + "loss": 1.427, + "theoretical_loss": 3.6928889394544173, + "tokens_seen": 957349888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003585456369107322, + "loss": 1.4176, + "theoretical_loss": 3.6928648751588007, + "tokens_seen": 957415424 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003585356068204614, + "loss": 1.49, + "theoretical_loss": 3.6928408129715384, + "tokens_seen": 957480960 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035852557673019057, + "loss": 1.4718, + "theoretical_loss": 3.692816752892302, + "tokens_seen": 957546496 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035851554663991975, + "loss": 1.5193, + "theoretical_loss": 3.6927926949207617, + "tokens_seen": 957612032 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358505516549649, + "loss": 1.4476, + "theoretical_loss": 3.692768639056589, + "tokens_seen": 957677568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003584954864593781, + "loss": 1.3406, + "theoretical_loss": 3.692744585299455, + "tokens_seen": 957743104 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035848545636910735, + "loss": 1.4305, + "theoretical_loss": 3.692720533649031, + "tokens_seen": 957808640 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035847542627883653, + "loss": 1.3918, + "theoretical_loss": 3.6926964841049887, + "tokens_seen": 957874176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003584653961885657, + "loss": 1.5438, + "theoretical_loss": 3.692672436666999, + "tokens_seen": 957939712 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003584553660982949, + "loss": 1.3997, + "theoretical_loss": 3.6926483913347337, + "tokens_seen": 958005248 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035844533600802407, + "loss": 1.4575, + "theoretical_loss": 3.692624348107864, + "tokens_seen": 958070784 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035843530591775325, + "loss": 1.5151, + "theoretical_loss": 3.692600306986063, + "tokens_seen": 958136320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003584252758274825, + "loss": 1.4862, + "theoretical_loss": 3.692576267969, + "tokens_seen": 958201856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003584152457372116, + "loss": 1.5111, + "theoretical_loss": 3.6925522310563488, + "tokens_seen": 958267392 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035840521564694085, + "loss": 1.5202, + "theoretical_loss": 3.6925281962477805, + "tokens_seen": 958332928 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035839518555667, + "loss": 1.4827, + "theoretical_loss": 3.692504163542967, + "tokens_seen": 958398464 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583851554663992, + "loss": 1.4371, + "theoretical_loss": 3.692480132941581, + "tokens_seen": 958464000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583751253761284, + "loss": 1.5044, + "theoretical_loss": 3.6924561044432935, + "tokens_seen": 958529536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583650952858576, + "loss": 1.3911, + "theoretical_loss": 3.6924320780477773, + "tokens_seen": 958595072 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035835506519558676, + "loss": 1.6399, + "theoretical_loss": 3.692408053754705, + "tokens_seen": 958660608 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035834503510531594, + "loss": 1.4565, + "theoretical_loss": 3.6923840315637477, + "tokens_seen": 958726144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583350050150451, + "loss": 1.5241, + "theoretical_loss": 3.6923600114745794, + "tokens_seen": 958791680 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035832497492477435, + "loss": 1.439, + "theoretical_loss": 3.6923359934868714, + "tokens_seen": 958857216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583149448345035, + "loss": 1.434, + "theoretical_loss": 3.692311977600297, + "tokens_seen": 958922752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003583049147442327, + "loss": 1.4795, + "theoretical_loss": 3.6922879638145276, + "tokens_seen": 958988288 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003582948846539619, + "loss": 1.4152, + "theoretical_loss": 3.6922639521292373, + "tokens_seen": 959053824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003582848545636911, + "loss": 1.4193, + "theoretical_loss": 3.6922399425440977, + "tokens_seen": 959119360 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035827482447342026, + "loss": 1.4449, + "theoretical_loss": 3.6922159350587824, + "tokens_seen": 959184896 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035826479438314944, + "loss": 1.4513, + "theoretical_loss": 3.692191929672964, + "tokens_seen": 959250432 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003582547642928786, + "loss": 1.4857, + "theoretical_loss": 3.6921679263863156, + "tokens_seen": 959315968 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035824473420260786, + "loss": 1.2947, + "theoretical_loss": 3.6921439251985104, + "tokens_seen": 959381504 + }, + { + "epoch": 0.29, + "learning_rate": 0.000358234704112337, + "loss": 1.4732, + "theoretical_loss": 3.6921199261092212, + "tokens_seen": 959447040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003582246740220662, + "loss": 1.3641, + "theoretical_loss": 3.6920959291181212, + "tokens_seen": 959512576 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035821464393179535, + "loss": 1.5697, + "theoretical_loss": 3.692071934224884, + "tokens_seen": 959578112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003582046138415246, + "loss": 1.4067, + "theoretical_loss": 3.6920479414291822, + "tokens_seen": 959643648 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035819458375125376, + "loss": 1.502, + "theoretical_loss": 3.6920239507306905, + "tokens_seen": 959709184 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035818455366098294, + "loss": 1.5136, + "theoretical_loss": 3.6919999621290813, + "tokens_seen": 959774720 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003581745235707121, + "loss": 1.5305, + "theoretical_loss": 3.691975975624028, + "tokens_seen": 959840256 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035816449348044136, + "loss": 1.5416, + "theoretical_loss": 3.6919519912152055, + "tokens_seen": 959905792 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003581544633901705, + "loss": 1.4106, + "theoretical_loss": 3.691928008902286, + "tokens_seen": 959971328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003581444332998997, + "loss": 1.5076, + "theoretical_loss": 3.6919040286849443, + "tokens_seen": 960036864 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 352874, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3872644901275635, + "objective/train/theoretical_loss": 3.6918800505628546, + "objective/train/tokens_used": 980562400, + "theoretical_loss": 3.6918800505628546, + "tokens_seen": 960102400 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035813440320962885, + "loss": 1.4203, + "theoretical_loss": 3.6918800505628546, + "tokens_seen": 960102400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003581243731193581, + "loss": 1.4639, + "theoretical_loss": 3.691856074535689, + "tokens_seen": 960167936 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035811434302908727, + "loss": 1.4291, + "theoretical_loss": 3.691832100603124, + "tokens_seen": 960233472 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035810431293881645, + "loss": 1.521, + "theoretical_loss": 3.691808128764831, + "tokens_seen": 960299008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580942828485457, + "loss": 1.4519, + "theoretical_loss": 3.6917841590204867, + "tokens_seen": 960364544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580842527582748, + "loss": 1.4161, + "theoretical_loss": 3.6917601913697635, + "tokens_seen": 960430080 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035807422266800404, + "loss": 1.3722, + "theoretical_loss": 3.6917362258123365, + "tokens_seen": 960495616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580641925777332, + "loss": 1.343, + "theoretical_loss": 3.69171226234788, + "tokens_seen": 960561152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580541624874624, + "loss": 1.406, + "theoretical_loss": 3.691688300976068, + "tokens_seen": 960626688 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580441323971916, + "loss": 1.4927, + "theoretical_loss": 3.6916643416965753, + "tokens_seen": 960692224 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035803410230692077, + "loss": 1.4425, + "theoretical_loss": 3.6916403845090766, + "tokens_seen": 960757760 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035802407221664995, + "loss": 1.3669, + "theoretical_loss": 3.691616429413247, + "tokens_seen": 960823296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580140421263792, + "loss": 1.4408, + "theoretical_loss": 3.6915924764087604, + "tokens_seen": 960888832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003580040120361083, + "loss": 1.4099, + "theoretical_loss": 3.6915685254952915, + "tokens_seen": 960954368 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035799398194583755, + "loss": 1.4148, + "theoretical_loss": 3.6915445766725163, + "tokens_seen": 961019904 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035798395185556673, + "loss": 1.4777, + "theoretical_loss": 3.691520629940108, + "tokens_seen": 961085440 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003579739217652959, + "loss": 1.4489, + "theoretical_loss": 3.6914966852977438, + "tokens_seen": 961150976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003579638916750251, + "loss": 1.3947, + "theoretical_loss": 3.691472742745097, + "tokens_seen": 961216512 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035795386158475427, + "loss": 1.4936, + "theoretical_loss": 3.6914488022818435, + "tokens_seen": 961282048 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035794383149448345, + "loss": 1.4834, + "theoretical_loss": 3.691424863907659, + "tokens_seen": 961347584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003579338014042127, + "loss": 1.4087, + "theoretical_loss": 3.691400927622217, + "tokens_seen": 961413120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003579237713139418, + "loss": 1.399, + "theoretical_loss": 3.6913769934251954, + "tokens_seen": 961478656 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035791374122367105, + "loss": 1.3227, + "theoretical_loss": 3.691353061316268, + "tokens_seen": 961544192 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003579037111334002, + "loss": 1.4457, + "theoretical_loss": 3.6913291312951104, + "tokens_seen": 961609728 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578936810431294, + "loss": 1.4758, + "theoretical_loss": 3.6913052033613987, + "tokens_seen": 961675264 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578836509528586, + "loss": 1.4185, + "theoretical_loss": 3.6912812775148085, + "tokens_seen": 961740800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578736208625878, + "loss": 1.4918, + "theoretical_loss": 3.691257353755015, + "tokens_seen": 961806336 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035786359077231696, + "loss": 1.4566, + "theoretical_loss": 3.691233432081695, + "tokens_seen": 961871872 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035785356068204614, + "loss": 1.5129, + "theoretical_loss": 3.691209512494524, + "tokens_seen": 961937408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578435305917753, + "loss": 1.5654, + "theoretical_loss": 3.6911855949931773, + "tokens_seen": 962002944 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035783350050150455, + "loss": 1.4005, + "theoretical_loss": 3.6911616795773314, + "tokens_seen": 962068480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578234704112337, + "loss": 1.447, + "theoretical_loss": 3.691137766246663, + "tokens_seen": 962134016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578134403209629, + "loss": 1.5059, + "theoretical_loss": 3.6911138550008475, + "tokens_seen": 962199552 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003578034102306921, + "loss": 1.4775, + "theoretical_loss": 3.691089945839561, + "tokens_seen": 962265088 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003577933801404213, + "loss": 1.443, + "theoretical_loss": 3.6910660387624805, + "tokens_seen": 962330624 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035778335005015046, + "loss": 1.3831, + "theoretical_loss": 3.6910421337692823, + "tokens_seen": 962396160 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035777331995987964, + "loss": 1.435, + "theoretical_loss": 3.6910182308596426, + "tokens_seen": 962461696 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003577632898696088, + "loss": 1.4889, + "theoretical_loss": 3.6909943300332384, + "tokens_seen": 962527232 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035775325977933806, + "loss": 1.4826, + "theoretical_loss": 3.6909704312897453, + "tokens_seen": 962592768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003577432296890672, + "loss": 1.3526, + "theoretical_loss": 3.690946534628841, + "tokens_seen": 962658304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003577331995987964, + "loss": 1.3398, + "theoretical_loss": 3.6909226400502018, + "tokens_seen": 962723840 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035772316950852555, + "loss": 1.4001, + "theoretical_loss": 3.6908987475535047, + "tokens_seen": 962789376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003577131394182548, + "loss": 1.4782, + "theoretical_loss": 3.690874857138426, + "tokens_seen": 962854912 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035770310932798396, + "loss": 1.4751, + "theoretical_loss": 3.690850968804644, + "tokens_seen": 962920448 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035769307923771314, + "loss": 1.351, + "theoretical_loss": 3.6908270825518343, + "tokens_seen": 962985984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003576830491474423, + "loss": 1.467, + "theoretical_loss": 3.690803198379675, + "tokens_seen": 963051520 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035767301905717156, + "loss": 1.346, + "theoretical_loss": 3.690779316287843, + "tokens_seen": 963117056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003576629889669007, + "loss": 1.3667, + "theoretical_loss": 3.6907554362760155, + "tokens_seen": 963182592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003576529588766299, + "loss": 1.4234, + "theoretical_loss": 3.6907315583438693, + "tokens_seen": 963248128 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035764292878635905, + "loss": 1.5062, + "theoretical_loss": 3.690707682491083, + "tokens_seen": 963313664 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 353969, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3839304447174072, + "objective/train/theoretical_loss": 3.690683808717333, + "objective/train/tokens_used": 983839200, + "theoretical_loss": 3.690683808717333, + "tokens_seen": 963379200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003576328986960883, + "loss": 1.3493, + "theoretical_loss": 3.690683808717333, + "tokens_seen": 963379200 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035762286860581747, + "loss": 1.4468, + "theoretical_loss": 3.6906599370222977, + "tokens_seen": 963444736 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035761283851554665, + "loss": 1.5983, + "theoretical_loss": 3.690636067405654, + "tokens_seen": 963510272 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035760280842527583, + "loss": 1.396, + "theoretical_loss": 3.69061219986708, + "tokens_seen": 963575808 + }, + { + "epoch": 0.29, + "learning_rate": 0.000357592778335005, + "loss": 1.4081, + "theoretical_loss": 3.6905883344062533, + "tokens_seen": 963641344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575827482447342, + "loss": 1.3758, + "theoretical_loss": 3.690564471022852, + "tokens_seen": 963706880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575727181544634, + "loss": 1.3639, + "theoretical_loss": 3.690540609716554, + "tokens_seen": 963772416 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035756268806419255, + "loss": 1.3996, + "theoretical_loss": 3.6905167504870366, + "tokens_seen": 963837952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575526579739218, + "loss": 1.4201, + "theoretical_loss": 3.6904928933339787, + "tokens_seen": 963903488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575426278836509, + "loss": 1.5436, + "theoretical_loss": 3.690469038257058, + "tokens_seen": 963969024 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035753259779338015, + "loss": 1.3841, + "theoretical_loss": 3.6904451852559528, + "tokens_seen": 964034560 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035752256770310933, + "loss": 1.5271, + "theoretical_loss": 3.6904213343303414, + "tokens_seen": 964100096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575125376128385, + "loss": 1.3967, + "theoretical_loss": 3.690397485479902, + "tokens_seen": 964165632 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003575025075225677, + "loss": 1.428, + "theoretical_loss": 3.690373638704313, + "tokens_seen": 964231168 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035749247743229693, + "loss": 1.46, + "theoretical_loss": 3.6903497940032532, + "tokens_seen": 964296704 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035748244734202606, + "loss": 1.2993, + "theoretical_loss": 3.690325951376401, + "tokens_seen": 964362240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003574724172517553, + "loss": 1.4691, + "theoretical_loss": 3.690302110823435, + "tokens_seen": 964427776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003574623871614844, + "loss": 1.5454, + "theoretical_loss": 3.6902782723440337, + "tokens_seen": 964493312 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035745235707121365, + "loss": 1.4598, + "theoretical_loss": 3.6902544359378755, + "tokens_seen": 964558848 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035744232698094283, + "loss": 1.3934, + "theoretical_loss": 3.6902306016046404, + "tokens_seen": 964624384 + }, + { + "epoch": 0.29, + "learning_rate": 0.000357432296890672, + "loss": 1.5414, + "theoretical_loss": 3.690206769344007, + "tokens_seen": 964689920 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003574222668004012, + "loss": 1.3808, + "theoretical_loss": 3.6901829391556538, + "tokens_seen": 964755456 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003574122367101304, + "loss": 1.5835, + "theoretical_loss": 3.69015911103926, + "tokens_seen": 964820992 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035740220661985956, + "loss": 1.3667, + "theoretical_loss": 3.690135284994504, + "tokens_seen": 964886528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573921765295888, + "loss": 1.4865, + "theoretical_loss": 3.6901114610210666, + "tokens_seen": 964952064 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573821464393179, + "loss": 1.4081, + "theoretical_loss": 3.6900876391186266, + "tokens_seen": 965017600 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035737211634904716, + "loss": 1.39, + "theoretical_loss": 3.6900638192868627, + "tokens_seen": 965083136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573620862587763, + "loss": 1.3264, + "theoretical_loss": 3.690040001525454, + "tokens_seen": 965148672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573520561685055, + "loss": 1.5069, + "theoretical_loss": 3.6900161858340814, + "tokens_seen": 965214208 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035734202607823475, + "loss": 1.3131, + "theoretical_loss": 3.689992372212423, + "tokens_seen": 965279744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573319959879639, + "loss": 1.557, + "theoretical_loss": 3.6899685606601595, + "tokens_seen": 965345280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573219658976931, + "loss": 1.6166, + "theoretical_loss": 3.68994475117697, + "tokens_seen": 965410816 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573119358074223, + "loss": 1.4599, + "theoretical_loss": 3.6899209437625347, + "tokens_seen": 965476352 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003573019057171515, + "loss": 1.3983, + "theoretical_loss": 3.6898971384165327, + "tokens_seen": 965541888 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035729187562688066, + "loss": 1.3949, + "theoretical_loss": 3.6898733351386444, + "tokens_seen": 965607424 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035728184553660984, + "loss": 1.4266, + "theoretical_loss": 3.68984953392855, + "tokens_seen": 965672960 + }, + { + "epoch": 0.29, + "learning_rate": 0.000357271815446339, + "loss": 1.4404, + "theoretical_loss": 3.68982573478593, + "tokens_seen": 965738496 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035726178535606826, + "loss": 1.4217, + "theoretical_loss": 3.689801937710463, + "tokens_seen": 965804032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003572517552657974, + "loss": 1.5289, + "theoretical_loss": 3.6897781427018304, + "tokens_seen": 965869568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003572417251755266, + "loss": 1.5216, + "theoretical_loss": 3.6897543497597125, + "tokens_seen": 965935104 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035723169508525575, + "loss": 1.4433, + "theoretical_loss": 3.689730558883789, + "tokens_seen": 966000640 + }, + { + "epoch": 0.29, + "learning_rate": 0.000357221664994985, + "loss": 1.4896, + "theoretical_loss": 3.68970677007374, + "tokens_seen": 966066176 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035721163490471416, + "loss": 1.3175, + "theoretical_loss": 3.6896829833292477, + "tokens_seen": 966131712 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035720160481444334, + "loss": 1.526, + "theoretical_loss": 3.6896591986499905, + "tokens_seen": 966197248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003571915747241725, + "loss": 1.444, + "theoretical_loss": 3.689635416035651, + "tokens_seen": 966262784 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035718154463390176, + "loss": 1.473, + "theoretical_loss": 3.6896116354859085, + "tokens_seen": 966328320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003571715145436309, + "loss": 1.5949, + "theoretical_loss": 3.6895878570004443, + "tokens_seen": 966393856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003571614844533601, + "loss": 1.4537, + "theoretical_loss": 3.689564080578939, + "tokens_seen": 966459392 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035715145436308925, + "loss": 1.4869, + "theoretical_loss": 3.689540306221074, + "tokens_seen": 966524928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003571414242728185, + "loss": 1.4524, + "theoretical_loss": 3.6895165339265295, + "tokens_seen": 966590464 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 355190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.541426658630371, + "objective/train/theoretical_loss": 3.6894927636949877, + "objective/train/tokens_used": 987116000, + "theoretical_loss": 3.6894927636949877, + "tokens_seen": 966656000 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035713139418254767, + "loss": 1.5599, + "theoretical_loss": 3.6894927636949877, + "tokens_seen": 966656000 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035712136409227685, + "loss": 1.5257, + "theoretical_loss": 3.689468995526128, + "tokens_seen": 966721536 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035711133400200603, + "loss": 1.5751, + "theoretical_loss": 3.6894452294196336, + "tokens_seen": 966787072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003571013039117352, + "loss": 1.4713, + "theoretical_loss": 3.6894214653751845, + "tokens_seen": 966852608 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003570912738214644, + "loss": 1.4342, + "theoretical_loss": 3.689397703392462, + "tokens_seen": 966918144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003570812437311936, + "loss": 1.4215, + "theoretical_loss": 3.689373943471148, + "tokens_seen": 966983680 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035707121364092275, + "loss": 1.3781, + "theoretical_loss": 3.689350185610924, + "tokens_seen": 967049216 + }, + { + "epoch": 0.29, + "learning_rate": 0.000357061183550652, + "loss": 1.3838, + "theoretical_loss": 3.6893264298114716, + "tokens_seen": 967114752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003570511534603811, + "loss": 1.4433, + "theoretical_loss": 3.689302676072472, + "tokens_seen": 967180288 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035704112337011035, + "loss": 1.4348, + "theoretical_loss": 3.689278924393607, + "tokens_seen": 967245824 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035703109327983953, + "loss": 1.4577, + "theoretical_loss": 3.689255174774558, + "tokens_seen": 967311360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003570210631895687, + "loss": 1.4184, + "theoretical_loss": 3.6892314272150077, + "tokens_seen": 967376896 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003570110330992979, + "loss": 1.3851, + "theoretical_loss": 3.6892076817146373, + "tokens_seen": 967442432 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035700100300902713, + "loss": 1.424, + "theoretical_loss": 3.6891839382731297, + "tokens_seen": 967507968 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035699097291875626, + "loss": 1.4767, + "theoretical_loss": 3.6891601968901657, + "tokens_seen": 967573504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003569809428284855, + "loss": 1.3717, + "theoretical_loss": 3.689136457565428, + "tokens_seen": 967639040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003569709127382146, + "loss": 1.4404, + "theoretical_loss": 3.689112720298599, + "tokens_seen": 967704576 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035696088264794385, + "loss": 1.5107, + "theoretical_loss": 3.6890889850893602, + "tokens_seen": 967770112 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035695085255767303, + "loss": 1.4977, + "theoretical_loss": 3.689065251937395, + "tokens_seen": 967835648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003569408224674022, + "loss": 1.4503, + "theoretical_loss": 3.6890415208423857, + "tokens_seen": 967901184 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003569307923771314, + "loss": 1.3878, + "theoretical_loss": 3.6890177918040132, + "tokens_seen": 967966720 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003569207622868606, + "loss": 1.4418, + "theoretical_loss": 3.688994064821962, + "tokens_seen": 968032256 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035691073219658976, + "loss": 1.5133, + "theoretical_loss": 3.6889703398959135, + "tokens_seen": 968097792 + }, + { + "epoch": 0.29, + "learning_rate": 0.000356900702106319, + "loss": 1.4172, + "theoretical_loss": 3.6889466170255503, + "tokens_seen": 968163328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568906720160481, + "loss": 1.532, + "theoretical_loss": 3.6889228962105562, + "tokens_seen": 968228864 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035688064192577736, + "loss": 1.4543, + "theoretical_loss": 3.6888991774506135, + "tokens_seen": 968294400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568706118355065, + "loss": 1.4295, + "theoretical_loss": 3.688875460745405, + "tokens_seen": 968359936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568605817452357, + "loss": 1.4974, + "theoretical_loss": 3.6888517460946133, + "tokens_seen": 968425472 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568505516549649, + "loss": 1.3001, + "theoretical_loss": 3.6888280334979218, + "tokens_seen": 968491008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568405215646941, + "loss": 1.4875, + "theoretical_loss": 3.688804322955014, + "tokens_seen": 968556544 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035683049147442326, + "loss": 1.4298, + "theoretical_loss": 3.688780614465572, + "tokens_seen": 968622080 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568204613841525, + "loss": 1.4249, + "theoretical_loss": 3.68875690802928, + "tokens_seen": 968687616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003568104312938816, + "loss": 1.3504, + "theoretical_loss": 3.688733203645821, + "tokens_seen": 968753152 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035680040120361086, + "loss": 1.5116, + "theoretical_loss": 3.6887095013148787, + "tokens_seen": 968818688 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035679037111334, + "loss": 1.4851, + "theoretical_loss": 3.6886858010361356, + "tokens_seen": 968884224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003567803410230692, + "loss": 1.5344, + "theoretical_loss": 3.6886621028092765, + "tokens_seen": 968949760 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003567703109327984, + "loss": 1.4755, + "theoretical_loss": 3.688638406633984, + "tokens_seen": 969015296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003567602808425276, + "loss": 1.3172, + "theoretical_loss": 3.688614712509942, + "tokens_seen": 969080832 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035675025075225677, + "loss": 1.33, + "theoretical_loss": 3.688591020436835, + "tokens_seen": 969146368 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035674022066198595, + "loss": 1.5218, + "theoretical_loss": 3.6885673304143456, + "tokens_seen": 969211904 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035673019057171513, + "loss": 1.4475, + "theoretical_loss": 3.6885436424421583, + "tokens_seen": 969277440 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035672016048144436, + "loss": 1.4586, + "theoretical_loss": 3.688519956519957, + "tokens_seen": 969342976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003567101303911735, + "loss": 1.4599, + "theoretical_loss": 3.6884962726474253, + "tokens_seen": 969408512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003567001003009027, + "loss": 1.4882, + "theoretical_loss": 3.688472590824248, + "tokens_seen": 969474048 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035669007021063185, + "loss": 1.4239, + "theoretical_loss": 3.6884489110501093, + "tokens_seen": 969539584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003566800401203611, + "loss": 1.3925, + "theoretical_loss": 3.6884252333246925, + "tokens_seen": 969605120 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035667001003009027, + "loss": 1.4456, + "theoretical_loss": 3.6884015576476825, + "tokens_seen": 969670656 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035665997993981945, + "loss": 1.4277, + "theoretical_loss": 3.688377884018764, + "tokens_seen": 969736192 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035664994984954863, + "loss": 1.4017, + "theoretical_loss": 3.6883542124376207, + "tokens_seen": 969801728 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035663991975927787, + "loss": 1.5431, + "theoretical_loss": 3.6883305429039375, + "tokens_seen": 969867264 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 356486, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3990880250930786, + "objective/train/theoretical_loss": 3.688306875417399, + "objective/train/tokens_used": 990392800, + "theoretical_loss": 3.688306875417399, + "tokens_seen": 969932800 + }, + { + "epoch": 0.29, + "learning_rate": 0.000356629889669007, + "loss": 1.4214, + "theoretical_loss": 3.688306875417399, + "tokens_seen": 969932800 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035661985957873623, + "loss": 1.392, + "theoretical_loss": 3.688283209977689, + "tokens_seen": 969998336 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035660982948846536, + "loss": 1.3883, + "theoretical_loss": 3.6882595465844936, + "tokens_seen": 970063872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565997993981946, + "loss": 1.5155, + "theoretical_loss": 3.6882358852374972, + "tokens_seen": 970129408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565897693079238, + "loss": 1.4917, + "theoretical_loss": 3.688212225936384, + "tokens_seen": 970194944 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035657973921765295, + "loss": 1.3664, + "theoretical_loss": 3.6881885686808396, + "tokens_seen": 970260480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565697091273822, + "loss": 1.4666, + "theoretical_loss": 3.688164913470548, + "tokens_seen": 970326016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565596790371113, + "loss": 1.5294, + "theoretical_loss": 3.6881412603051955, + "tokens_seen": 970391552 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035654964894684055, + "loss": 1.3706, + "theoretical_loss": 3.6881176091844665, + "tokens_seen": 970457088 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035653961885656973, + "loss": 1.3938, + "theoretical_loss": 3.6880939601080467, + "tokens_seen": 970522624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565295887662989, + "loss": 1.4637, + "theoretical_loss": 3.688070313075621, + "tokens_seen": 970588160 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003565195586760281, + "loss": 1.3705, + "theoretical_loss": 3.6880466680868746, + "tokens_seen": 970653696 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035650952858575733, + "loss": 1.5717, + "theoretical_loss": 3.6880230251414936, + "tokens_seen": 970719232 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035649949849548646, + "loss": 1.479, + "theoretical_loss": 3.6879993842391627, + "tokens_seen": 970784768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564894684052157, + "loss": 1.421, + "theoretical_loss": 3.6879757453795676, + "tokens_seen": 970850304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564794383149448, + "loss": 1.4372, + "theoretical_loss": 3.6879521085623943, + "tokens_seen": 970915840 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035646940822467405, + "loss": 1.4794, + "theoretical_loss": 3.687928473787328, + "tokens_seen": 970981376 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035645937813440323, + "loss": 1.4415, + "theoretical_loss": 3.6879048410540554, + "tokens_seen": 971046912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564493480441324, + "loss": 1.4051, + "theoretical_loss": 3.687881210362261, + "tokens_seen": 971112448 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564393179538616, + "loss": 1.5114, + "theoretical_loss": 3.687857581711631, + "tokens_seen": 971177984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564292878635908, + "loss": 1.4259, + "theoretical_loss": 3.687833955101852, + "tokens_seen": 971243520 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035641925777331996, + "loss": 1.5216, + "theoretical_loss": 3.68781033053261, + "tokens_seen": 971309056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003564092276830492, + "loss": 1.3935, + "theoretical_loss": 3.687786708003591, + "tokens_seen": 971374592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563991975927783, + "loss": 1.4946, + "theoretical_loss": 3.6877630875144805, + "tokens_seen": 971440128 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035638916750250756, + "loss": 1.5024, + "theoretical_loss": 3.687739469064966, + "tokens_seen": 971505664 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563791374122367, + "loss": 1.4311, + "theoretical_loss": 3.6877158526547325, + "tokens_seen": 971571200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563691073219659, + "loss": 1.5041, + "theoretical_loss": 3.6876922382834674, + "tokens_seen": 971636736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563590772316951, + "loss": 1.4146, + "theoretical_loss": 3.687668625950856, + "tokens_seen": 971702272 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563490471414243, + "loss": 1.3916, + "theoretical_loss": 3.687645015656586, + "tokens_seen": 971767808 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035633901705115346, + "loss": 1.4733, + "theoretical_loss": 3.6876214074003437, + "tokens_seen": 971833344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563289869608827, + "loss": 1.4569, + "theoretical_loss": 3.6875978011818153, + "tokens_seen": 971898880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003563189568706118, + "loss": 1.4101, + "theoretical_loss": 3.6875741970006883, + "tokens_seen": 971964416 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035630892678034106, + "loss": 1.4821, + "theoretical_loss": 3.6875505948566487, + "tokens_seen": 972029952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562988966900702, + "loss": 1.4003, + "theoretical_loss": 3.6875269947493834, + "tokens_seen": 972095488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562888665997994, + "loss": 1.5144, + "theoretical_loss": 3.68750339667858, + "tokens_seen": 972161024 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562788365095286, + "loss": 1.4227, + "theoretical_loss": 3.687479800643925, + "tokens_seen": 972226560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562688064192578, + "loss": 1.3428, + "theoretical_loss": 3.687456206645106, + "tokens_seen": 972292096 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035625877632898697, + "loss": 1.4156, + "theoretical_loss": 3.6874326146818097, + "tokens_seen": 972357632 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035624874623871615, + "loss": 1.5006, + "theoretical_loss": 3.687409024753723, + "tokens_seen": 972423168 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035623871614844533, + "loss": 1.3227, + "theoretical_loss": 3.6873854368605334, + "tokens_seen": 972488704 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035622868605817456, + "loss": 1.5146, + "theoretical_loss": 3.687361851001929, + "tokens_seen": 972554240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562186559679037, + "loss": 1.3838, + "theoretical_loss": 3.687338267177596, + "tokens_seen": 972619776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003562086258776329, + "loss": 1.458, + "theoretical_loss": 3.687314685387223, + "tokens_seen": 972685312 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035619859578736205, + "loss": 1.5107, + "theoretical_loss": 3.6872911056304973, + "tokens_seen": 972750848 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003561885656970913, + "loss": 1.3889, + "theoretical_loss": 3.687267527907106, + "tokens_seen": 972816384 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035617853560682047, + "loss": 1.5406, + "theoretical_loss": 3.687243952216737, + "tokens_seen": 972881920 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035616850551654965, + "loss": 1.3086, + "theoretical_loss": 3.6872203785590782, + "tokens_seen": 972947456 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035615847542627883, + "loss": 1.4215, + "theoretical_loss": 3.6871968069338177, + "tokens_seen": 973012992 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035614844533600807, + "loss": 1.5611, + "theoretical_loss": 3.6871732373406436, + "tokens_seen": 973078528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003561384152457372, + "loss": 1.5613, + "theoretical_loss": 3.6871496697792425, + "tokens_seen": 973144064 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 357717, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5072946548461914, + "objective/train/theoretical_loss": 3.6871261042493035, + "objective/train/tokens_used": 993669600, + "theoretical_loss": 3.6871261042493035, + "tokens_seen": 973209600 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035612838515546643, + "loss": 1.5493, + "theoretical_loss": 3.6871261042493035, + "tokens_seen": 973209600 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035611835506519556, + "loss": 1.4724, + "theoretical_loss": 3.687102540750515, + "tokens_seen": 973275136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0003561083249749248, + "loss": 1.3616, + "theoretical_loss": 3.687078979282564, + "tokens_seen": 973340672 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035609829488465397, + "loss": 1.3912, + "theoretical_loss": 3.6870554198451404, + "tokens_seen": 973406208 + }, + { + "epoch": 0.29, + "learning_rate": 0.00035608826479438315, + "loss": 1.477, + "theoretical_loss": 3.687031862437931, + "tokens_seen": 973471744 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035607823470411233, + "loss": 1.3878, + "theoretical_loss": 3.687008307060625, + "tokens_seen": 973537280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003560682046138415, + "loss": 1.361, + "theoretical_loss": 3.6869847537129106, + "tokens_seen": 973602816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003560581745235707, + "loss": 1.445, + "theoretical_loss": 3.6869612023944764, + "tokens_seen": 973668352 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035604814443329993, + "loss": 1.4746, + "theoretical_loss": 3.6869376531050113, + "tokens_seen": 973733888 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035603811434302906, + "loss": 1.4456, + "theoretical_loss": 3.686914105844203, + "tokens_seen": 973799424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003560280842527583, + "loss": 1.339, + "theoretical_loss": 3.686890560611741, + "tokens_seen": 973864960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003560180541624874, + "loss": 1.4436, + "theoretical_loss": 3.686867017407314, + "tokens_seen": 973930496 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035600802407221666, + "loss": 1.3668, + "theoretical_loss": 3.686843476230611, + "tokens_seen": 973996032 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035599799398194584, + "loss": 1.4969, + "theoretical_loss": 3.686819937081321, + "tokens_seen": 974061568 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355987963891675, + "loss": 1.325, + "theoretical_loss": 3.6867963999591327, + "tokens_seen": 974127104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003559779338014042, + "loss": 1.3895, + "theoretical_loss": 3.686772864863735, + "tokens_seen": 974192640 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035596790371113344, + "loss": 1.4492, + "theoretical_loss": 3.6867493317948172, + "tokens_seen": 974258176 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035595787362086256, + "loss": 1.4832, + "theoretical_loss": 3.686725800752069, + "tokens_seen": 974323712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003559478435305918, + "loss": 1.3166, + "theoretical_loss": 3.6867022717351787, + "tokens_seen": 974389248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003559378134403209, + "loss": 1.5419, + "theoretical_loss": 3.686678744743837, + "tokens_seen": 974454784 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035592778335005016, + "loss": 1.4158, + "theoretical_loss": 3.686655219777732, + "tokens_seen": 974520320 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035591775325977934, + "loss": 1.4484, + "theoretical_loss": 3.686631696836554, + "tokens_seen": 974585856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003559077231695085, + "loss": 1.574, + "theoretical_loss": 3.686608175919992, + "tokens_seen": 974651392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003558976930792377, + "loss": 1.3583, + "theoretical_loss": 3.6865846570277365, + "tokens_seen": 974716928 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003558876629889669, + "loss": 1.4559, + "theoretical_loss": 3.6865611401594762, + "tokens_seen": 974782464 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035587763289869607, + "loss": 1.4087, + "theoretical_loss": 3.6865376253149016, + "tokens_seen": 974848000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003558676028084253, + "loss": 1.4265, + "theoretical_loss": 3.6865141124937013, + "tokens_seen": 974913536 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035585757271815443, + "loss": 1.4863, + "theoretical_loss": 3.686490601695567, + "tokens_seen": 974979072 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035584754262788366, + "loss": 1.4904, + "theoretical_loss": 3.6864670929201875, + "tokens_seen": 975044608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003558375125376129, + "loss": 1.4562, + "theoretical_loss": 3.686443586167253, + "tokens_seen": 975110144 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355827482447342, + "loss": 1.5381, + "theoretical_loss": 3.686420081436454, + "tokens_seen": 975175680 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035581745235707126, + "loss": 1.447, + "theoretical_loss": 3.6863965787274804, + "tokens_seen": 975241216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003558074222668004, + "loss": 1.4884, + "theoretical_loss": 3.686373078040022, + "tokens_seen": 975306752 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003557973921765296, + "loss": 1.4456, + "theoretical_loss": 3.6863495793737697, + "tokens_seen": 975372288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003557873620862588, + "loss": 1.5332, + "theoretical_loss": 3.6863260827284137, + "tokens_seen": 975437824 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355777331995988, + "loss": 1.4466, + "theoretical_loss": 3.6863025881036444, + "tokens_seen": 975503360 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035576730190571717, + "loss": 1.3402, + "theoretical_loss": 3.6862790954991524, + "tokens_seen": 975568896 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035575727181544635, + "loss": 1.4081, + "theoretical_loss": 3.686255604914628, + "tokens_seen": 975634432 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035574724172517553, + "loss": 1.519, + "theoretical_loss": 3.6862321163497622, + "tokens_seen": 975699968 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035573721163490476, + "loss": 1.4897, + "theoretical_loss": 3.686208629804246, + "tokens_seen": 975765504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003557271815446339, + "loss": 1.4828, + "theoretical_loss": 3.686185145277769, + "tokens_seen": 975831040 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003557171514543631, + "loss": 1.4529, + "theoretical_loss": 3.686161662770023, + "tokens_seen": 975896576 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035570712136409225, + "loss": 1.306, + "theoretical_loss": 3.686138182280699, + "tokens_seen": 975962112 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003556970912738215, + "loss": 1.4617, + "theoretical_loss": 3.686114703809488, + "tokens_seen": 976027648 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035568706118355067, + "loss": 1.4501, + "theoretical_loss": 3.68609122735608, + "tokens_seen": 976093184 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035567703109327985, + "loss": 1.3901, + "theoretical_loss": 3.6860677529201675, + "tokens_seen": 976158720 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035566700100300903, + "loss": 1.4301, + "theoretical_loss": 3.686044280501441, + "tokens_seen": 976224256 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035565697091273827, + "loss": 1.4219, + "theoretical_loss": 3.686020810099592, + "tokens_seen": 976289792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003556469408224674, + "loss": 1.4021, + "theoretical_loss": 3.6859973417143115, + "tokens_seen": 976355328 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035563691073219663, + "loss": 1.4546, + "theoretical_loss": 3.685973875345291, + "tokens_seen": 976420864 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 358882, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.267359733581543, + "objective/train/theoretical_loss": 3.6859504109922225, + "objective/train/tokens_used": 996946400, + "theoretical_loss": 3.6859504109922225, + "tokens_seen": 976486400 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035562688064192576, + "loss": 1.4088, + "theoretical_loss": 3.6859504109922225, + "tokens_seen": 976486400 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355616850551655, + "loss": 1.4784, + "theoretical_loss": 3.6859269486547968, + "tokens_seen": 976551936 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035560682046138417, + "loss": 1.395, + "theoretical_loss": 3.6859034883327055, + "tokens_seen": 976617472 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035559679037111335, + "loss": 1.4117, + "theoretical_loss": 3.6858800300256407, + "tokens_seen": 976683008 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035558676028084253, + "loss": 1.3995, + "theoretical_loss": 3.685856573733294, + "tokens_seen": 976748544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003555767301905717, + "loss": 1.4376, + "theoretical_loss": 3.6858331194553573, + "tokens_seen": 976814080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003555667001003009, + "loss": 1.4456, + "theoretical_loss": 3.6858096671915224, + "tokens_seen": 976879616 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035555667001003013, + "loss": 1.4705, + "theoretical_loss": 3.685786216941481, + "tokens_seen": 976945152 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035554663991975926, + "loss": 1.4058, + "theoretical_loss": 3.6857627687049255, + "tokens_seen": 977010688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003555366098294885, + "loss": 1.5828, + "theoretical_loss": 3.6857393224815476, + "tokens_seen": 977076224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003555265797392176, + "loss": 1.5275, + "theoretical_loss": 3.68571587827104, + "tokens_seen": 977141760 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035551654964894686, + "loss": 1.4012, + "theoretical_loss": 3.6856924360730936, + "tokens_seen": 977207296 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035550651955867604, + "loss": 1.4651, + "theoretical_loss": 3.6856689958874025, + "tokens_seen": 977272832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003554964894684052, + "loss": 1.5205, + "theoretical_loss": 3.6856455577136584, + "tokens_seen": 977338368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003554864593781344, + "loss": 1.5293, + "theoretical_loss": 3.6856221215515528, + "tokens_seen": 977403904 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035547642928786364, + "loss": 1.3603, + "theoretical_loss": 3.685598687400779, + "tokens_seen": 977469440 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035546639919759276, + "loss": 1.5137, + "theoretical_loss": 3.6855752552610292, + "tokens_seen": 977534976 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355456369107322, + "loss": 1.4653, + "theoretical_loss": 3.6855518251319968, + "tokens_seen": 977600512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003554463390170511, + "loss": 1.3851, + "theoretical_loss": 3.6855283970133734, + "tokens_seen": 977666048 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035543630892678036, + "loss": 1.3804, + "theoretical_loss": 3.6855049709048524, + "tokens_seen": 977731584 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035542627883650954, + "loss": 1.3819, + "theoretical_loss": 3.685481546806127, + "tokens_seen": 977797120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003554162487462387, + "loss": 1.3956, + "theoretical_loss": 3.685458124716889, + "tokens_seen": 977862656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003554062186559679, + "loss": 1.3582, + "theoretical_loss": 3.6854347046368314, + "tokens_seen": 977928192 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003553961885656971, + "loss": 1.5256, + "theoretical_loss": 3.6854112865656488, + "tokens_seen": 977993728 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035538615847542627, + "loss": 1.5924, + "theoretical_loss": 3.6853878705030327, + "tokens_seen": 978059264 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003553761283851555, + "loss": 1.4653, + "theoretical_loss": 3.685364456448677, + "tokens_seen": 978124800 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035536609829488463, + "loss": 1.3676, + "theoretical_loss": 3.685341044402274, + "tokens_seen": 978190336 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035535606820461386, + "loss": 1.4653, + "theoretical_loss": 3.6853176343635186, + "tokens_seen": 978255872 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035534603811434304, + "loss": 1.4755, + "theoretical_loss": 3.6852942263321022, + "tokens_seen": 978321408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003553360080240722, + "loss": 1.3995, + "theoretical_loss": 3.6852708203077205, + "tokens_seen": 978386944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003553259779338014, + "loss": 1.3926, + "theoretical_loss": 3.685247416290065, + "tokens_seen": 978452480 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003553159478435306, + "loss": 1.4055, + "theoretical_loss": 3.6852240142788304, + "tokens_seen": 978518016 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035530591775325977, + "loss": 1.4251, + "theoretical_loss": 3.6852006142737093, + "tokens_seen": 978583552 + }, + { + "epoch": 0.3, + "learning_rate": 0.000355295887662989, + "loss": 1.4616, + "theoretical_loss": 3.685177216274397, + "tokens_seen": 978649088 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035528585757271813, + "loss": 1.424, + "theoretical_loss": 3.6851538202805854, + "tokens_seen": 978714624 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035527582748244737, + "loss": 1.3731, + "theoretical_loss": 3.6851304262919697, + "tokens_seen": 978780160 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003552657973921765, + "loss": 1.5555, + "theoretical_loss": 3.6851070343082433, + "tokens_seen": 978845696 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035525576730190573, + "loss": 1.4321, + "theoretical_loss": 3.6850836443291004, + "tokens_seen": 978911232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003552457372116349, + "loss": 1.2651, + "theoretical_loss": 3.6850602563542347, + "tokens_seen": 978976768 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003552357071213641, + "loss": 1.4747, + "theoretical_loss": 3.6850368703833403, + "tokens_seen": 979042304 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035522567703109327, + "loss": 1.4037, + "theoretical_loss": 3.6850134864161115, + "tokens_seen": 979107840 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035521564694082245, + "loss": 1.4091, + "theoretical_loss": 3.684990104452243, + "tokens_seen": 979173376 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035520561685055163, + "loss": 1.3234, + "theoretical_loss": 3.684966724491428, + "tokens_seen": 979238912 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035519558676028087, + "loss": 1.339, + "theoretical_loss": 3.6849433465333616, + "tokens_seen": 979304448 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035518555667001, + "loss": 1.4201, + "theoretical_loss": 3.6849199705777385, + "tokens_seen": 979369984 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035517552657973923, + "loss": 1.4126, + "theoretical_loss": 3.6848965966242524, + "tokens_seen": 979435520 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003551654964894684, + "loss": 1.4861, + "theoretical_loss": 3.684873224672599, + "tokens_seen": 979501056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003551554663991976, + "loss": 1.5262, + "theoretical_loss": 3.6848498547224717, + "tokens_seen": 979566592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003551454363089268, + "loss": 1.5132, + "theoretical_loss": 3.6848264867735656, + "tokens_seen": 979632128 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035513540621865596, + "loss": 1.5634, + "theoretical_loss": 3.684803120825576, + "tokens_seen": 979697664 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 360128, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.526136875152588, + "objective/train/theoretical_loss": 3.6847797568781973, + "objective/train/tokens_used": 1000223200, + "theoretical_loss": 3.6847797568781973, + "tokens_seen": 979763200 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035512537612838514, + "loss": 1.4635, + "theoretical_loss": 3.6847797568781973, + "tokens_seen": 979763200 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035511534603811437, + "loss": 1.4501, + "theoretical_loss": 3.6847563949311244, + "tokens_seen": 979828736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003551053159478435, + "loss": 1.4047, + "theoretical_loss": 3.684733034984052, + "tokens_seen": 979894272 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035509528585757273, + "loss": 1.4027, + "theoretical_loss": 3.684709677036676, + "tokens_seen": 979959808 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003550852557673019, + "loss": 1.4787, + "theoretical_loss": 3.6846863210886904, + "tokens_seen": 980025344 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003550752256770311, + "loss": 1.4703, + "theoretical_loss": 3.6846629671397912, + "tokens_seen": 980090880 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035506519558676033, + "loss": 1.4305, + "theoretical_loss": 3.684639615189673, + "tokens_seen": 980156416 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035505516549648946, + "loss": 1.3889, + "theoretical_loss": 3.6846162652380325, + "tokens_seen": 980221952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003550451354062187, + "loss": 1.472, + "theoretical_loss": 3.6845929172845633, + "tokens_seen": 980287488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003550351053159478, + "loss": 1.4041, + "theoretical_loss": 3.6845695713289617, + "tokens_seen": 980353024 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035502507522567706, + "loss": 1.4703, + "theoretical_loss": 3.684546227370923, + "tokens_seen": 980418560 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035501504513540624, + "loss": 1.3893, + "theoretical_loss": 3.684522885410143, + "tokens_seen": 980484096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003550050150451354, + "loss": 1.4099, + "theoretical_loss": 3.684499545446317, + "tokens_seen": 980549632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549949849548646, + "loss": 1.3492, + "theoretical_loss": 3.684476207479141, + "tokens_seen": 980615168 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035498495486459384, + "loss": 1.4607, + "theoretical_loss": 3.6844528715083107, + "tokens_seen": 980680704 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035497492477432296, + "loss": 1.4476, + "theoretical_loss": 3.6844295375335223, + "tokens_seen": 980746240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549648946840522, + "loss": 1.4223, + "theoretical_loss": 3.684406205554471, + "tokens_seen": 980811776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549548645937813, + "loss": 1.4165, + "theoretical_loss": 3.6843828755708525, + "tokens_seen": 980877312 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035494483450351056, + "loss": 1.3992, + "theoretical_loss": 3.684359547582364, + "tokens_seen": 980942848 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035493480441323974, + "loss": 1.4817, + "theoretical_loss": 3.684336221588701, + "tokens_seen": 981008384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549247743229689, + "loss": 1.4515, + "theoretical_loss": 3.6843128975895603, + "tokens_seen": 981073920 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549147442326981, + "loss": 1.5134, + "theoretical_loss": 3.6842895755846365, + "tokens_seen": 981139456 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003549047141424273, + "loss": 1.4364, + "theoretical_loss": 3.684266255573627, + "tokens_seen": 981204992 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035489468405215647, + "loss": 1.4609, + "theoretical_loss": 3.684242937556229, + "tokens_seen": 981270528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003548846539618857, + "loss": 1.4666, + "theoretical_loss": 3.6842196215321374, + "tokens_seen": 981336064 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035487462387161483, + "loss": 1.5323, + "theoretical_loss": 3.6841963075010495, + "tokens_seen": 981401600 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035486459378134406, + "loss": 1.3696, + "theoretical_loss": 3.684172995462662, + "tokens_seen": 981467136 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035485456369107324, + "loss": 1.3961, + "theoretical_loss": 3.6841496854166706, + "tokens_seen": 981532672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003548445336008024, + "loss": 1.546, + "theoretical_loss": 3.6841263773627735, + "tokens_seen": 981598208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003548345035105316, + "loss": 1.4741, + "theoretical_loss": 3.684103071300666, + "tokens_seen": 981663744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003548244734202608, + "loss": 1.4276, + "theoretical_loss": 3.6840797672300463, + "tokens_seen": 981729280 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035481444332998997, + "loss": 1.4632, + "theoretical_loss": 3.6840564651506096, + "tokens_seen": 981794816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003548044132397192, + "loss": 1.3786, + "theoretical_loss": 3.684033165062055, + "tokens_seen": 981860352 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035479438314944833, + "loss": 1.4269, + "theoretical_loss": 3.6840098669640775, + "tokens_seen": 981925888 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035478435305917757, + "loss": 1.4517, + "theoretical_loss": 3.6839865708563755, + "tokens_seen": 981991424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003547743229689067, + "loss": 1.4398, + "theoretical_loss": 3.6839632767386457, + "tokens_seen": 982056960 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035476429287863593, + "loss": 1.5153, + "theoretical_loss": 3.6839399846105847, + "tokens_seen": 982122496 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003547542627883651, + "loss": 1.4536, + "theoretical_loss": 3.6839166944718915, + "tokens_seen": 982188032 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003547442326980943, + "loss": 1.4932, + "theoretical_loss": 3.683893406322262, + "tokens_seen": 982253568 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035473420260782347, + "loss": 1.4157, + "theoretical_loss": 3.6838701201613944, + "tokens_seen": 982319104 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035472417251755265, + "loss": 1.4033, + "theoretical_loss": 3.6838468359889855, + "tokens_seen": 982384640 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035471414242728183, + "loss": 1.5031, + "theoretical_loss": 3.683823553804733, + "tokens_seen": 982450176 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035470411233701107, + "loss": 1.4969, + "theoretical_loss": 3.6838002736083353, + "tokens_seen": 982515712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003546940822467402, + "loss": 1.4925, + "theoretical_loss": 3.6837769953994894, + "tokens_seen": 982581248 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035468405215646943, + "loss": 1.5246, + "theoretical_loss": 3.6837537191778935, + "tokens_seen": 982646784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003546740220661986, + "loss": 1.4868, + "theoretical_loss": 3.6837304449432446, + "tokens_seen": 982712320 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003546639919759278, + "loss": 1.4907, + "theoretical_loss": 3.683707172695242, + "tokens_seen": 982777856 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354653961885657, + "loss": 1.4806, + "theoretical_loss": 3.6836839024335823, + "tokens_seen": 982843392 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035464393179538616, + "loss": 1.4135, + "theoretical_loss": 3.683660634157964, + "tokens_seen": 982908928 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035463390170511534, + "loss": 1.377, + "theoretical_loss": 3.6836373678680854, + "tokens_seen": 982974464 + }, + { + "debugging/distinct-1-grams": 0.7488293247094114, + "debugging/entropy-1-grams": 5.349564433264487, + "debugging/length": 462.46153846153845, + "debugging/num_segments": 13, + "debugging/score": 0.0073770168052465215, + "debugging/score_std": 0.0054551366535427, + "epoch": 0.3, + "objective/train/docs_used": 361383, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3466198444366455, + "objective/train/theoretical_loss": 3.6836141035636447, + "objective/train/tokens_used": 1003500000, + "theoretical_loss": 3.6836141035636447, + "tokens_seen": 983040000 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035462387161484457, + "loss": 1.4312, + "theoretical_loss": 3.6836141035636447, + "tokens_seen": 983040000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003546138415245737, + "loss": 1.4563, + "theoretical_loss": 3.6835908412443397, + "tokens_seen": 983105536 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035460381143430293, + "loss": 1.5214, + "theoretical_loss": 3.683567580909869, + "tokens_seen": 983171072 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035459378134403206, + "loss": 1.4388, + "theoretical_loss": 3.683544322559931, + "tokens_seen": 983236608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003545837512537613, + "loss": 1.4071, + "theoretical_loss": 3.683521066194224, + "tokens_seen": 983302144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003545737211634905, + "loss": 1.4178, + "theoretical_loss": 3.6834978118124466, + "tokens_seen": 983367680 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035456369107321966, + "loss": 1.4034, + "theoretical_loss": 3.683474559414297, + "tokens_seen": 983433216 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035455366098294884, + "loss": 1.5274, + "theoretical_loss": 3.6834513089994743, + "tokens_seen": 983498752 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354543630892678, + "loss": 1.4617, + "theoretical_loss": 3.683428060567677, + "tokens_seen": 983564288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003545336008024072, + "loss": 1.4988, + "theoretical_loss": 3.6834048141186035, + "tokens_seen": 983629824 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035452357071213644, + "loss": 1.4856, + "theoretical_loss": 3.683381569651953, + "tokens_seen": 983695360 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035451354062186556, + "loss": 1.437, + "theoretical_loss": 3.683358327167425, + "tokens_seen": 983760896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003545035105315948, + "loss": 1.48, + "theoretical_loss": 3.683335086664717, + "tokens_seen": 983826432 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354493480441324, + "loss": 1.4385, + "theoretical_loss": 3.6833118481435294, + "tokens_seen": 983891968 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035448345035105316, + "loss": 1.2812, + "theoretical_loss": 3.6832886116035604, + "tokens_seen": 983957504 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035447342026078234, + "loss": 1.3823, + "theoretical_loss": 3.6832653770445094, + "tokens_seen": 984023040 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003544633901705115, + "loss": 1.3908, + "theoretical_loss": 3.6832421444660755, + "tokens_seen": 984088576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003544533600802407, + "loss": 1.4291, + "theoretical_loss": 3.6832189138679583, + "tokens_seen": 984154112 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035444332998996994, + "loss": 1.4916, + "theoretical_loss": 3.6831956852498573, + "tokens_seen": 984219648 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035443329989969907, + "loss": 1.382, + "theoretical_loss": 3.6831724586114714, + "tokens_seen": 984285184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003544232698094283, + "loss": 1.4875, + "theoretical_loss": 3.6831492339525, + "tokens_seen": 984350720 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035441323971915743, + "loss": 1.4052, + "theoretical_loss": 3.683126011272643, + "tokens_seen": 984416256 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035440320962888667, + "loss": 1.4576, + "theoretical_loss": 3.6831027905716, + "tokens_seen": 984481792 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035439317953861585, + "loss": 1.4749, + "theoretical_loss": 3.6830795718490705, + "tokens_seen": 984547328 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035438314944834503, + "loss": 1.5715, + "theoretical_loss": 3.683056355104754, + "tokens_seen": 984612864 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003543731193580742, + "loss": 1.389, + "theoretical_loss": 3.683033140338351, + "tokens_seen": 984678400 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035436308926780344, + "loss": 1.5292, + "theoretical_loss": 3.6830099275495614, + "tokens_seen": 984743936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003543530591775326, + "loss": 1.4781, + "theoretical_loss": 3.6829867167380845, + "tokens_seen": 984809472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003543430290872618, + "loss": 1.4616, + "theoretical_loss": 3.6829635079036205, + "tokens_seen": 984875008 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354332998996991, + "loss": 1.386, + "theoretical_loss": 3.682940301045869, + "tokens_seen": 984940544 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035432296890672017, + "loss": 1.5102, + "theoretical_loss": 3.6829170961645312, + "tokens_seen": 985006080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003543129388164494, + "loss": 1.4877, + "theoretical_loss": 3.6828938932593065, + "tokens_seen": 985071616 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035430290872617853, + "loss": 1.4484, + "theoretical_loss": 3.6828706923298955, + "tokens_seen": 985137152 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035429287863590777, + "loss": 1.4919, + "theoretical_loss": 3.6828474933759985, + "tokens_seen": 985202688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003542828485456369, + "loss": 1.4175, + "theoretical_loss": 3.6828242963973157, + "tokens_seen": 985268224 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035427281845536613, + "loss": 1.4525, + "theoretical_loss": 3.6828011013935473, + "tokens_seen": 985333760 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003542627883650953, + "loss": 1.3852, + "theoretical_loss": 3.682777908364395, + "tokens_seen": 985399296 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003542527582748245, + "loss": 1.4986, + "theoretical_loss": 3.682754717309558, + "tokens_seen": 985464832 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035424272818455367, + "loss": 1.4732, + "theoretical_loss": 3.6827315282287376, + "tokens_seen": 985530368 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035423269809428285, + "loss": 1.4202, + "theoretical_loss": 3.682708341121635, + "tokens_seen": 985595904 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035422266800401203, + "loss": 1.5315, + "theoretical_loss": 3.68268515598795, + "tokens_seen": 985661440 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035421263791374127, + "loss": 1.4536, + "theoretical_loss": 3.6826619728273835, + "tokens_seen": 985726976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003542026078234704, + "loss": 1.5138, + "theoretical_loss": 3.6826387916396373, + "tokens_seen": 985792512 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035419257773319963, + "loss": 1.6359, + "theoretical_loss": 3.6826156124244123, + "tokens_seen": 985858048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003541825476429288, + "loss": 1.395, + "theoretical_loss": 3.6825924351814088, + "tokens_seen": 985923584 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354172517552658, + "loss": 1.502, + "theoretical_loss": 3.682569259910328, + "tokens_seen": 985989120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003541624874623872, + "loss": 1.4104, + "theoretical_loss": 3.6825460866108717, + "tokens_seen": 986054656 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035415245737211636, + "loss": 1.3775, + "theoretical_loss": 3.6825229152827403, + "tokens_seen": 986120192 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035414242728184554, + "loss": 1.4138, + "theoretical_loss": 3.682499745925636, + "tokens_seen": 986185728 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035413239719157477, + "loss": 1.3502, + "theoretical_loss": 3.6824765785392595, + "tokens_seen": 986251264 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 362627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4603009223937988, + "objective/train/theoretical_loss": 3.682453413123313, + "objective/train/tokens_used": 1006776800, + "theoretical_loss": 3.682453413123313, + "tokens_seen": 986316800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003541223671013039, + "loss": 1.4308, + "theoretical_loss": 3.682453413123313, + "tokens_seen": 986316800 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035411233701103314, + "loss": 1.4612, + "theoretical_loss": 3.682430249677497, + "tokens_seen": 986382336 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035410230692076226, + "loss": 1.4394, + "theoretical_loss": 3.682407088201514, + "tokens_seen": 986447872 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003540922768304915, + "loss": 1.4366, + "theoretical_loss": 3.6823839286950646, + "tokens_seen": 986513408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003540822467402207, + "loss": 1.3698, + "theoretical_loss": 3.6823607711578514, + "tokens_seen": 986578944 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035407221664994986, + "loss": 1.3528, + "theoretical_loss": 3.682337615589576, + "tokens_seen": 986644480 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035406218655967904, + "loss": 1.3896, + "theoretical_loss": 3.68231446198994, + "tokens_seen": 986710016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003540521564694082, + "loss": 1.4534, + "theoretical_loss": 3.682291310358645, + "tokens_seen": 986775552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003540421263791374, + "loss": 1.5164, + "theoretical_loss": 3.682268160695394, + "tokens_seen": 986841088 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035403209628886664, + "loss": 1.4597, + "theoretical_loss": 3.6822450129998883, + "tokens_seen": 986906624 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035402206619859577, + "loss": 1.4869, + "theoretical_loss": 3.6822218672718297, + "tokens_seen": 986972160 + }, + { + "epoch": 0.3, + "learning_rate": 0.000354012036108325, + "loss": 1.4286, + "theoretical_loss": 3.6821987235109206, + "tokens_seen": 987037696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003540020060180542, + "loss": 1.5196, + "theoretical_loss": 3.6821755817168635, + "tokens_seen": 987103232 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035399197592778336, + "loss": 1.3567, + "theoretical_loss": 3.6821524418893605, + "tokens_seen": 987168768 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035398194583751254, + "loss": 1.3943, + "theoretical_loss": 3.682129304028114, + "tokens_seen": 987234304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003539719157472417, + "loss": 1.5481, + "theoretical_loss": 3.682106168132826, + "tokens_seen": 987299840 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003539618856569709, + "loss": 1.4475, + "theoretical_loss": 3.6820830342032, + "tokens_seen": 987365376 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035395185556670014, + "loss": 1.4163, + "theoretical_loss": 3.6820599022389375, + "tokens_seen": 987430912 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035394182547642927, + "loss": 1.4244, + "theoretical_loss": 3.6820367722397416, + "tokens_seen": 987496448 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003539317953861585, + "loss": 1.5037, + "theoretical_loss": 3.6820136442053144, + "tokens_seen": 987561984 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035392176529588763, + "loss": 1.4459, + "theoretical_loss": 3.6819905181353594, + "tokens_seen": 987627520 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035391173520561687, + "loss": 1.4471, + "theoretical_loss": 3.6819673940295794, + "tokens_seen": 987693056 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035390170511534605, + "loss": 1.419, + "theoretical_loss": 3.6819442718876765, + "tokens_seen": 987758592 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035389167502507523, + "loss": 1.554, + "theoretical_loss": 3.6819211517093544, + "tokens_seen": 987824128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003538816449348044, + "loss": 1.4249, + "theoretical_loss": 3.6818980334943157, + "tokens_seen": 987889664 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035387161484453364, + "loss": 1.4499, + "theoretical_loss": 3.681874917242263, + "tokens_seen": 987955200 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035386158475426277, + "loss": 1.4727, + "theoretical_loss": 3.6818518029529006, + "tokens_seen": 988020736 + }, + { + "epoch": 0.3, + "learning_rate": 0.000353851554663992, + "loss": 1.4795, + "theoretical_loss": 3.681828690625931, + "tokens_seen": 988086272 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035384152457372113, + "loss": 1.4341, + "theoretical_loss": 3.681805580261057, + "tokens_seen": 988151808 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035383149448345037, + "loss": 1.3571, + "theoretical_loss": 3.681782471857983, + "tokens_seen": 988217344 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035382146439317955, + "loss": 1.3514, + "theoretical_loss": 3.6817593654164114, + "tokens_seen": 988282880 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035381143430290873, + "loss": 1.5143, + "theoretical_loss": 3.681736260936046, + "tokens_seen": 988348416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003538014042126379, + "loss": 1.3134, + "theoretical_loss": 3.6817131584165907, + "tokens_seen": 988413952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537913741223671, + "loss": 1.4171, + "theoretical_loss": 3.6816900578577485, + "tokens_seen": 988479488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537813440320963, + "loss": 1.3706, + "theoretical_loss": 3.681666959259223, + "tokens_seen": 988545024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537713139418255, + "loss": 1.4813, + "theoretical_loss": 3.681643862620719, + "tokens_seen": 988610560 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035376128385155464, + "loss": 1.4586, + "theoretical_loss": 3.681620767941939, + "tokens_seen": 988676096 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035375125376128387, + "loss": 1.5137, + "theoretical_loss": 3.6815976752225867, + "tokens_seen": 988741632 + }, + { + "epoch": 0.3, + "learning_rate": 0.000353741223671013, + "loss": 1.5609, + "theoretical_loss": 3.681574584462367, + "tokens_seen": 988807168 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035373119358074223, + "loss": 1.5733, + "theoretical_loss": 3.6815514956609836, + "tokens_seen": 988872704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537211634904714, + "loss": 1.4064, + "theoretical_loss": 3.6815284088181404, + "tokens_seen": 988938240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537111334002006, + "loss": 1.3826, + "theoretical_loss": 3.681505323933541, + "tokens_seen": 989003776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003537011033099298, + "loss": 1.4821, + "theoretical_loss": 3.68148224100689, + "tokens_seen": 989069312 + }, + { + "epoch": 0.3, + "learning_rate": 0.000353691073219659, + "loss": 1.3275, + "theoretical_loss": 3.681459160037892, + "tokens_seen": 989134848 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035368104312938814, + "loss": 1.4949, + "theoretical_loss": 3.6814360810262508, + "tokens_seen": 989200384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536710130391174, + "loss": 1.4549, + "theoretical_loss": 3.681413003971671, + "tokens_seen": 989265920 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536609829488465, + "loss": 1.5786, + "theoretical_loss": 3.6813899288738563, + "tokens_seen": 989331456 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035365095285857574, + "loss": 1.5212, + "theoretical_loss": 3.681366855732512, + "tokens_seen": 989396992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536409227683049, + "loss": 1.4499, + "theoretical_loss": 3.6813437845473427, + "tokens_seen": 989462528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536308926780341, + "loss": 1.4051, + "theoretical_loss": 3.6813207153180523, + "tokens_seen": 989528064 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 363941, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.389134407043457, + "objective/train/theoretical_loss": 3.681297648044346, + "objective/train/tokens_used": 1010053600, + "theoretical_loss": 3.681297648044346, + "tokens_seen": 989593600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536208625877633, + "loss": 1.3767, + "theoretical_loss": 3.681297648044346, + "tokens_seen": 989593600 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035361083249749246, + "loss": 1.379, + "theoretical_loss": 3.681274582725928, + "tokens_seen": 989659136 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003536008024072217, + "loss": 1.3186, + "theoretical_loss": 3.681251519362504, + "tokens_seen": 989724672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003535907723169509, + "loss": 1.3402, + "theoretical_loss": 3.681228457953778, + "tokens_seen": 989790208 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035358074222668006, + "loss": 1.5278, + "theoretical_loss": 3.681205398499456, + "tokens_seen": 989855744 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035357071213640924, + "loss": 1.3266, + "theoretical_loss": 3.681182340999241, + "tokens_seen": 989921280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003535606820461384, + "loss": 1.5176, + "theoretical_loss": 3.68115928545284, + "tokens_seen": 989986816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003535506519558676, + "loss": 1.5466, + "theoretical_loss": 3.6811362318599574, + "tokens_seen": 990052352 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035354062186559684, + "loss": 1.4886, + "theoretical_loss": 3.681113180220298, + "tokens_seen": 990117888 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035353059177532597, + "loss": 1.4644, + "theoretical_loss": 3.681090130533568, + "tokens_seen": 990183424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003535205616850552, + "loss": 1.5087, + "theoretical_loss": 3.6810670827994723, + "tokens_seen": 990248960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003535105315947844, + "loss": 1.4839, + "theoretical_loss": 3.6810440370177155, + "tokens_seen": 990314496 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035350050150451356, + "loss": 1.4745, + "theoretical_loss": 3.681020993188004, + "tokens_seen": 990380032 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035349047141424274, + "loss": 1.4166, + "theoretical_loss": 3.680997951310043, + "tokens_seen": 990445568 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003534804413239719, + "loss": 1.4422, + "theoretical_loss": 3.6809749113835375, + "tokens_seen": 990511104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003534704112337011, + "loss": 1.5408, + "theoretical_loss": 3.680951873408194, + "tokens_seen": 990576640 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035346038114343034, + "loss": 1.3984, + "theoretical_loss": 3.680928837383718, + "tokens_seen": 990642176 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035345035105315947, + "loss": 1.3514, + "theoretical_loss": 3.680905803309815, + "tokens_seen": 990707712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003534403209628887, + "loss": 1.4388, + "theoretical_loss": 3.680882771186191, + "tokens_seen": 990773248 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035343029087261783, + "loss": 1.4095, + "theoretical_loss": 3.6808597410125516, + "tokens_seen": 990838784 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035342026078234707, + "loss": 1.4842, + "theoretical_loss": 3.680836712788603, + "tokens_seen": 990904320 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035341023069207625, + "loss": 1.4173, + "theoretical_loss": 3.6808136865140515, + "tokens_seen": 990969856 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035340020060180543, + "loss": 1.3716, + "theoretical_loss": 3.6807906621886026, + "tokens_seen": 991035392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003533901705115346, + "loss": 1.5284, + "theoretical_loss": 3.6807676398119624, + "tokens_seen": 991100928 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035338014042126384, + "loss": 1.3797, + "theoretical_loss": 3.6807446193838373, + "tokens_seen": 991166464 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035337011033099297, + "loss": 1.3813, + "theoretical_loss": 3.6807216009039343, + "tokens_seen": 991232000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003533600802407222, + "loss": 1.7834, + "theoretical_loss": 3.680698584371959, + "tokens_seen": 991297536 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035335005015045133, + "loss": 1.3518, + "theoretical_loss": 3.680675569787617, + "tokens_seen": 991363072 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035334002006018057, + "loss": 1.5138, + "theoretical_loss": 3.6806525571506166, + "tokens_seen": 991428608 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035332998996990975, + "loss": 1.4987, + "theoretical_loss": 3.680629546460663, + "tokens_seen": 991494144 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035331995987963893, + "loss": 1.4053, + "theoretical_loss": 3.680606537717463, + "tokens_seen": 991559680 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003533099297893681, + "loss": 1.4414, + "theoretical_loss": 3.6805835309207238, + "tokens_seen": 991625216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532998996990973, + "loss": 1.436, + "theoretical_loss": 3.680560526070151, + "tokens_seen": 991690752 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532898696088265, + "loss": 1.4416, + "theoretical_loss": 3.6805375231654525, + "tokens_seen": 991756288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532798395185557, + "loss": 1.5542, + "theoretical_loss": 3.6805145222063347, + "tokens_seen": 991821824 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035326980942828484, + "loss": 1.5084, + "theoretical_loss": 3.6804915231925044, + "tokens_seen": 991887360 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035325977933801407, + "loss": 1.633, + "theoretical_loss": 3.680468526123669, + "tokens_seen": 991952896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532497492477432, + "loss": 1.5314, + "theoretical_loss": 3.6804455309995348, + "tokens_seen": 992018432 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035323971915747243, + "loss": 1.5619, + "theoretical_loss": 3.6804225378198097, + "tokens_seen": 992083968 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532296890672016, + "loss": 1.3748, + "theoretical_loss": 3.6803995465842, + "tokens_seen": 992149504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003532196589769308, + "loss": 1.415, + "theoretical_loss": 3.680376557292414, + "tokens_seen": 992215040 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035320962888666, + "loss": 1.5022, + "theoretical_loss": 3.680353569944158, + "tokens_seen": 992280576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531995987963892, + "loss": 1.3905, + "theoretical_loss": 3.6803305845391394, + "tokens_seen": 992346112 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035318956870611834, + "loss": 1.4972, + "theoretical_loss": 3.680307601077066, + "tokens_seen": 992411648 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531795386158476, + "loss": 1.3812, + "theoretical_loss": 3.6802846195576455, + "tokens_seen": 992477184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531695085255767, + "loss": 1.4592, + "theoretical_loss": 3.6802616399805848, + "tokens_seen": 992542720 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035315947843530594, + "loss": 1.4744, + "theoretical_loss": 3.680238662345592, + "tokens_seen": 992608256 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531494483450351, + "loss": 1.5432, + "theoretical_loss": 3.6802156866523745, + "tokens_seen": 992673792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531394182547643, + "loss": 1.4659, + "theoretical_loss": 3.6801927129006407, + "tokens_seen": 992739328 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003531293881644935, + "loss": 1.3977, + "theoretical_loss": 3.680169741090097, + "tokens_seen": 992804864 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 365105, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4565749168395996, + "objective/train/theoretical_loss": 3.6801467712204525, + "objective/train/tokens_used": 1013330400, + "theoretical_loss": 3.6801467712204525, + "tokens_seen": 992870400 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035311935807422266, + "loss": 1.4048, + "theoretical_loss": 3.6801467712204525, + "tokens_seen": 992870400 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035310932798395184, + "loss": 1.5382, + "theoretical_loss": 3.6801238032914148, + "tokens_seen": 992935936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003530992978936811, + "loss": 1.538, + "theoretical_loss": 3.6801008373026916, + "tokens_seen": 993001472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003530892678034102, + "loss": 1.5549, + "theoretical_loss": 3.6800778732539907, + "tokens_seen": 993067008 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035307923771313944, + "loss": 1.4661, + "theoretical_loss": 3.680054911145021, + "tokens_seen": 993132544 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035306920762286857, + "loss": 1.3552, + "theoretical_loss": 3.68003195097549, + "tokens_seen": 993198080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003530591775325978, + "loss": 1.421, + "theoretical_loss": 3.6800089927451065, + "tokens_seen": 993263616 + }, + { + "epoch": 0.3, + "learning_rate": 0.000353049147442327, + "loss": 1.473, + "theoretical_loss": 3.6799860364535784, + "tokens_seen": 993329152 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035303911735205617, + "loss": 1.4878, + "theoretical_loss": 3.679963082100614, + "tokens_seen": 993394688 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035302908726178535, + "loss": 1.4059, + "theoretical_loss": 3.6799401296859227, + "tokens_seen": 993460224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003530190571715146, + "loss": 1.3921, + "theoretical_loss": 3.6799171792092116, + "tokens_seen": 993525760 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003530090270812437, + "loss": 1.2774, + "theoretical_loss": 3.6798942306701896, + "tokens_seen": 993591296 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035299899699097294, + "loss": 1.3999, + "theoretical_loss": 3.6798712840685655, + "tokens_seen": 993656832 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035298896690070207, + "loss": 1.4492, + "theoretical_loss": 3.679848339404048, + "tokens_seen": 993722368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003529789368104313, + "loss": 1.5039, + "theoretical_loss": 3.679825396676346, + "tokens_seen": 993787904 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003529689067201605, + "loss": 1.3637, + "theoretical_loss": 3.6798024558851683, + "tokens_seen": 993853440 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035295887662988967, + "loss": 1.4419, + "theoretical_loss": 3.679779517030224, + "tokens_seen": 993918976 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035294884653961885, + "loss": 1.4047, + "theoretical_loss": 3.6797565801112206, + "tokens_seen": 993984512 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035293881644934803, + "loss": 1.5407, + "theoretical_loss": 3.679733645127869, + "tokens_seen": 994050048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003529287863590772, + "loss": 1.3946, + "theoretical_loss": 3.679710712079877, + "tokens_seen": 994115584 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035291875626880645, + "loss": 1.4405, + "theoretical_loss": 3.679687780966954, + "tokens_seen": 994181120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003529087261785356, + "loss": 1.3107, + "theoretical_loss": 3.679664851788809, + "tokens_seen": 994246656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003528986960882648, + "loss": 1.3621, + "theoretical_loss": 3.6796419245451526, + "tokens_seen": 994312192 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035288866599799394, + "loss": 1.5294, + "theoretical_loss": 3.679618999235692, + "tokens_seen": 994377728 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035287863590772317, + "loss": 1.4976, + "theoretical_loss": 3.679596075860138, + "tokens_seen": 994443264 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035286860581745235, + "loss": 1.5345, + "theoretical_loss": 3.6795731544181995, + "tokens_seen": 994508800 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035285857572718153, + "loss": 1.3841, + "theoretical_loss": 3.679550234909586, + "tokens_seen": 994574336 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035284854563691077, + "loss": 1.4815, + "theoretical_loss": 3.6795273173340073, + "tokens_seen": 994639872 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035283851554663995, + "loss": 1.4586, + "theoretical_loss": 3.6795044016911724, + "tokens_seen": 994705408 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035282848545636913, + "loss": 1.4184, + "theoretical_loss": 3.679481487980792, + "tokens_seen": 994770944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003528184553660983, + "loss": 1.4291, + "theoretical_loss": 3.6794585762025753, + "tokens_seen": 994836480 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003528084252758275, + "loss": 1.4718, + "theoretical_loss": 3.679435666356232, + "tokens_seen": 994902016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527983951855567, + "loss": 1.5213, + "theoretical_loss": 3.679412758441472, + "tokens_seen": 994967552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527883650952859, + "loss": 1.4254, + "theoretical_loss": 3.679389852458005, + "tokens_seen": 995033088 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035277833500501504, + "loss": 1.4381, + "theoretical_loss": 3.6793669484055416, + "tokens_seen": 995098624 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035276830491474427, + "loss": 1.4852, + "theoretical_loss": 3.6793440462837923, + "tokens_seen": 995164160 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527582748244734, + "loss": 1.4318, + "theoretical_loss": 3.6793211460924655, + "tokens_seen": 995229696 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035274824473420263, + "loss": 1.4402, + "theoretical_loss": 3.679298247831272, + "tokens_seen": 995295232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527382146439318, + "loss": 1.4356, + "theoretical_loss": 3.679275351499924, + "tokens_seen": 995360768 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352728184553661, + "loss": 1.3998, + "theoretical_loss": 3.679252457098129, + "tokens_seen": 995426304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527181544633902, + "loss": 1.4865, + "theoretical_loss": 3.6792295646255986, + "tokens_seen": 995491840 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003527081243731194, + "loss": 1.5625, + "theoretical_loss": 3.6792066740820433, + "tokens_seen": 995557376 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035269809428284854, + "loss": 1.5029, + "theoretical_loss": 3.679183785467173, + "tokens_seen": 995622912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526880641925778, + "loss": 1.4772, + "theoretical_loss": 3.6791608987807, + "tokens_seen": 995688448 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526780341023069, + "loss": 1.4727, + "theoretical_loss": 3.6791380140223326, + "tokens_seen": 995753984 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035266800401203614, + "loss": 1.5863, + "theoretical_loss": 3.679115131191783, + "tokens_seen": 995819520 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526579739217653, + "loss": 1.394, + "theoretical_loss": 3.6790922502887615, + "tokens_seen": 995885056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526479438314945, + "loss": 1.3305, + "theoretical_loss": 3.6790693713129787, + "tokens_seen": 995950592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526379137412237, + "loss": 1.4746, + "theoretical_loss": 3.679046494264146, + "tokens_seen": 996016128 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035262788365095286, + "loss": 1.5114, + "theoretical_loss": 3.679023619141974, + "tokens_seen": 996081664 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 366329, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.550077199935913, + "objective/train/theoretical_loss": 3.6790007459461735, + "objective/train/tokens_used": 1016607200, + "theoretical_loss": 3.6790007459461735, + "tokens_seen": 996147200 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035261785356068204, + "loss": 1.4266, + "theoretical_loss": 3.6790007459461735, + "tokens_seen": 996147200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003526078234704113, + "loss": 1.4743, + "theoretical_loss": 3.678977874676456, + "tokens_seen": 996212736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003525977933801404, + "loss": 1.5531, + "theoretical_loss": 3.678955005332532, + "tokens_seen": 996278272 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035258776328986964, + "loss": 1.252, + "theoretical_loss": 3.6789321379141136, + "tokens_seen": 996343808 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035257773319959877, + "loss": 1.4793, + "theoretical_loss": 3.6789092724209107, + "tokens_seen": 996409344 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352567703109328, + "loss": 1.3224, + "theoretical_loss": 3.6788864088526356, + "tokens_seen": 996474880 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003525576730190572, + "loss": 1.3725, + "theoretical_loss": 3.678863547209, + "tokens_seen": 996540416 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035254764292878637, + "loss": 1.4499, + "theoretical_loss": 3.6788406874897146, + "tokens_seen": 996605952 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035253761283851555, + "loss": 1.5127, + "theoretical_loss": 3.6788178296944913, + "tokens_seen": 996671488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003525275827482448, + "loss": 1.5487, + "theoretical_loss": 3.6787949738230408, + "tokens_seen": 996737024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003525175526579739, + "loss": 1.5075, + "theoretical_loss": 3.678772119875076, + "tokens_seen": 996802560 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035250752256770314, + "loss": 1.3767, + "theoretical_loss": 3.678749267850308, + "tokens_seen": 996868096 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035249749247743227, + "loss": 1.5456, + "theoretical_loss": 3.678726417748448, + "tokens_seen": 996933632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003524874623871615, + "loss": 1.4243, + "theoretical_loss": 3.6787035695692083, + "tokens_seen": 996999168 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003524774322968907, + "loss": 1.4841, + "theoretical_loss": 3.678680723312301, + "tokens_seen": 997064704 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035246740220661987, + "loss": 1.4258, + "theoretical_loss": 3.678657878977438, + "tokens_seen": 997130240 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035245737211634905, + "loss": 1.3447, + "theoretical_loss": 3.6786350365643306, + "tokens_seen": 997195776 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035244734202607823, + "loss": 1.4506, + "theoretical_loss": 3.6786121960726916, + "tokens_seen": 997261312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003524373119358074, + "loss": 1.4588, + "theoretical_loss": 3.678589357502233, + "tokens_seen": 997326848 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035242728184553665, + "loss": 1.4306, + "theoretical_loss": 3.6785665208526668, + "tokens_seen": 997392384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003524172517552658, + "loss": 1.3479, + "theoretical_loss": 3.678543686123705, + "tokens_seen": 997457920 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352407221664995, + "loss": 1.461, + "theoretical_loss": 3.67852085331506, + "tokens_seen": 997523456 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035239719157472414, + "loss": 1.4827, + "theoretical_loss": 3.6784980224264445, + "tokens_seen": 997588992 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035238716148445337, + "loss": 1.4013, + "theoretical_loss": 3.6784751934575706, + "tokens_seen": 997654528 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035237713139418255, + "loss": 1.508, + "theoretical_loss": 3.6784523664081514, + "tokens_seen": 997720064 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035236710130391173, + "loss": 1.5471, + "theoretical_loss": 3.6784295412778985, + "tokens_seen": 997785600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003523570712136409, + "loss": 1.4689, + "theoretical_loss": 3.678406718066525, + "tokens_seen": 997851136 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035234704112337015, + "loss": 1.5526, + "theoretical_loss": 3.6783838967737443, + "tokens_seen": 997916672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003523370110330993, + "loss": 1.4633, + "theoretical_loss": 3.6783610773992677, + "tokens_seen": 997982208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003523269809428285, + "loss": 1.5866, + "theoretical_loss": 3.6783382599428087, + "tokens_seen": 998047744 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035231695085255764, + "loss": 1.3077, + "theoretical_loss": 3.67831544440408, + "tokens_seen": 998113280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003523069207622869, + "loss": 1.4567, + "theoretical_loss": 3.6782926307827952, + "tokens_seen": 998178816 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035229689067201606, + "loss": 1.5338, + "theoretical_loss": 3.6782698190786665, + "tokens_seen": 998244352 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035228686058174524, + "loss": 1.5405, + "theoretical_loss": 3.6782470092914075, + "tokens_seen": 998309888 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003522768304914744, + "loss": 1.5172, + "theoretical_loss": 3.67822420142073, + "tokens_seen": 998375424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003522668004012036, + "loss": 1.577, + "theoretical_loss": 3.678201395466349, + "tokens_seen": 998440960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003522567703109328, + "loss": 1.4943, + "theoretical_loss": 3.6781785914279768, + "tokens_seen": 998506496 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352246740220662, + "loss": 1.4724, + "theoretical_loss": 3.6781557893053267, + "tokens_seen": 998572032 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035223671013039114, + "loss": 1.4124, + "theoretical_loss": 3.6781329890981125, + "tokens_seen": 998637568 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003522266800401204, + "loss": 1.34, + "theoretical_loss": 3.6781101908060467, + "tokens_seen": 998703104 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035221664994984956, + "loss": 1.4689, + "theoretical_loss": 3.678087394428843, + "tokens_seen": 998768640 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035220661985957874, + "loss": 1.4186, + "theoretical_loss": 3.6780645999662163, + "tokens_seen": 998834176 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521965897693079, + "loss": 1.613, + "theoretical_loss": 3.6780418074178782, + "tokens_seen": 998899712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521865596790371, + "loss": 1.4776, + "theoretical_loss": 3.678019016783544, + "tokens_seen": 998965248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521765295887663, + "loss": 1.5578, + "theoretical_loss": 3.677996228062926, + "tokens_seen": 999030784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521664994984955, + "loss": 1.4376, + "theoretical_loss": 3.677973441255739, + "tokens_seen": 999096320 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035215646940822465, + "loss": 1.4802, + "theoretical_loss": 3.6779506563616966, + "tokens_seen": 999161856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521464393179539, + "loss": 1.4398, + "theoretical_loss": 3.677927873380513, + "tokens_seen": 999227392 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352136409227683, + "loss": 1.4569, + "theoretical_loss": 3.6779050923119017, + "tokens_seen": 999292928 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035212637913741224, + "loss": 1.4683, + "theoretical_loss": 3.6778823131555765, + "tokens_seen": 999358464 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 367579, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6011443138122559, + "objective/train/theoretical_loss": 3.677859535911252, + "objective/train/tokens_used": 1019884000, + "theoretical_loss": 3.677859535911252, + "tokens_seen": 999424000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521163490471414, + "loss": 1.5106, + "theoretical_loss": 3.677859535911252, + "tokens_seen": 999424000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003521063189568706, + "loss": 1.39, + "theoretical_loss": 3.677836760578643, + "tokens_seen": 999489536 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035209628886659984, + "loss": 1.4714, + "theoretical_loss": 3.677813987157462, + "tokens_seen": 999555072 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035208625877632897, + "loss": 1.4354, + "theoretical_loss": 3.6777912156474244, + "tokens_seen": 999620608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003520762286860582, + "loss": 1.4221, + "theoretical_loss": 3.677768446048244, + "tokens_seen": 999686144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003520661985957874, + "loss": 1.3934, + "theoretical_loss": 3.677745678359636, + "tokens_seen": 999751680 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035205616850551657, + "loss": 1.4202, + "theoretical_loss": 3.6777229125813142, + "tokens_seen": 999817216 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035204613841524575, + "loss": 1.4027, + "theoretical_loss": 3.6777001487129937, + "tokens_seen": 999882752 + }, + { + "epoch": 0.3, + "learning_rate": 0.000352036108324975, + "loss": 1.5443, + "theoretical_loss": 3.677677386754388, + "tokens_seen": 999948288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003520260782347041, + "loss": 1.4259, + "theoretical_loss": 3.677654626705213, + "tokens_seen": 1000013824 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035201604814443334, + "loss": 1.4712, + "theoretical_loss": 3.6776318685651823, + "tokens_seen": 1000079360 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035200601805416247, + "loss": 1.4619, + "theoretical_loss": 3.6776091123340118, + "tokens_seen": 1000144896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003519959879638917, + "loss": 1.4493, + "theoretical_loss": 3.6775863580114154, + "tokens_seen": 1000210432 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003519859578736209, + "loss": 1.5499, + "theoretical_loss": 3.6775636055971086, + "tokens_seen": 1000275968 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035197592778335007, + "loss": 1.4278, + "theoretical_loss": 3.6775408550908057, + "tokens_seen": 1000341504 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035196589769307925, + "loss": 1.289, + "theoretical_loss": 3.677518106492222, + "tokens_seen": 1000407040 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035195586760280843, + "loss": 1.528, + "theoretical_loss": 3.6774953598010733, + "tokens_seen": 1000472576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003519458375125376, + "loss": 1.4246, + "theoretical_loss": 3.6774726150170736, + "tokens_seen": 1000538112 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035193580742226685, + "loss": 1.5043, + "theoretical_loss": 3.6774498721399387, + "tokens_seen": 1000603648 + }, + { + "epoch": 0.3, + "learning_rate": 0.000351925777331996, + "loss": 1.5268, + "theoretical_loss": 3.6774271311693836, + "tokens_seen": 1000669184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003519157472417252, + "loss": 1.4814, + "theoretical_loss": 3.6774043921051236, + "tokens_seen": 1000734720 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035190571715145434, + "loss": 1.3731, + "theoretical_loss": 3.677381654946875, + "tokens_seen": 1000800256 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035189568706118357, + "loss": 1.4457, + "theoretical_loss": 3.6773589196943517, + "tokens_seen": 1000865792 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035188565697091275, + "loss": 1.3861, + "theoretical_loss": 3.6773361863472704, + "tokens_seen": 1000931328 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035187562688064193, + "loss": 1.5023, + "theoretical_loss": 3.677313454905346, + "tokens_seen": 1000996864 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003518655967903711, + "loss": 1.4259, + "theoretical_loss": 3.6772907253682945, + "tokens_seen": 1001062400 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035185556670010035, + "loss": 1.3487, + "theoretical_loss": 3.6772679977358314, + "tokens_seen": 1001127936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003518455366098295, + "loss": 1.3376, + "theoretical_loss": 3.6772452720076725, + "tokens_seen": 1001193472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003518355065195587, + "loss": 1.3061, + "theoretical_loss": 3.6772225481835337, + "tokens_seen": 1001259008 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035182547642928784, + "loss": 1.4127, + "theoretical_loss": 3.6771998262631307, + "tokens_seen": 1001324544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003518154463390171, + "loss": 1.4009, + "theoretical_loss": 3.6771771062461793, + "tokens_seen": 1001390080 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035180541624874626, + "loss": 1.4313, + "theoretical_loss": 3.6771543881323963, + "tokens_seen": 1001455616 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035179538615847544, + "loss": 1.5448, + "theoretical_loss": 3.6771316719214964, + "tokens_seen": 1001521152 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003517853560682046, + "loss": 1.4116, + "theoretical_loss": 3.677108957613197, + "tokens_seen": 1001586688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003517753259779338, + "loss": 1.3685, + "theoretical_loss": 3.6770862452072133, + "tokens_seen": 1001652224 + }, + { + "epoch": 0.3, + "learning_rate": 0.000351765295887663, + "loss": 1.4394, + "theoretical_loss": 3.6770635347032616, + "tokens_seen": 1001717760 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003517552657973922, + "loss": 1.4651, + "theoretical_loss": 3.6770408261010594, + "tokens_seen": 1001783296 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035174523570712134, + "loss": 1.364, + "theoretical_loss": 3.677018119400322, + "tokens_seen": 1001848832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003517352056168506, + "loss": 1.4255, + "theoretical_loss": 3.6769954146007655, + "tokens_seen": 1001914368 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035172517552657976, + "loss": 1.4617, + "theoretical_loss": 3.676972711702107, + "tokens_seen": 1001979904 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035171514543630894, + "loss": 1.3918, + "theoretical_loss": 3.676950010704063, + "tokens_seen": 1002045440 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003517051153460381, + "loss": 1.466, + "theoretical_loss": 3.6769273116063506, + "tokens_seen": 1002110976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516950852557673, + "loss": 1.552, + "theoretical_loss": 3.676904614408685, + "tokens_seen": 1002176512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516850551654965, + "loss": 1.425, + "theoretical_loss": 3.6768819191107847, + "tokens_seen": 1002242048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516750250752257, + "loss": 1.4672, + "theoretical_loss": 3.6768592257123647, + "tokens_seen": 1002307584 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035166499498495485, + "loss": 1.5729, + "theoretical_loss": 3.6768365342131437, + "tokens_seen": 1002373120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516549648946841, + "loss": 1.4336, + "theoretical_loss": 3.6768138446128367, + "tokens_seen": 1002438656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516449348044132, + "loss": 1.367, + "theoretical_loss": 3.676791156911162, + "tokens_seen": 1002504192 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035163490471414244, + "loss": 1.5328, + "theoretical_loss": 3.6767684711078363, + "tokens_seen": 1002569728 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516248746238716, + "loss": 1.3804, + "theoretical_loss": 3.6767457872025764, + "tokens_seen": 1002635264 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 368716, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5347790718078613, + "objective/train/theoretical_loss": 3.6767231051950997, + "objective/train/tokens_used": 1023160800, + "theoretical_loss": 3.6767231051950997, + "tokens_seen": 1002700800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003516148445336008, + "loss": 1.4908, + "theoretical_loss": 3.6767231051950997, + "tokens_seen": 1002700800 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035160481444333, + "loss": 1.4552, + "theoretical_loss": 3.676700425085123, + "tokens_seen": 1002766336 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035159478435305917, + "loss": 1.3196, + "theoretical_loss": 3.6766777468723646, + "tokens_seen": 1002831872 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035158475426278835, + "loss": 1.4355, + "theoretical_loss": 3.6766550705565404, + "tokens_seen": 1002897408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003515747241725176, + "loss": 1.4515, + "theoretical_loss": 3.676632396137369, + "tokens_seen": 1002962944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003515646940822467, + "loss": 1.4206, + "theoretical_loss": 3.676609723614567, + "tokens_seen": 1003028480 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035155466399197595, + "loss": 1.3835, + "theoretical_loss": 3.6765870529878524, + "tokens_seen": 1003094016 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035154463390170513, + "loss": 1.3629, + "theoretical_loss": 3.6765643842569427, + "tokens_seen": 1003159552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003515346038114343, + "loss": 1.4389, + "theoretical_loss": 3.6765417174215553, + "tokens_seen": 1003225088 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003515245737211635, + "loss": 1.522, + "theoretical_loss": 3.6765190524814084, + "tokens_seen": 1003290624 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035151454363089267, + "loss": 1.4985, + "theoretical_loss": 3.6764963894362195, + "tokens_seen": 1003356160 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035150451354062185, + "loss": 1.4782, + "theoretical_loss": 3.676473728285706, + "tokens_seen": 1003421696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514944834503511, + "loss": 1.3838, + "theoretical_loss": 3.676451069029586, + "tokens_seen": 1003487232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514844533600802, + "loss": 1.5199, + "theoretical_loss": 3.676428411667578, + "tokens_seen": 1003552768 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035147442326980945, + "loss": 1.5046, + "theoretical_loss": 3.676405756199399, + "tokens_seen": 1003618304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514643931795386, + "loss": 1.5432, + "theoretical_loss": 3.6763831026247678, + "tokens_seen": 1003683840 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514543630892678, + "loss": 1.5494, + "theoretical_loss": 3.6763604509434025, + "tokens_seen": 1003749376 + }, + { + "epoch": 0.3, + "learning_rate": 0.000351444332998997, + "loss": 1.4509, + "theoretical_loss": 3.6763378011550207, + "tokens_seen": 1003814912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514343029087262, + "loss": 1.4842, + "theoretical_loss": 3.676315153259341, + "tokens_seen": 1003880448 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035142427281845536, + "loss": 1.5108, + "theoretical_loss": 3.6762925072560817, + "tokens_seen": 1003945984 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035141424272818454, + "loss": 1.4409, + "theoretical_loss": 3.6762698631449613, + "tokens_seen": 1004011520 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003514042126379137, + "loss": 1.3522, + "theoretical_loss": 3.676247220925698, + "tokens_seen": 1004077056 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035139418254764295, + "loss": 1.3614, + "theoretical_loss": 3.67622458059801, + "tokens_seen": 1004142592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513841524573721, + "loss": 1.5043, + "theoretical_loss": 3.6762019421616166, + "tokens_seen": 1004208128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513741223671013, + "loss": 1.5364, + "theoretical_loss": 3.6761793056162357, + "tokens_seen": 1004273664 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513640922768305, + "loss": 1.3099, + "theoretical_loss": 3.676156670961586, + "tokens_seen": 1004339200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513540621865597, + "loss": 1.4993, + "theoretical_loss": 3.676134038197387, + "tokens_seen": 1004404736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513440320962889, + "loss": 1.4487, + "theoretical_loss": 3.6761114073233565, + "tokens_seen": 1004470272 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035133400200601804, + "loss": 1.5669, + "theoretical_loss": 3.6760887783392135, + "tokens_seen": 1004535808 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003513239719157473, + "loss": 1.4401, + "theoretical_loss": 3.6760661512446777, + "tokens_seen": 1004601344 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035131394182547646, + "loss": 1.3722, + "theoretical_loss": 3.676043526039467, + "tokens_seen": 1004666880 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035130391173520564, + "loss": 1.3655, + "theoretical_loss": 3.6760209027233013, + "tokens_seen": 1004732416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512938816449348, + "loss": 1.3582, + "theoretical_loss": 3.6759982812958993, + "tokens_seen": 1004797952 + }, + { + "epoch": 0.3, + "learning_rate": 0.000351283851554664, + "loss": 1.3907, + "theoretical_loss": 3.67597566175698, + "tokens_seen": 1004863488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512738214643932, + "loss": 1.5831, + "theoretical_loss": 3.6759530441062624, + "tokens_seen": 1004929024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512637913741224, + "loss": 1.3889, + "theoretical_loss": 3.675930428343466, + "tokens_seen": 1004994560 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035125376128385154, + "loss": 1.4396, + "theoretical_loss": 3.6759078144683106, + "tokens_seen": 1005060096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512437311935808, + "loss": 1.4276, + "theoretical_loss": 3.675885202480515, + "tokens_seen": 1005125632 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035123370110330996, + "loss": 1.4493, + "theoretical_loss": 3.675862592379799, + "tokens_seen": 1005191168 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035122367101303914, + "loss": 1.457, + "theoretical_loss": 3.6758399841658815, + "tokens_seen": 1005256704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512136409227683, + "loss": 1.4716, + "theoretical_loss": 3.6758173778384826, + "tokens_seen": 1005322240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003512036108324975, + "loss": 1.5552, + "theoretical_loss": 3.675794773397322, + "tokens_seen": 1005387776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511935807422267, + "loss": 1.4509, + "theoretical_loss": 3.675772170842119, + "tokens_seen": 1005453312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511835506519559, + "loss": 1.4375, + "theoretical_loss": 3.6757495701725933, + "tokens_seen": 1005518848 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035117352056168505, + "loss": 1.4534, + "theoretical_loss": 3.675726971388465, + "tokens_seen": 1005584384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511634904714143, + "loss": 1.6032, + "theoretical_loss": 3.6757043744894538, + "tokens_seen": 1005649920 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511534603811434, + "loss": 1.404, + "theoretical_loss": 3.67568177947528, + "tokens_seen": 1005715456 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035114343029087264, + "loss": 1.5206, + "theoretical_loss": 3.675659186345663, + "tokens_seen": 1005780992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511334002006018, + "loss": 1.3666, + "theoretical_loss": 3.6756365951003227, + "tokens_seen": 1005846528 + }, + { + "epoch": 0.3, + "learning_rate": 0.000351123370110331, + "loss": 1.4626, + "theoretical_loss": 3.6756140057389795, + "tokens_seen": 1005912064 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 369909, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.582378625869751, + "objective/train/theoretical_loss": 3.6755914182613543, + "objective/train/tokens_used": 1026437600, + "theoretical_loss": 3.6755914182613543, + "tokens_seen": 1005977600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003511133400200602, + "loss": 1.5704, + "theoretical_loss": 3.6755914182613543, + "tokens_seen": 1005977600 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035110330992978937, + "loss": 1.4892, + "theoretical_loss": 3.675568832667166, + "tokens_seen": 1006043136 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035109327983951855, + "loss": 1.4781, + "theoretical_loss": 3.6755462489561355, + "tokens_seen": 1006108672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003510832497492478, + "loss": 1.4709, + "theoretical_loss": 3.6755236671279836, + "tokens_seen": 1006174208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003510732196589769, + "loss": 1.3948, + "theoretical_loss": 3.67550108718243, + "tokens_seen": 1006239744 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035106318956870615, + "loss": 1.422, + "theoretical_loss": 3.6754785091191957, + "tokens_seen": 1006305280 + }, + { + "epoch": 0.3, + "learning_rate": 0.00035105315947843533, + "loss": 1.5089, + "theoretical_loss": 3.675455932938, + "tokens_seen": 1006370816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0003510431293881645, + "loss": 1.4229, + "theoretical_loss": 3.675433358638565, + "tokens_seen": 1006436352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003510330992978937, + "loss": 1.4174, + "theoretical_loss": 3.6754107862206107, + "tokens_seen": 1006501888 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035102306920762287, + "loss": 1.3858, + "theoretical_loss": 3.6753882156838578, + "tokens_seen": 1006567424 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035101303911735205, + "loss": 1.4922, + "theoretical_loss": 3.675365647028027, + "tokens_seen": 1006632960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003510030090270813, + "loss": 1.3466, + "theoretical_loss": 3.675343080252839, + "tokens_seen": 1006698496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003509929789368104, + "loss": 1.5046, + "theoretical_loss": 3.675320515358015, + "tokens_seen": 1006764032 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035098294884653965, + "loss": 1.5243, + "theoretical_loss": 3.675297952343276, + "tokens_seen": 1006829568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003509729187562688, + "loss": 1.4512, + "theoretical_loss": 3.6752753912083422, + "tokens_seen": 1006895104 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350962888665998, + "loss": 1.4271, + "theoretical_loss": 3.6752528319529354, + "tokens_seen": 1006960640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003509528585757272, + "loss": 1.4946, + "theoretical_loss": 3.675230274576777, + "tokens_seen": 1007026176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003509428284854564, + "loss": 1.4181, + "theoretical_loss": 3.6752077190795873, + "tokens_seen": 1007091712 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035093279839518556, + "loss": 1.4587, + "theoretical_loss": 3.6751851654610874, + "tokens_seen": 1007157248 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035092276830491474, + "loss": 1.396, + "theoretical_loss": 3.6751626137209996, + "tokens_seen": 1007222784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003509127382146439, + "loss": 1.5448, + "theoretical_loss": 3.675140063859045, + "tokens_seen": 1007288320 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035090270812437315, + "loss": 1.5151, + "theoretical_loss": 3.6751175158749447, + "tokens_seen": 1007353856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508926780341023, + "loss": 1.4495, + "theoretical_loss": 3.67509496976842, + "tokens_seen": 1007419392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508826479438315, + "loss": 1.4041, + "theoretical_loss": 3.6750724255391924, + "tokens_seen": 1007484928 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508726178535607, + "loss": 1.6128, + "theoretical_loss": 3.6750498831869844, + "tokens_seen": 1007550464 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508625877632899, + "loss": 1.5006, + "theoretical_loss": 3.6750273427115165, + "tokens_seen": 1007616000 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035085255767301906, + "loss": 1.4953, + "theoretical_loss": 3.675004804112511, + "tokens_seen": 1007681536 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035084252758274824, + "loss": 1.5349, + "theoretical_loss": 3.67498226738969, + "tokens_seen": 1007747072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508324974924774, + "loss": 1.4563, + "theoretical_loss": 3.674959732542774, + "tokens_seen": 1007812608 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035082246740220666, + "loss": 1.4546, + "theoretical_loss": 3.6749371995714863, + "tokens_seen": 1007878144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003508124373119358, + "loss": 1.4221, + "theoretical_loss": 3.674914668475548, + "tokens_seen": 1007943680 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350802407221665, + "loss": 1.3967, + "theoretical_loss": 3.674892139254682, + "tokens_seen": 1008009216 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035079237713139415, + "loss": 1.568, + "theoretical_loss": 3.6748696119086093, + "tokens_seen": 1008074752 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003507823470411234, + "loss": 1.5332, + "theoretical_loss": 3.674847086437052, + "tokens_seen": 1008140288 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035077231695085256, + "loss": 1.3947, + "theoretical_loss": 3.6748245628397336, + "tokens_seen": 1008205824 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035076228686058174, + "loss": 1.4901, + "theoretical_loss": 3.6748020411163753, + "tokens_seen": 1008271360 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003507522567703109, + "loss": 1.5725, + "theoretical_loss": 3.674779521266699, + "tokens_seen": 1008336896 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035074222668004016, + "loss": 1.5099, + "theoretical_loss": 3.674757003290429, + "tokens_seen": 1008402432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003507321965897693, + "loss": 1.5125, + "theoretical_loss": 3.674734487187285, + "tokens_seen": 1008467968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003507221664994985, + "loss": 1.3891, + "theoretical_loss": 3.6747119729569917, + "tokens_seen": 1008533504 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035071213640922765, + "loss": 1.411, + "theoretical_loss": 3.6746894605992706, + "tokens_seen": 1008599040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003507021063189569, + "loss": 1.4272, + "theoretical_loss": 3.674666950113844, + "tokens_seen": 1008664576 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035069207622868607, + "loss": 1.5266, + "theoretical_loss": 3.674644441500435, + "tokens_seen": 1008730112 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035068204613841525, + "loss": 1.4089, + "theoretical_loss": 3.674621934758767, + "tokens_seen": 1008795648 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035067201604814443, + "loss": 1.5692, + "theoretical_loss": 3.674599429888562, + "tokens_seen": 1008861184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003506619859578736, + "loss": 1.4315, + "theoretical_loss": 3.6745769268895425, + "tokens_seen": 1008926720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003506519558676028, + "loss": 1.5205, + "theoretical_loss": 3.674554425761432, + "tokens_seen": 1008992256 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350641925777332, + "loss": 1.376, + "theoretical_loss": 3.6745319265039535, + "tokens_seen": 1009057792 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035063189568706115, + "loss": 1.3639, + "theoretical_loss": 3.6745094291168297, + "tokens_seen": 1009123328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003506218655967904, + "loss": 1.4678, + "theoretical_loss": 3.6744869335997836, + "tokens_seen": 1009188864 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 370943, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5496997833251953, + "objective/train/theoretical_loss": 3.6744644399525384, + "objective/train/tokens_used": 1029714400, + "theoretical_loss": 3.6744644399525384, + "tokens_seen": 1009254400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003506118355065195, + "loss": 1.47, + "theoretical_loss": 3.6744644399525384, + "tokens_seen": 1009254400 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035060180541624875, + "loss": 1.504, + "theoretical_loss": 3.6744419481748176, + "tokens_seen": 1009319936 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350591775325978, + "loss": 1.468, + "theoretical_loss": 3.674419458266344, + "tokens_seen": 1009385472 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003505817452357071, + "loss": 1.4311, + "theoretical_loss": 3.674396970226841, + "tokens_seen": 1009451008 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035057171514543635, + "loss": 1.382, + "theoretical_loss": 3.674374484056032, + "tokens_seen": 1009516544 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035056168505516553, + "loss": 1.4246, + "theoretical_loss": 3.6743519997536414, + "tokens_seen": 1009582080 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003505516549648947, + "loss": 1.5169, + "theoretical_loss": 3.674329517319391, + "tokens_seen": 1009647616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003505416248746239, + "loss": 1.3471, + "theoretical_loss": 3.674307036753005, + "tokens_seen": 1009713152 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035053159478435307, + "loss": 1.4387, + "theoretical_loss": 3.6742845580542074, + "tokens_seen": 1009778688 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035052156469408225, + "loss": 1.4591, + "theoretical_loss": 3.6742620812227216, + "tokens_seen": 1009844224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003505115346038115, + "loss": 1.4126, + "theoretical_loss": 3.6742396062582703, + "tokens_seen": 1009909760 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003505015045135406, + "loss": 1.4524, + "theoretical_loss": 3.674217133160579, + "tokens_seen": 1009975296 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035049147442326985, + "loss": 1.4825, + "theoretical_loss": 3.6741946619293713, + "tokens_seen": 1010040832 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350481444332999, + "loss": 1.484, + "theoretical_loss": 3.6741721925643698, + "tokens_seen": 1010106368 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003504714142427282, + "loss": 1.4001, + "theoretical_loss": 3.6741497250652997, + "tokens_seen": 1010171904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003504613841524574, + "loss": 1.437, + "theoretical_loss": 3.6741272594318843, + "tokens_seen": 1010237440 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003504513540621866, + "loss": 1.418, + "theoretical_loss": 3.6741047956638475, + "tokens_seen": 1010302976 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035044132397191576, + "loss": 1.4794, + "theoretical_loss": 3.6740823337609143, + "tokens_seen": 1010368512 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035043129388164494, + "loss": 1.4526, + "theoretical_loss": 3.6740598737228085, + "tokens_seen": 1010434048 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003504212637913741, + "loss": 1.5527, + "theoretical_loss": 3.674037415549254, + "tokens_seen": 1010499584 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035041123370110335, + "loss": 1.4538, + "theoretical_loss": 3.674014959239975, + "tokens_seen": 1010565120 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003504012036108325, + "loss": 1.5071, + "theoretical_loss": 3.6739925047946964, + "tokens_seen": 1010630656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003503911735205617, + "loss": 1.4652, + "theoretical_loss": 3.6739700522131424, + "tokens_seen": 1010696192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003503811434302909, + "loss": 1.4108, + "theoretical_loss": 3.673947601495038, + "tokens_seen": 1010761728 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003503711133400201, + "loss": 1.4296, + "theoretical_loss": 3.6739251526401064, + "tokens_seen": 1010827264 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035036108324974926, + "loss": 1.3751, + "theoretical_loss": 3.6739027056480733, + "tokens_seen": 1010892800 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035035105315947844, + "loss": 1.4121, + "theoretical_loss": 3.673880260518663, + "tokens_seen": 1010958336 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003503410230692076, + "loss": 1.4094, + "theoretical_loss": 3.6738578172516005, + "tokens_seen": 1011023872 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035033099297893686, + "loss": 1.4519, + "theoretical_loss": 3.6738353758466102, + "tokens_seen": 1011089408 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350320962888666, + "loss": 1.583, + "theoretical_loss": 3.6738129363034178, + "tokens_seen": 1011154944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003503109327983952, + "loss": 1.4961, + "theoretical_loss": 3.6737904986217464, + "tokens_seen": 1011220480 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035030090270812435, + "loss": 1.5481, + "theoretical_loss": 3.6737680628013223, + "tokens_seen": 1011286016 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003502908726178536, + "loss": 1.5232, + "theoretical_loss": 3.6737456288418704, + "tokens_seen": 1011351552 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035028084252758276, + "loss": 1.5023, + "theoretical_loss": 3.673723196743116, + "tokens_seen": 1011417088 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035027081243731194, + "loss": 1.4189, + "theoretical_loss": 3.673700766504783, + "tokens_seen": 1011482624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003502607823470411, + "loss": 1.5105, + "theoretical_loss": 3.6736783381265976, + "tokens_seen": 1011548160 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035025075225677036, + "loss": 1.3852, + "theoretical_loss": 3.673655911608285, + "tokens_seen": 1011613696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003502407221664995, + "loss": 1.5305, + "theoretical_loss": 3.67363348694957, + "tokens_seen": 1011679232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003502306920762287, + "loss": 1.5108, + "theoretical_loss": 3.6736110641501787, + "tokens_seen": 1011744768 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035022066198595785, + "loss": 1.5133, + "theoretical_loss": 3.673588643209836, + "tokens_seen": 1011810304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003502106318956871, + "loss": 1.5252, + "theoretical_loss": 3.673566224128267, + "tokens_seen": 1011875840 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035020060180541627, + "loss": 1.4356, + "theoretical_loss": 3.673543806905198, + "tokens_seen": 1011941376 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035019057171514545, + "loss": 1.4305, + "theoretical_loss": 3.6735213915403544, + "tokens_seen": 1012006912 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035018054162487463, + "loss": 1.4886, + "theoretical_loss": 3.6734989780334617, + "tokens_seen": 1012072448 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003501705115346038, + "loss": 1.4504, + "theoretical_loss": 3.6734765663842452, + "tokens_seen": 1012137984 + }, + { + "epoch": 0.31, + "learning_rate": 0.000350160481444333, + "loss": 1.4175, + "theoretical_loss": 3.6734541565924315, + "tokens_seen": 1012203520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003501504513540622, + "loss": 1.4459, + "theoretical_loss": 3.673431748657746, + "tokens_seen": 1012269056 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035014042126379135, + "loss": 1.3849, + "theoretical_loss": 3.673409342579914, + "tokens_seen": 1012334592 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003501303911735206, + "loss": 1.6315, + "theoretical_loss": 3.6733869383586626, + "tokens_seen": 1012400128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003501203610832497, + "loss": 1.4677, + "theoretical_loss": 3.6733645359937173, + "tokens_seen": 1012465664 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 372166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4745606184005737, + "objective/train/theoretical_loss": 3.6733421354848037, + "objective/train/tokens_used": 1032991200, + "theoretical_loss": 3.6733421354848037, + "tokens_seen": 1012531200 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035011033099297895, + "loss": 1.5079, + "theoretical_loss": 3.6733421354848037, + "tokens_seen": 1012531200 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035010030090270813, + "loss": 1.4612, + "theoretical_loss": 3.6733197368316484, + "tokens_seen": 1012596736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003500902708124373, + "loss": 1.4673, + "theoretical_loss": 3.6732973400339772, + "tokens_seen": 1012662272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003500802407221665, + "loss": 1.4645, + "theoretical_loss": 3.6732749450915168, + "tokens_seen": 1012727808 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035007021063189573, + "loss": 1.5517, + "theoretical_loss": 3.6732525520039934, + "tokens_seen": 1012793344 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035006018054162486, + "loss": 1.4475, + "theoretical_loss": 3.6732301607711326, + "tokens_seen": 1012858880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003500501504513541, + "loss": 1.4019, + "theoretical_loss": 3.6732077713926623, + "tokens_seen": 1012924416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003500401203610832, + "loss": 1.4555, + "theoretical_loss": 3.673185383868307, + "tokens_seen": 1012989952 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035003009027081245, + "loss": 1.4757, + "theoretical_loss": 3.673162998197795, + "tokens_seen": 1013055488 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035002006018054163, + "loss": 1.4421, + "theoretical_loss": 3.673140614380852, + "tokens_seen": 1013121024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003500100300902708, + "loss": 1.5147, + "theoretical_loss": 3.673118232417205, + "tokens_seen": 1013186560 + }, + { + "epoch": 0.31, + "learning_rate": 0.00035, + "loss": 1.5441, + "theoretical_loss": 3.6730958523065804, + "tokens_seen": 1013252096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499899699097292, + "loss": 1.4118, + "theoretical_loss": 3.6730734740487048, + "tokens_seen": 1013317632 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034997993981945836, + "loss": 1.5535, + "theoretical_loss": 3.6730510976433055, + "tokens_seen": 1013383168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499699097291876, + "loss": 1.5184, + "theoretical_loss": 3.673028723090109, + "tokens_seen": 1013448704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499598796389167, + "loss": 1.431, + "theoretical_loss": 3.6730063503888424, + "tokens_seen": 1013514240 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034994984954864596, + "loss": 1.4967, + "theoretical_loss": 3.6729839795392323, + "tokens_seen": 1013579776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499398194583751, + "loss": 1.3555, + "theoretical_loss": 3.672961610541006, + "tokens_seen": 1013645312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499297893681043, + "loss": 1.4633, + "theoretical_loss": 3.672939243393891, + "tokens_seen": 1013710848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499197592778335, + "loss": 1.3919, + "theoretical_loss": 3.672916878097614, + "tokens_seen": 1013776384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003499097291875627, + "loss": 1.3841, + "theoretical_loss": 3.6728945146519023, + "tokens_seen": 1013841920 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034989969909729186, + "loss": 1.4175, + "theoretical_loss": 3.6728721530564834, + "tokens_seen": 1013907456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498896690070211, + "loss": 1.5232, + "theoretical_loss": 3.672849793311084, + "tokens_seen": 1013972992 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498796389167502, + "loss": 1.4529, + "theoretical_loss": 3.6728274354154324, + "tokens_seen": 1014038528 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034986960882647946, + "loss": 1.4509, + "theoretical_loss": 3.672805079369255, + "tokens_seen": 1014104064 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498595787362086, + "loss": 1.4848, + "theoretical_loss": 3.67278272517228, + "tokens_seen": 1014169600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498495486459378, + "loss": 1.4012, + "theoretical_loss": 3.6727603728242353, + "tokens_seen": 1014235136 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034983951855566706, + "loss": 1.3936, + "theoretical_loss": 3.6727380223248476, + "tokens_seen": 1014300672 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498294884653962, + "loss": 1.5198, + "theoretical_loss": 3.672715673673845, + "tokens_seen": 1014366208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003498194583751254, + "loss": 1.5382, + "theoretical_loss": 3.672693326870955, + "tokens_seen": 1014431744 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034980942828485455, + "loss": 1.4072, + "theoretical_loss": 3.672670981915906, + "tokens_seen": 1014497280 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003497993981945838, + "loss": 1.4478, + "theoretical_loss": 3.6726486388084254, + "tokens_seen": 1014562816 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034978936810431296, + "loss": 1.3657, + "theoretical_loss": 3.6726262975482413, + "tokens_seen": 1014628352 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034977933801404214, + "loss": 1.4016, + "theoretical_loss": 3.6726039581350816, + "tokens_seen": 1014693888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003497693079237713, + "loss": 1.5544, + "theoretical_loss": 3.672581620568674, + "tokens_seen": 1014759424 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034975927783350056, + "loss": 1.3857, + "theoretical_loss": 3.672559284848747, + "tokens_seen": 1014824960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003497492477432297, + "loss": 1.559, + "theoretical_loss": 3.6725369509750285, + "tokens_seen": 1014890496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003497392176529589, + "loss": 1.3813, + "theoretical_loss": 3.672514618947247, + "tokens_seen": 1014956032 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034972918756268805, + "loss": 1.46, + "theoretical_loss": 3.67249228876513, + "tokens_seen": 1015021568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003497191574724173, + "loss": 1.4279, + "theoretical_loss": 3.672469960428407, + "tokens_seen": 1015087104 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034970912738214647, + "loss": 1.4894, + "theoretical_loss": 3.6724476339368053, + "tokens_seen": 1015152640 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034969909729187565, + "loss": 1.437, + "theoretical_loss": 3.672425309290054, + "tokens_seen": 1015218176 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034968906720160483, + "loss": 1.4752, + "theoretical_loss": 3.6724029864878807, + "tokens_seen": 1015283712 + }, + { + "epoch": 0.31, + "learning_rate": 0.000349679037111334, + "loss": 1.4434, + "theoretical_loss": 3.6723806655300155, + "tokens_seen": 1015349248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003496690070210632, + "loss": 1.4659, + "theoretical_loss": 3.6723583464161855, + "tokens_seen": 1015414784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003496589769307924, + "loss": 1.5005, + "theoretical_loss": 3.67233602914612, + "tokens_seen": 1015480320 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034964894684052155, + "loss": 1.5089, + "theoretical_loss": 3.6723137137195474, + "tokens_seen": 1015545856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003496389167502508, + "loss": 1.4195, + "theoretical_loss": 3.672291400136197, + "tokens_seen": 1015611392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003496288866599799, + "loss": 1.4499, + "theoretical_loss": 3.6722690883957974, + "tokens_seen": 1015676928 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034961885656970915, + "loss": 1.5633, + "theoretical_loss": 3.672246778498077, + "tokens_seen": 1015742464 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 372984, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4236130714416504, + "objective/train/theoretical_loss": 3.6722244704427656, + "objective/train/tokens_used": 1036268000, + "theoretical_loss": 3.6722244704427656, + "tokens_seen": 1015808000 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034960882647943833, + "loss": 1.4463, + "theoretical_loss": 3.6722244704427656, + "tokens_seen": 1015808000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003495987963891675, + "loss": 1.3884, + "theoretical_loss": 3.6722021642295912, + "tokens_seen": 1015873536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003495887662988967, + "loss": 1.4336, + "theoretical_loss": 3.6721798598582835, + "tokens_seen": 1015939072 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034957873620862593, + "loss": 1.4859, + "theoretical_loss": 3.6721575573285716, + "tokens_seen": 1016004608 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034956870611835506, + "loss": 1.329, + "theoretical_loss": 3.6721352566401846, + "tokens_seen": 1016070144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003495586760280843, + "loss": 1.4722, + "theoretical_loss": 3.6721129577928524, + "tokens_seen": 1016135680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003495486459378134, + "loss": 1.5466, + "theoretical_loss": 3.672090660786303, + "tokens_seen": 1016201216 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034953861584754265, + "loss": 1.5206, + "theoretical_loss": 3.672068365620266, + "tokens_seen": 1016266752 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034952858575727183, + "loss": 1.4023, + "theoretical_loss": 3.6720460722944717, + "tokens_seen": 1016332288 + }, + { + "epoch": 0.31, + "learning_rate": 0.000349518555667001, + "loss": 1.5139, + "theoretical_loss": 3.672023780808648, + "tokens_seen": 1016397824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003495085255767302, + "loss": 1.4447, + "theoretical_loss": 3.6720014911625265, + "tokens_seen": 1016463360 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494984954864594, + "loss": 1.3974, + "theoretical_loss": 3.671979203355835, + "tokens_seen": 1016528896 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034948846539618856, + "loss": 1.4181, + "theoretical_loss": 3.671956917388304, + "tokens_seen": 1016594432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494784353059178, + "loss": 1.485, + "theoretical_loss": 3.6719346332596636, + "tokens_seen": 1016659968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494684052156469, + "loss": 1.5108, + "theoretical_loss": 3.6719123509696425, + "tokens_seen": 1016725504 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034945837512537616, + "loss": 1.5544, + "theoretical_loss": 3.6718900705179705, + "tokens_seen": 1016791040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494483450351053, + "loss": 1.3783, + "theoretical_loss": 3.6718677919043783, + "tokens_seen": 1016856576 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494383149448345, + "loss": 1.4883, + "theoretical_loss": 3.671845515128595, + "tokens_seen": 1016922112 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494282848545637, + "loss": 1.5358, + "theoretical_loss": 3.671823240190352, + "tokens_seen": 1016987648 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003494182547642929, + "loss": 1.4023, + "theoretical_loss": 3.6718009670893776, + "tokens_seen": 1017053184 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034940822467402206, + "loss": 1.473, + "theoretical_loss": 3.6717786958254024, + "tokens_seen": 1017118720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003493981945837513, + "loss": 1.4635, + "theoretical_loss": 3.671756426398157, + "tokens_seen": 1017184256 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003493881644934804, + "loss": 1.4402, + "theoretical_loss": 3.6717341588073715, + "tokens_seen": 1017249792 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034937813440320966, + "loss": 1.4995, + "theoretical_loss": 3.671711893052776, + "tokens_seen": 1017315328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003493681043129388, + "loss": 1.4083, + "theoretical_loss": 3.6716896291341, + "tokens_seen": 1017380864 + }, + { + "epoch": 0.31, + "learning_rate": 0.000349358074222668, + "loss": 1.4066, + "theoretical_loss": 3.671667367051076, + "tokens_seen": 1017446400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003493480441323972, + "loss": 1.3574, + "theoretical_loss": 3.671645106803432, + "tokens_seen": 1017511936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003493380140421264, + "loss": 1.484, + "theoretical_loss": 3.6716228483909, + "tokens_seen": 1017577472 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034932798395185557, + "loss": 1.4238, + "theoretical_loss": 3.6716005918132097, + "tokens_seen": 1017643008 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034931795386158475, + "loss": 1.5143, + "theoretical_loss": 3.671578337070093, + "tokens_seen": 1017708544 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034930792377131393, + "loss": 1.3851, + "theoretical_loss": 3.6715560841612795, + "tokens_seen": 1017774080 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034929789368104316, + "loss": 1.4629, + "theoretical_loss": 3.6715338330864995, + "tokens_seen": 1017839616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003492878635907723, + "loss": 1.4293, + "theoretical_loss": 3.6715115838454846, + "tokens_seen": 1017905152 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003492778335005015, + "loss": 1.4809, + "theoretical_loss": 3.6714893364379657, + "tokens_seen": 1017970688 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034926780341023065, + "loss": 1.4951, + "theoretical_loss": 3.671467090863673, + "tokens_seen": 1018036224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003492577733199599, + "loss": 1.4825, + "theoretical_loss": 3.6714448471223378, + "tokens_seen": 1018101760 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034924774322968907, + "loss": 1.4423, + "theoretical_loss": 3.6714226052136913, + "tokens_seen": 1018167296 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034923771313941825, + "loss": 1.4731, + "theoretical_loss": 3.6714003651374645, + "tokens_seen": 1018232832 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034922768304914743, + "loss": 1.3845, + "theoretical_loss": 3.671378126893388, + "tokens_seen": 1018298368 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034921765295887667, + "loss": 1.4938, + "theoretical_loss": 3.671355890481194, + "tokens_seen": 1018363904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003492076228686058, + "loss": 1.3915, + "theoretical_loss": 3.6713336559006127, + "tokens_seen": 1018429440 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034919759277833503, + "loss": 1.5349, + "theoretical_loss": 3.6713114231513755, + "tokens_seen": 1018494976 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034918756268806416, + "loss": 1.4014, + "theoretical_loss": 3.671289192233214, + "tokens_seen": 1018560512 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003491775325977934, + "loss": 1.4829, + "theoretical_loss": 3.6712669631458597, + "tokens_seen": 1018626048 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034916750250752257, + "loss": 1.4476, + "theoretical_loss": 3.671244735889044, + "tokens_seen": 1018691584 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034915747241725175, + "loss": 1.5239, + "theoretical_loss": 3.6712225104624983, + "tokens_seen": 1018757120 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034914744232698093, + "loss": 1.5808, + "theoretical_loss": 3.6712002868659543, + "tokens_seen": 1018822656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003491374122367101, + "loss": 1.3767, + "theoretical_loss": 3.6711780650991432, + "tokens_seen": 1018888192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003491273821464393, + "loss": 1.4492, + "theoretical_loss": 3.671155845161797, + "tokens_seen": 1018953728 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034911735205616853, + "loss": 1.5118, + "theoretical_loss": 3.671133627053648, + "tokens_seen": 1019019264 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 374190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.405362606048584, + "objective/train/theoretical_loss": 3.6711114107744267, + "objective/train/tokens_used": 1039544800, + "theoretical_loss": 3.6711114107744267, + "tokens_seen": 1019084800 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034910732196589766, + "loss": 1.468, + "theoretical_loss": 3.6711114107744267, + "tokens_seen": 1019084800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490972918756269, + "loss": 1.5317, + "theoretical_loss": 3.671089196323866, + "tokens_seen": 1019150336 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034908726178535613, + "loss": 1.414, + "theoretical_loss": 3.6710669837016976, + "tokens_seen": 1019215872 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034907723169508526, + "loss": 1.3819, + "theoretical_loss": 3.671044772907653, + "tokens_seen": 1019281408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490672016048145, + "loss": 1.5093, + "theoretical_loss": 3.671022563941465, + "tokens_seen": 1019346944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490571715145436, + "loss": 1.4252, + "theoretical_loss": 3.671000356802865, + "tokens_seen": 1019412480 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034904714142427285, + "loss": 1.5645, + "theoretical_loss": 3.6709781514915854, + "tokens_seen": 1019478016 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034903711133400203, + "loss": 1.4723, + "theoretical_loss": 3.670955948007358, + "tokens_seen": 1019543552 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490270812437312, + "loss": 1.3948, + "theoretical_loss": 3.6709337463499154, + "tokens_seen": 1019609088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490170511534604, + "loss": 1.4976, + "theoretical_loss": 3.6709115465189903, + "tokens_seen": 1019674624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003490070210631896, + "loss": 1.5424, + "theoretical_loss": 3.670889348514314, + "tokens_seen": 1019740160 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034899699097291876, + "loss": 1.4719, + "theoretical_loss": 3.67086715233562, + "tokens_seen": 1019805696 + }, + { + "epoch": 0.31, + "learning_rate": 0.000348986960882648, + "loss": 1.3321, + "theoretical_loss": 3.6708449579826405, + "tokens_seen": 1019871232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489769307923771, + "loss": 1.5336, + "theoretical_loss": 3.6708227654551076, + "tokens_seen": 1019936768 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034896690070210636, + "loss": 1.52, + "theoretical_loss": 3.6708005747527537, + "tokens_seen": 1020002304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489568706118355, + "loss": 1.5259, + "theoretical_loss": 3.6707783858753125, + "tokens_seen": 1020067840 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489468405215647, + "loss": 1.4754, + "theoretical_loss": 3.6707561988225157, + "tokens_seen": 1020133376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489368104312939, + "loss": 1.3068, + "theoretical_loss": 3.670734013594096, + "tokens_seen": 1020198912 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489267803410231, + "loss": 1.5114, + "theoretical_loss": 3.670711830189787, + "tokens_seen": 1020264448 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034891675025075226, + "loss": 1.5121, + "theoretical_loss": 3.6706896486093212, + "tokens_seen": 1020329984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003489067201604815, + "loss": 1.4864, + "theoretical_loss": 3.6706674688524314, + "tokens_seen": 1020395520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003488966900702106, + "loss": 1.5016, + "theoretical_loss": 3.670645290918851, + "tokens_seen": 1020461056 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034888665997993986, + "loss": 1.4269, + "theoretical_loss": 3.670623114808312, + "tokens_seen": 1020526592 + }, + { + "epoch": 0.31, + "learning_rate": 0.000348876629889669, + "loss": 1.4127, + "theoretical_loss": 3.670600940520549, + "tokens_seen": 1020592128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003488665997993982, + "loss": 1.4482, + "theoretical_loss": 3.670578768055294, + "tokens_seen": 1020657664 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003488565697091274, + "loss": 1.4463, + "theoretical_loss": 3.67055659741228, + "tokens_seen": 1020723200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003488465396188566, + "loss": 1.3992, + "theoretical_loss": 3.670534428591241, + "tokens_seen": 1020788736 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034883650952858577, + "loss": 1.3346, + "theoretical_loss": 3.6705122615919104, + "tokens_seen": 1020854272 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034882647943831495, + "loss": 1.5576, + "theoretical_loss": 3.6704900964140212, + "tokens_seen": 1020919808 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034881644934804413, + "loss": 1.5888, + "theoretical_loss": 3.6704679330573065, + "tokens_seen": 1020985344 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034880641925777336, + "loss": 1.4318, + "theoretical_loss": 3.6704457715215004, + "tokens_seen": 1021050880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003487963891675025, + "loss": 1.534, + "theoretical_loss": 3.6704236118063363, + "tokens_seen": 1021116416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003487863590772317, + "loss": 1.5598, + "theoretical_loss": 3.6704014539115475, + "tokens_seen": 1021181952 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034877632898696085, + "loss": 1.4113, + "theoretical_loss": 3.670379297836868, + "tokens_seen": 1021247488 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003487662988966901, + "loss": 1.3676, + "theoretical_loss": 3.6703571435820312, + "tokens_seen": 1021313024 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034875626880641927, + "loss": 1.4976, + "theoretical_loss": 3.6703349911467713, + "tokens_seen": 1021378560 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034874623871614845, + "loss": 1.5185, + "theoretical_loss": 3.670312840530821, + "tokens_seen": 1021444096 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034873620862587763, + "loss": 1.3611, + "theoretical_loss": 3.670290691733916, + "tokens_seen": 1021509632 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034872617853560687, + "loss": 1.3218, + "theoretical_loss": 3.6702685447557886, + "tokens_seen": 1021575168 + }, + { + "epoch": 0.31, + "learning_rate": 0.000348716148445336, + "loss": 1.4674, + "theoretical_loss": 3.6702463995961736, + "tokens_seen": 1021640704 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034870611835506523, + "loss": 1.3989, + "theoretical_loss": 3.6702242562548046, + "tokens_seen": 1021706240 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034869608826479436, + "loss": 1.4568, + "theoretical_loss": 3.670202114731416, + "tokens_seen": 1021771776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003486860581745236, + "loss": 1.2957, + "theoretical_loss": 3.670179975025742, + "tokens_seen": 1021837312 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034867602808425277, + "loss": 1.3817, + "theoretical_loss": 3.670157837137517, + "tokens_seen": 1021902848 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034866599799398195, + "loss": 1.4925, + "theoretical_loss": 3.6701357010664744, + "tokens_seen": 1021968384 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034865596790371113, + "loss": 1.4387, + "theoretical_loss": 3.6701135668123497, + "tokens_seen": 1022033920 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003486459378134403, + "loss": 1.4251, + "theoretical_loss": 3.670091434374876, + "tokens_seen": 1022099456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003486359077231695, + "loss": 1.4995, + "theoretical_loss": 3.670069303753789, + "tokens_seen": 1022164992 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034862587763289873, + "loss": 1.519, + "theoretical_loss": 3.670047174948822, + "tokens_seen": 1022230528 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034861584754262786, + "loss": 1.5067, + "theoretical_loss": 3.6700250479597107, + "tokens_seen": 1022296064 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 375283, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4147348403930664, + "objective/train/theoretical_loss": 3.6700029227861886, + "objective/train/tokens_used": 1042821600, + "theoretical_loss": 3.6700029227861886, + "tokens_seen": 1022361600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003486058174523571, + "loss": 1.4682, + "theoretical_loss": 3.6700029227861886, + "tokens_seen": 1022361600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003485957873620862, + "loss": 1.4034, + "theoretical_loss": 3.669980799427991, + "tokens_seen": 1022427136 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034858575727181546, + "loss": 1.4405, + "theoretical_loss": 3.6699586778848525, + "tokens_seen": 1022492672 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034857572718154464, + "loss": 1.3539, + "theoretical_loss": 3.6699365581565084, + "tokens_seen": 1022558208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003485656970912738, + "loss": 1.4336, + "theoretical_loss": 3.6699144402426924, + "tokens_seen": 1022623744 + }, + { + "epoch": 0.31, + "learning_rate": 0.000348555667001003, + "loss": 1.4904, + "theoretical_loss": 3.66989232414314, + "tokens_seen": 1022689280 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034854563691073223, + "loss": 1.4171, + "theoretical_loss": 3.6698702098575864, + "tokens_seen": 1022754816 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034853560682046136, + "loss": 1.343, + "theoretical_loss": 3.669848097385766, + "tokens_seen": 1022820352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003485255767301906, + "loss": 1.459, + "theoretical_loss": 3.669825986727414, + "tokens_seen": 1022885888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003485155466399197, + "loss": 1.5477, + "theoretical_loss": 3.669803877882266, + "tokens_seen": 1022951424 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034850551654964896, + "loss": 1.6056, + "theoretical_loss": 3.6697817708500566, + "tokens_seen": 1023016960 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034849548645937814, + "loss": 1.4639, + "theoretical_loss": 3.6697596656305214, + "tokens_seen": 1023082496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484854563691073, + "loss": 1.3897, + "theoretical_loss": 3.6697375622233954, + "tokens_seen": 1023148032 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484754262788365, + "loss": 1.4537, + "theoretical_loss": 3.6697154606284137, + "tokens_seen": 1023213568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484653961885657, + "loss": 1.4442, + "theoretical_loss": 3.669693360845312, + "tokens_seen": 1023279104 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034845536609829486, + "loss": 1.4451, + "theoretical_loss": 3.669671262873826, + "tokens_seen": 1023344640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484453360080241, + "loss": 1.5244, + "theoretical_loss": 3.6696491667136906, + "tokens_seen": 1023410176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484353059177532, + "loss": 1.5925, + "theoretical_loss": 3.6696270723646416, + "tokens_seen": 1023475712 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034842527582748246, + "loss": 1.5209, + "theoretical_loss": 3.669604979826415, + "tokens_seen": 1023541248 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034841524573721164, + "loss": 1.4236, + "theoretical_loss": 3.669582889098746, + "tokens_seen": 1023606784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003484052156469408, + "loss": 1.4733, + "theoretical_loss": 3.6695608001813707, + "tokens_seen": 1023672320 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034839518555667, + "loss": 1.56, + "theoretical_loss": 3.669538713074024, + "tokens_seen": 1023737856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003483851554663992, + "loss": 1.526, + "theoretical_loss": 3.669516627776442, + "tokens_seen": 1023803392 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034837512537612837, + "loss": 1.3743, + "theoretical_loss": 3.669494544288362, + "tokens_seen": 1023868928 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003483650952858576, + "loss": 1.455, + "theoretical_loss": 3.6694724626095185, + "tokens_seen": 1023934464 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034835506519558673, + "loss": 1.298, + "theoretical_loss": 3.6694503827396474, + "tokens_seen": 1024000000 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034834503510531597, + "loss": 1.3959, + "theoretical_loss": 3.669428304678485, + "tokens_seen": 1024065536 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034833500501504515, + "loss": 1.527, + "theoretical_loss": 3.669406228425768, + "tokens_seen": 1024131072 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034832497492477433, + "loss": 1.5245, + "theoretical_loss": 3.669384153981232, + "tokens_seen": 1024196608 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034831494483450356, + "loss": 1.4598, + "theoretical_loss": 3.6693620813446133, + "tokens_seen": 1024262144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003483049147442327, + "loss": 1.4899, + "theoretical_loss": 3.669340010515648, + "tokens_seen": 1024327680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003482948846539619, + "loss": 1.3647, + "theoretical_loss": 3.6693179414940733, + "tokens_seen": 1024393216 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034828485456369105, + "loss": 1.5221, + "theoretical_loss": 3.6692958742796242, + "tokens_seen": 1024458752 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003482748244734203, + "loss": 1.3586, + "theoretical_loss": 3.6692738088720382, + "tokens_seen": 1024524288 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034826479438314947, + "loss": 1.4999, + "theoretical_loss": 3.669251745271051, + "tokens_seen": 1024589824 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034825476429287865, + "loss": 1.5189, + "theoretical_loss": 3.6692296834764, + "tokens_seen": 1024655360 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034824473420260783, + "loss": 1.438, + "theoretical_loss": 3.6692076234878215, + "tokens_seen": 1024720896 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034823470411233707, + "loss": 1.4438, + "theoretical_loss": 3.6691855653050514, + "tokens_seen": 1024786432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003482246740220662, + "loss": 1.4661, + "theoretical_loss": 3.669163508927827, + "tokens_seen": 1024851968 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034821464393179543, + "loss": 1.5118, + "theoretical_loss": 3.6691414543558856, + "tokens_seen": 1024917504 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034820461384152456, + "loss": 1.5266, + "theoretical_loss": 3.669119401588963, + "tokens_seen": 1024983040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003481945837512538, + "loss": 1.3813, + "theoretical_loss": 3.669097350626797, + "tokens_seen": 1025048576 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034818455366098297, + "loss": 1.4428, + "theoretical_loss": 3.669075301469124, + "tokens_seen": 1025114112 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034817452357071215, + "loss": 1.3636, + "theoretical_loss": 3.6690532541156813, + "tokens_seen": 1025179648 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034816449348044133, + "loss": 1.3834, + "theoretical_loss": 3.6690312085662056, + "tokens_seen": 1025245184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003481544633901705, + "loss": 1.4962, + "theoretical_loss": 3.6690091648204337, + "tokens_seen": 1025310720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003481444332998997, + "loss": 1.4101, + "theoretical_loss": 3.6689871228781037, + "tokens_seen": 1025376256 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034813440320962893, + "loss": 1.4471, + "theoretical_loss": 3.6689650827389517, + "tokens_seen": 1025441792 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034812437311935806, + "loss": 1.352, + "theoretical_loss": 3.6689430444027162, + "tokens_seen": 1025507328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003481143430290873, + "loss": 1.4878, + "theoretical_loss": 3.668921007869134, + "tokens_seen": 1025572864 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 376516, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2511804103851318, + "objective/train/theoretical_loss": 3.6688989731379413, + "objective/train/tokens_used": 1046098400, + "theoretical_loss": 3.6688989731379413, + "tokens_seen": 1025638400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003481043129388164, + "loss": 1.3347, + "theoretical_loss": 3.6688989731379413, + "tokens_seen": 1025638400 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034809428284854566, + "loss": 1.3183, + "theoretical_loss": 3.6688769402088774, + "tokens_seen": 1025703936 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034808425275827484, + "loss": 1.531, + "theoretical_loss": 3.668854909081679, + "tokens_seen": 1025769472 + }, + { + "epoch": 0.31, + "learning_rate": 0.000348074222668004, + "loss": 1.3272, + "theoretical_loss": 3.668832879756083, + "tokens_seen": 1025835008 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003480641925777332, + "loss": 1.4793, + "theoretical_loss": 3.668810852231828, + "tokens_seen": 1025900544 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034805416248746243, + "loss": 1.428, + "theoretical_loss": 3.668788826508651, + "tokens_seen": 1025966080 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034804413239719156, + "loss": 1.4178, + "theoretical_loss": 3.6687668025862905, + "tokens_seen": 1026031616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003480341023069208, + "loss": 1.4685, + "theoretical_loss": 3.668744780464483, + "tokens_seen": 1026097152 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003480240722166499, + "loss": 1.3827, + "theoretical_loss": 3.668722760142967, + "tokens_seen": 1026162688 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034801404212637916, + "loss": 1.3275, + "theoretical_loss": 3.668700741621481, + "tokens_seen": 1026228224 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034800401203610834, + "loss": 1.4884, + "theoretical_loss": 3.6686787248997614, + "tokens_seen": 1026293760 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003479939819458375, + "loss": 1.3606, + "theoretical_loss": 3.6686567099775473, + "tokens_seen": 1026359296 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003479839518555667, + "loss": 1.4718, + "theoretical_loss": 3.6686346968545767, + "tokens_seen": 1026424832 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003479739217652959, + "loss": 1.4734, + "theoretical_loss": 3.6686126855305874, + "tokens_seen": 1026490368 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034796389167502506, + "loss": 1.4488, + "theoretical_loss": 3.6685906760053175, + "tokens_seen": 1026555904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003479538615847543, + "loss": 1.4488, + "theoretical_loss": 3.6685686682785055, + "tokens_seen": 1026621440 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034794383149448343, + "loss": 1.3822, + "theoretical_loss": 3.6685466623498892, + "tokens_seen": 1026686976 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034793380140421266, + "loss": 1.4342, + "theoretical_loss": 3.6685246582192077, + "tokens_seen": 1026752512 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034792377131394184, + "loss": 1.4763, + "theoretical_loss": 3.668502655886198, + "tokens_seen": 1026818048 + }, + { + "epoch": 0.31, + "learning_rate": 0.000347913741223671, + "loss": 1.4596, + "theoretical_loss": 3.6684806553505998, + "tokens_seen": 1026883584 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003479037111334002, + "loss": 1.4754, + "theoretical_loss": 3.6684586566121506, + "tokens_seen": 1026949120 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003478936810431294, + "loss": 1.3854, + "theoretical_loss": 3.66843665967059, + "tokens_seen": 1027014656 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034788365095285857, + "loss": 1.5441, + "theoretical_loss": 3.668414664525656, + "tokens_seen": 1027080192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003478736208625878, + "loss": 1.4318, + "theoretical_loss": 3.668392671177087, + "tokens_seen": 1027145728 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034786359077231693, + "loss": 1.5365, + "theoretical_loss": 3.668370679624622, + "tokens_seen": 1027211264 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034785356068204617, + "loss": 1.4834, + "theoretical_loss": 3.6683486898679996, + "tokens_seen": 1027276800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003478435305917753, + "loss": 1.4267, + "theoretical_loss": 3.6683267019069588, + "tokens_seen": 1027342336 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034783350050150453, + "loss": 1.5143, + "theoretical_loss": 3.668304715741238, + "tokens_seen": 1027407872 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003478234704112337, + "loss": 1.3702, + "theoretical_loss": 3.6682827313705766, + "tokens_seen": 1027473408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003478134403209629, + "loss": 1.523, + "theoretical_loss": 3.6682607487947134, + "tokens_seen": 1027538944 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034780341023069207, + "loss": 1.4245, + "theoretical_loss": 3.668238768013387, + "tokens_seen": 1027604480 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034779338014042125, + "loss": 1.5289, + "theoretical_loss": 3.6682167890263377, + "tokens_seen": 1027670016 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034778335005015043, + "loss": 1.4243, + "theoretical_loss": 3.668194811833303, + "tokens_seen": 1027735552 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034777331995987967, + "loss": 1.42, + "theoretical_loss": 3.6681728364340236, + "tokens_seen": 1027801088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003477632898696088, + "loss": 1.4803, + "theoretical_loss": 3.668150862828237, + "tokens_seen": 1027866624 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034775325977933803, + "loss": 1.3167, + "theoretical_loss": 3.668128891015684, + "tokens_seen": 1027932160 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003477432296890672, + "loss": 1.5083, + "theoretical_loss": 3.6681069209961037, + "tokens_seen": 1027997696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003477331995987964, + "loss": 1.5271, + "theoretical_loss": 3.668084952769235, + "tokens_seen": 1028063232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003477231695085256, + "loss": 1.4602, + "theoretical_loss": 3.6680629863348178, + "tokens_seen": 1028128768 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034771313941825476, + "loss": 1.4055, + "theoretical_loss": 3.6680410216925914, + "tokens_seen": 1028194304 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034770310932798394, + "loss": 1.3968, + "theoretical_loss": 3.6680190588422947, + "tokens_seen": 1028259840 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034769307923771317, + "loss": 1.4384, + "theoretical_loss": 3.6679970977836684, + "tokens_seen": 1028325376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003476830491474423, + "loss": 1.5182, + "theoretical_loss": 3.667975138516452, + "tokens_seen": 1028390912 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034767301905717153, + "loss": 1.3969, + "theoretical_loss": 3.6679531810403843, + "tokens_seen": 1028456448 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034766298896690066, + "loss": 1.5326, + "theoretical_loss": 3.6679312253552063, + "tokens_seen": 1028521984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003476529588766299, + "loss": 1.4147, + "theoretical_loss": 3.6679092714606574, + "tokens_seen": 1028587520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003476429287863591, + "loss": 1.3998, + "theoretical_loss": 3.6678873193564767, + "tokens_seen": 1028653056 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034763289869608826, + "loss": 1.5102, + "theoretical_loss": 3.6678653690424055, + "tokens_seen": 1028718592 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034762286860581744, + "loss": 1.3488, + "theoretical_loss": 3.6678434205181825, + "tokens_seen": 1028784128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003476128385155466, + "loss": 1.4282, + "theoretical_loss": 3.6678214737835484, + "tokens_seen": 1028849664 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 377682, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3727624416351318, + "objective/train/theoretical_loss": 3.667799528838244, + "objective/train/tokens_used": 1049375200, + "theoretical_loss": 3.667799528838244, + "tokens_seen": 1028915200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003476028084252758, + "loss": 1.4011, + "theoretical_loss": 3.667799528838244, + "tokens_seen": 1028915200 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034759277833500504, + "loss": 1.4929, + "theoretical_loss": 3.667777585682008, + "tokens_seen": 1028980736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003475827482447342, + "loss": 1.5492, + "theoretical_loss": 3.667755644314581, + "tokens_seen": 1029046272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003475727181544634, + "loss": 1.3525, + "theoretical_loss": 3.667733704735704, + "tokens_seen": 1029111808 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034756268806419263, + "loss": 1.4212, + "theoretical_loss": 3.667711766945117, + "tokens_seen": 1029177344 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034755265797392176, + "loss": 1.4939, + "theoretical_loss": 3.6676898309425603, + "tokens_seen": 1029242880 + }, + { + "epoch": 0.31, + "learning_rate": 0.000347542627883651, + "loss": 1.5147, + "theoretical_loss": 3.6676678967277745, + "tokens_seen": 1029308416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003475325977933801, + "loss": 1.5172, + "theoretical_loss": 3.6676459643004993, + "tokens_seen": 1029373952 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034752256770310936, + "loss": 1.4993, + "theoretical_loss": 3.667624033660476, + "tokens_seen": 1029439488 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034751253761283854, + "loss": 1.4962, + "theoretical_loss": 3.6676021048074454, + "tokens_seen": 1029505024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003475025075225677, + "loss": 1.3825, + "theoretical_loss": 3.6675801777411476, + "tokens_seen": 1029570560 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474924774322969, + "loss": 1.4369, + "theoretical_loss": 3.6675582524613235, + "tokens_seen": 1029636096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474824473420261, + "loss": 1.4941, + "theoretical_loss": 3.667536328967714, + "tokens_seen": 1029701632 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034747241725175527, + "loss": 1.5106, + "theoretical_loss": 3.6675144072600596, + "tokens_seen": 1029767168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474623871614845, + "loss": 1.4161, + "theoretical_loss": 3.667492487338101, + "tokens_seen": 1029832704 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034745235707121363, + "loss": 1.5473, + "theoretical_loss": 3.66747056920158, + "tokens_seen": 1029898240 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034744232698094286, + "loss": 1.4296, + "theoretical_loss": 3.667448652850237, + "tokens_seen": 1029963776 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034743229689067204, + "loss": 1.5402, + "theoretical_loss": 3.6674267382838126, + "tokens_seen": 1030029312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474222668004012, + "loss": 1.3851, + "theoretical_loss": 3.6674048255020484, + "tokens_seen": 1030094848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474122367101304, + "loss": 1.361, + "theoretical_loss": 3.6673829145046852, + "tokens_seen": 1030160384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003474022066198596, + "loss": 1.3974, + "theoretical_loss": 3.667361005291465, + "tokens_seen": 1030225920 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034739217652958877, + "loss": 1.4481, + "theoretical_loss": 3.6673390978621283, + "tokens_seen": 1030291456 + }, + { + "epoch": 0.31, + "learning_rate": 0.000347382146439318, + "loss": 1.4609, + "theoretical_loss": 3.6673171922164163, + "tokens_seen": 1030356992 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034737211634904713, + "loss": 1.3793, + "theoretical_loss": 3.6672952883540706, + "tokens_seen": 1030422528 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034736208625877637, + "loss": 1.4512, + "theoretical_loss": 3.6672733862748332, + "tokens_seen": 1030488064 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003473520561685055, + "loss": 1.3585, + "theoretical_loss": 3.6672514859784444, + "tokens_seen": 1030553600 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034734202607823473, + "loss": 1.4588, + "theoretical_loss": 3.6672295874646466, + "tokens_seen": 1030619136 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003473319959879639, + "loss": 1.4416, + "theoretical_loss": 3.6672076907331808, + "tokens_seen": 1030684672 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003473219658976931, + "loss": 1.588, + "theoretical_loss": 3.6671857957837886, + "tokens_seen": 1030750208 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034731193580742227, + "loss": 1.4388, + "theoretical_loss": 3.6671639026162124, + "tokens_seen": 1030815744 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034730190571715145, + "loss": 1.4314, + "theoretical_loss": 3.6671420112301933, + "tokens_seen": 1030881280 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034729187562688063, + "loss": 1.5277, + "theoretical_loss": 3.6671201216254734, + "tokens_seen": 1030946816 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034728184553660987, + "loss": 1.4575, + "theoretical_loss": 3.667098233801794, + "tokens_seen": 1031012352 + }, + { + "epoch": 0.31, + "learning_rate": 0.000347271815446339, + "loss": 1.3487, + "theoretical_loss": 3.667076347758898, + "tokens_seen": 1031077888 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034726178535606823, + "loss": 1.4653, + "theoretical_loss": 3.6670544634965263, + "tokens_seen": 1031143424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003472517552657974, + "loss": 1.5627, + "theoretical_loss": 3.6670325810144213, + "tokens_seen": 1031208960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003472417251755266, + "loss": 1.4307, + "theoretical_loss": 3.6670107003123253, + "tokens_seen": 1031274496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003472316950852558, + "loss": 1.3841, + "theoretical_loss": 3.66698882138998, + "tokens_seen": 1031340032 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034722166499498496, + "loss": 1.565, + "theoretical_loss": 3.666966944247128, + "tokens_seen": 1031405568 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034721163490471414, + "loss": 1.5333, + "theoretical_loss": 3.666945068883511, + "tokens_seen": 1031471104 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034720160481444337, + "loss": 1.6147, + "theoretical_loss": 3.6669231952988715, + "tokens_seen": 1031536640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003471915747241725, + "loss": 1.4652, + "theoretical_loss": 3.6669013234929517, + "tokens_seen": 1031602176 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034718154463390173, + "loss": 1.4722, + "theoretical_loss": 3.6668794534654943, + "tokens_seen": 1031667712 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034717151454363086, + "loss": 1.3876, + "theoretical_loss": 3.6668575852162415, + "tokens_seen": 1031733248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003471614844533601, + "loss": 1.4658, + "theoretical_loss": 3.6668357187449354, + "tokens_seen": 1031798784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003471514543630893, + "loss": 1.3827, + "theoretical_loss": 3.6668138540513198, + "tokens_seen": 1031864320 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034714142427281846, + "loss": 1.4655, + "theoretical_loss": 3.6667919911351357, + "tokens_seen": 1031929856 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034713139418254764, + "loss": 1.4282, + "theoretical_loss": 3.6667701299961264, + "tokens_seen": 1031995392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003471213640922768, + "loss": 1.3619, + "theoretical_loss": 3.6667482706340353, + "tokens_seen": 1032060928 + }, + { + "epoch": 0.31, + "learning_rate": 0.000347111334002006, + "loss": 1.3751, + "theoretical_loss": 3.666726413048604, + "tokens_seen": 1032126464 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 378946, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.434868574142456, + "objective/train/theoretical_loss": 3.6667045572395756, + "objective/train/tokens_used": 1052652000, + "theoretical_loss": 3.6667045572395756, + "tokens_seen": 1032192000 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034710130391173524, + "loss": 1.3707, + "theoretical_loss": 3.6667045572395756, + "tokens_seen": 1032192000 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034709127382146436, + "loss": 1.325, + "theoretical_loss": 3.6666827032066935, + "tokens_seen": 1032257536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003470812437311936, + "loss": 1.413, + "theoretical_loss": 3.6666608509497003, + "tokens_seen": 1032323072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003470712136409228, + "loss": 1.3247, + "theoretical_loss": 3.666639000468339, + "tokens_seen": 1032388608 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034706118355065196, + "loss": 1.4129, + "theoretical_loss": 3.6666171517623525, + "tokens_seen": 1032454144 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034705115346038114, + "loss": 1.368, + "theoretical_loss": 3.6665953048314837, + "tokens_seen": 1032519680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003470411233701103, + "loss": 1.4667, + "theoretical_loss": 3.666573459675476, + "tokens_seen": 1032585216 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003470310932798395, + "loss": 1.4098, + "theoretical_loss": 3.666551616294073, + "tokens_seen": 1032650752 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034702106318956874, + "loss": 1.5339, + "theoretical_loss": 3.6665297746870165, + "tokens_seen": 1032716288 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034701103309929787, + "loss": 1.5518, + "theoretical_loss": 3.6665079348540512, + "tokens_seen": 1032781824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003470010030090271, + "loss": 1.4807, + "theoretical_loss": 3.6664860967949204, + "tokens_seen": 1032847360 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034699097291875623, + "loss": 1.5089, + "theoretical_loss": 3.6664642605093665, + "tokens_seen": 1032912896 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034698094282848547, + "loss": 1.377, + "theoretical_loss": 3.6664424259971335, + "tokens_seen": 1032978432 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034697091273821465, + "loss": 1.2642, + "theoretical_loss": 3.666420593257965, + "tokens_seen": 1033043968 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034696088264794383, + "loss": 1.3213, + "theoretical_loss": 3.6663987622916046, + "tokens_seen": 1033109504 + }, + { + "epoch": 0.31, + "learning_rate": 0.000346950852557673, + "loss": 1.2633, + "theoretical_loss": 3.6663769330977956, + "tokens_seen": 1033175040 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034694082246740224, + "loss": 1.3781, + "theoretical_loss": 3.6663551056762818, + "tokens_seen": 1033240576 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034693079237713137, + "loss": 1.5084, + "theoretical_loss": 3.666333280026807, + "tokens_seen": 1033306112 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003469207622868606, + "loss": 1.412, + "theoretical_loss": 3.6663114561491144, + "tokens_seen": 1033371648 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034691073219658973, + "loss": 1.4894, + "theoretical_loss": 3.6662896340429487, + "tokens_seen": 1033437184 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034690070210631897, + "loss": 1.4604, + "theoretical_loss": 3.6662678137080533, + "tokens_seen": 1033502720 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034689067201604815, + "loss": 1.2709, + "theoretical_loss": 3.6662459951441724, + "tokens_seen": 1033568256 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034688064192577733, + "loss": 1.3528, + "theoretical_loss": 3.6662241783510496, + "tokens_seen": 1033633792 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003468706118355065, + "loss": 1.3751, + "theoretical_loss": 3.6662023633284293, + "tokens_seen": 1033699328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003468605817452357, + "loss": 1.367, + "theoretical_loss": 3.6661805500760547, + "tokens_seen": 1033764864 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003468505516549649, + "loss": 1.4897, + "theoretical_loss": 3.6661587385936714, + "tokens_seen": 1033830400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003468405215646941, + "loss": 1.5571, + "theoretical_loss": 3.6661369288810226, + "tokens_seen": 1033895936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003468304914744233, + "loss": 1.3924, + "theoretical_loss": 3.666115120937852, + "tokens_seen": 1033961472 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034682046138415247, + "loss": 1.4176, + "theoretical_loss": 3.6660933147639057, + "tokens_seen": 1034027008 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034681043129388165, + "loss": 1.2617, + "theoretical_loss": 3.6660715103589263, + "tokens_seen": 1034092544 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034680040120361083, + "loss": 1.3183, + "theoretical_loss": 3.6660497077226593, + "tokens_seen": 1034158080 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034679037111334007, + "loss": 1.448, + "theoretical_loss": 3.6660279068548487, + "tokens_seen": 1034223616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003467803410230692, + "loss": 1.4675, + "theoretical_loss": 3.666006107755239, + "tokens_seen": 1034289152 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034677031093279843, + "loss": 1.5198, + "theoretical_loss": 3.665984310423574, + "tokens_seen": 1034354688 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003467602808425276, + "loss": 1.4181, + "theoretical_loss": 3.6659625148596002, + "tokens_seen": 1034420224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003467502507522568, + "loss": 1.4341, + "theoretical_loss": 3.665940721063061, + "tokens_seen": 1034485760 + }, + { + "epoch": 0.31, + "learning_rate": 0.000346740220661986, + "loss": 1.5485, + "theoretical_loss": 3.6659189290337006, + "tokens_seen": 1034551296 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034673019057171516, + "loss": 1.467, + "theoretical_loss": 3.665897138771265, + "tokens_seen": 1034616832 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034672016048144434, + "loss": 1.3374, + "theoretical_loss": 3.6658753502754986, + "tokens_seen": 1034682368 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034671013039117357, + "loss": 1.4046, + "theoretical_loss": 3.665853563546146, + "tokens_seen": 1034747904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003467001003009027, + "loss": 1.4734, + "theoretical_loss": 3.6658317785829526, + "tokens_seen": 1034813440 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034669007021063193, + "loss": 1.4541, + "theoretical_loss": 3.665809995385663, + "tokens_seen": 1034878976 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034668004012036106, + "loss": 1.4436, + "theoretical_loss": 3.665788213954022, + "tokens_seen": 1034944512 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003466700100300903, + "loss": 1.4805, + "theoretical_loss": 3.6657664342877756, + "tokens_seen": 1035010048 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003466599799398195, + "loss": 1.4356, + "theoretical_loss": 3.665744656386668, + "tokens_seen": 1035075584 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034664994984954866, + "loss": 1.4552, + "theoretical_loss": 3.6657228802504447, + "tokens_seen": 1035141120 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034663991975927784, + "loss": 1.4436, + "theoretical_loss": 3.6657011058788513, + "tokens_seen": 1035206656 + }, + { + "epoch": 0.31, + "learning_rate": 0.000346629889669007, + "loss": 1.3698, + "theoretical_loss": 3.665679333271633, + "tokens_seen": 1035272192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003466198595787362, + "loss": 1.6126, + "theoretical_loss": 3.6656575624285344, + "tokens_seen": 1035337728 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034660982948846544, + "loss": 1.4804, + "theoretical_loss": 3.665635793349302, + "tokens_seen": 1035403264 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 379908, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4578062295913696, + "objective/train/theoretical_loss": 3.6656140260336807, + "objective/train/tokens_used": 1055928800, + "theoretical_loss": 3.6656140260336807, + "tokens_seen": 1035468800 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034659979939819456, + "loss": 1.4135, + "theoretical_loss": 3.6656140260336807, + "tokens_seen": 1035468800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003465897693079238, + "loss": 1.4023, + "theoretical_loss": 3.665592260481416, + "tokens_seen": 1035534336 + }, + { + "epoch": 0.31, + "learning_rate": 0.000346579739217653, + "loss": 1.3236, + "theoretical_loss": 3.6655704966922533, + "tokens_seen": 1035599872 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034656970912738216, + "loss": 1.4413, + "theoretical_loss": 3.665548734665939, + "tokens_seen": 1035665408 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034655967903711134, + "loss": 1.4937, + "theoretical_loss": 3.6655269744022174, + "tokens_seen": 1035730944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003465496489468405, + "loss": 1.4747, + "theoretical_loss": 3.665505215900836, + "tokens_seen": 1035796480 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003465396188565697, + "loss": 1.4867, + "theoretical_loss": 3.665483459161539, + "tokens_seen": 1035862016 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034652958876629894, + "loss": 1.4207, + "theoretical_loss": 3.6654617041840734, + "tokens_seen": 1035927552 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034651955867602807, + "loss": 1.5081, + "theoretical_loss": 3.6654399509681843, + "tokens_seen": 1035993088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003465095285857573, + "loss": 1.4605, + "theoretical_loss": 3.665418199513618, + "tokens_seen": 1036058624 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034649949849548643, + "loss": 1.4243, + "theoretical_loss": 3.6653964498201197, + "tokens_seen": 1036124160 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034648946840521567, + "loss": 1.304, + "theoretical_loss": 3.665374701887437, + "tokens_seen": 1036189696 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034647943831494485, + "loss": 1.5204, + "theoretical_loss": 3.6653529557153144, + "tokens_seen": 1036255232 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034646940822467403, + "loss": 1.4107, + "theoretical_loss": 3.6653312113034993, + "tokens_seen": 1036320768 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003464593781344032, + "loss": 1.3679, + "theoretical_loss": 3.665309468651737, + "tokens_seen": 1036386304 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034644934804413244, + "loss": 1.4977, + "theoretical_loss": 3.6652877277597744, + "tokens_seen": 1036451840 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034643931795386157, + "loss": 1.5167, + "theoretical_loss": 3.6652659886273575, + "tokens_seen": 1036517376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003464292878635908, + "loss": 1.4033, + "theoretical_loss": 3.6652442512542325, + "tokens_seen": 1036582912 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034641925777331993, + "loss": 1.4691, + "theoretical_loss": 3.6652225156401457, + "tokens_seen": 1036648448 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034640922768304917, + "loss": 1.5206, + "theoretical_loss": 3.6652007817848444, + "tokens_seen": 1036713984 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034639919759277835, + "loss": 1.4236, + "theoretical_loss": 3.6651790496880747, + "tokens_seen": 1036779520 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034638916750250753, + "loss": 1.5234, + "theoretical_loss": 3.6651573193495826, + "tokens_seen": 1036845056 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003463791374122367, + "loss": 1.4054, + "theoretical_loss": 3.665135590769115, + "tokens_seen": 1036910592 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003463691073219659, + "loss": 1.4835, + "theoretical_loss": 3.665113863946419, + "tokens_seen": 1036976128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003463590772316951, + "loss": 1.4747, + "theoretical_loss": 3.6650921388812416, + "tokens_seen": 1037041664 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003463490471414243, + "loss": 1.4026, + "theoretical_loss": 3.6650704155733287, + "tokens_seen": 1037107200 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034633901705115344, + "loss": 1.5397, + "theoretical_loss": 3.6650486940224267, + "tokens_seen": 1037172736 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034632898696088267, + "loss": 1.5448, + "theoretical_loss": 3.6650269742282844, + "tokens_seen": 1037238272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003463189568706118, + "loss": 1.3709, + "theoretical_loss": 3.6650052561906468, + "tokens_seen": 1037303808 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034630892678034103, + "loss": 1.5112, + "theoretical_loss": 3.664983539909262, + "tokens_seen": 1037369344 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462988966900702, + "loss": 1.4997, + "theoretical_loss": 3.6649618253838767, + "tokens_seen": 1037434880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462888665997994, + "loss": 1.2741, + "theoretical_loss": 3.664940112614238, + "tokens_seen": 1037500416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462788365095286, + "loss": 1.4184, + "theoretical_loss": 3.664918401600093, + "tokens_seen": 1037565952 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462688064192578, + "loss": 1.3634, + "theoretical_loss": 3.664896692341189, + "tokens_seen": 1037631488 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034625877632898694, + "loss": 1.4614, + "theoretical_loss": 3.6648749848372733, + "tokens_seen": 1037697024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462487462387162, + "loss": 1.4311, + "theoretical_loss": 3.664853279088093, + "tokens_seen": 1037762560 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462387161484453, + "loss": 1.3876, + "theoretical_loss": 3.6648315750933955, + "tokens_seen": 1037828096 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034622868605817454, + "loss": 1.392, + "theoretical_loss": 3.664809872852928, + "tokens_seen": 1037893632 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462186559679037, + "loss": 1.4437, + "theoretical_loss": 3.6647881723664386, + "tokens_seen": 1037959168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003462086258776329, + "loss": 1.3731, + "theoretical_loss": 3.6647664736336747, + "tokens_seen": 1038024704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003461985957873621, + "loss": 1.4359, + "theoretical_loss": 3.6647447766543833, + "tokens_seen": 1038090240 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034618856569709126, + "loss": 1.423, + "theoretical_loss": 3.664723081428312, + "tokens_seen": 1038155776 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034617853560682044, + "loss": 1.3386, + "theoretical_loss": 3.6647013879552093, + "tokens_seen": 1038221312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003461685055165497, + "loss": 1.4893, + "theoretical_loss": 3.664679696234822, + "tokens_seen": 1038286848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003461584754262788, + "loss": 1.4499, + "theoretical_loss": 3.6646580062668983, + "tokens_seen": 1038352384 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034614844533600804, + "loss": 1.43, + "theoretical_loss": 3.664636318051186, + "tokens_seen": 1038417920 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034613841524573717, + "loss": 1.346, + "theoretical_loss": 3.6646146315874333, + "tokens_seen": 1038483456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003461283851554664, + "loss": 1.3031, + "theoretical_loss": 3.6645929468753877, + "tokens_seen": 1038548992 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003461183550651956, + "loss": 1.5257, + "theoretical_loss": 3.664571263914797, + "tokens_seen": 1038614528 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034610832497492476, + "loss": 1.4466, + "theoretical_loss": 3.66454958270541, + "tokens_seen": 1038680064 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 381080, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2649497985839844, + "objective/train/theoretical_loss": 3.6645279032469746, + "objective/train/tokens_used": 1059205600, + "theoretical_loss": 3.6645279032469746, + "tokens_seen": 1038745600 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034609829488465395, + "loss": 1.3717, + "theoretical_loss": 3.6645279032469746, + "tokens_seen": 1038745600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003460882647943832, + "loss": 1.4507, + "theoretical_loss": 3.664506225539238, + "tokens_seen": 1038811136 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034607823470411236, + "loss": 1.4484, + "theoretical_loss": 3.664484549581949, + "tokens_seen": 1038876672 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034606820461384154, + "loss": 1.4958, + "theoretical_loss": 3.6644628753748565, + "tokens_seen": 1038942208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003460581745235707, + "loss": 1.417, + "theoretical_loss": 3.6644412029177076, + "tokens_seen": 1039007744 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003460481444332999, + "loss": 1.3637, + "theoretical_loss": 3.6644195322102515, + "tokens_seen": 1039073280 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034603811434302914, + "loss": 1.4431, + "theoretical_loss": 3.664397863252236, + "tokens_seen": 1039138816 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034602808425275827, + "loss": 1.3876, + "theoretical_loss": 3.6643761960434107, + "tokens_seen": 1039204352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0003460180541624875, + "loss": 1.4623, + "theoretical_loss": 3.6643545305835223, + "tokens_seen": 1039269888 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034600802407221663, + "loss": 1.4763, + "theoretical_loss": 3.664332866872321, + "tokens_seen": 1039335424 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034599799398194587, + "loss": 1.4043, + "theoretical_loss": 3.6643112049095543, + "tokens_seen": 1039400960 + }, + { + "epoch": 0.31, + "learning_rate": 0.00034598796389167505, + "loss": 1.4111, + "theoretical_loss": 3.6642895446949715, + "tokens_seen": 1039466496 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034597793380140423, + "loss": 1.4889, + "theoretical_loss": 3.664267886228321, + "tokens_seen": 1039532032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003459679037111334, + "loss": 1.4662, + "theoretical_loss": 3.6642462295093523, + "tokens_seen": 1039597568 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034595787362086264, + "loss": 1.3469, + "theoretical_loss": 3.664224574537813, + "tokens_seen": 1039663104 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034594784353059177, + "loss": 1.3727, + "theoretical_loss": 3.6642029213134526, + "tokens_seen": 1039728640 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345937813440321, + "loss": 1.4425, + "theoretical_loss": 3.66418126983602, + "tokens_seen": 1039794176 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034592778335005013, + "loss": 1.4043, + "theoretical_loss": 3.6641596201052637, + "tokens_seen": 1039859712 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034591775325977937, + "loss": 1.3934, + "theoretical_loss": 3.6641379721209337, + "tokens_seen": 1039925248 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034590772316950855, + "loss": 1.4013, + "theoretical_loss": 3.6641163258827785, + "tokens_seen": 1039990784 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034589769307923773, + "loss": 1.4475, + "theoretical_loss": 3.664094681390547, + "tokens_seen": 1040056320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003458876629889669, + "loss": 1.4221, + "theoretical_loss": 3.6640730386439895, + "tokens_seen": 1040121856 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003458776328986961, + "loss": 1.4125, + "theoretical_loss": 3.6640513976428535, + "tokens_seen": 1040187392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003458676028084253, + "loss": 1.4926, + "theoretical_loss": 3.6640297583868895, + "tokens_seen": 1040252928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003458575727181545, + "loss": 1.33, + "theoretical_loss": 3.664008120875846, + "tokens_seen": 1040318464 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034584754262788364, + "loss": 1.5571, + "theoretical_loss": 3.663986485109474, + "tokens_seen": 1040384000 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034583751253761287, + "loss": 1.4828, + "theoretical_loss": 3.6639648510875205, + "tokens_seen": 1040449536 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345827482447342, + "loss": 1.4116, + "theoretical_loss": 3.663943218809737, + "tokens_seen": 1040515072 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034581745235707123, + "loss": 1.375, + "theoretical_loss": 3.6639215882758727, + "tokens_seen": 1040580608 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003458074222668004, + "loss": 1.444, + "theoretical_loss": 3.663899959485676, + "tokens_seen": 1040646144 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457973921765296, + "loss": 1.4891, + "theoretical_loss": 3.6638783324388977, + "tokens_seen": 1040711680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457873620862588, + "loss": 1.4672, + "theoretical_loss": 3.663856707135287, + "tokens_seen": 1040777216 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345777331995988, + "loss": 1.511, + "theoretical_loss": 3.6638350835745945, + "tokens_seen": 1040842752 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034576730190571714, + "loss": 1.5512, + "theoretical_loss": 3.663813461756569, + "tokens_seen": 1040908288 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457572718154464, + "loss": 1.3771, + "theoretical_loss": 3.6637918416809603, + "tokens_seen": 1040973824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457472417251755, + "loss": 1.4296, + "theoretical_loss": 3.663770223347519, + "tokens_seen": 1041039360 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034573721163490474, + "loss": 1.4415, + "theoretical_loss": 3.663748606755995, + "tokens_seen": 1041104896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457271815446339, + "loss": 1.4644, + "theoretical_loss": 3.6637269919061373, + "tokens_seen": 1041170432 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457171514543631, + "loss": 1.3874, + "theoretical_loss": 3.6637053787976974, + "tokens_seen": 1041235968 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003457071213640923, + "loss": 1.4133, + "theoretical_loss": 3.663683767430424, + "tokens_seen": 1041301504 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034569709127382146, + "loss": 1.359, + "theoretical_loss": 3.6636621578040685, + "tokens_seen": 1041367040 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034568706118355064, + "loss": 1.3701, + "theoretical_loss": 3.66364054991838, + "tokens_seen": 1041432576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003456770310932799, + "loss": 1.3697, + "theoretical_loss": 3.66361894377311, + "tokens_seen": 1041498112 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345667001003009, + "loss": 1.3915, + "theoretical_loss": 3.6635973393680077, + "tokens_seen": 1041563648 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034565697091273824, + "loss": 1.3316, + "theoretical_loss": 3.6635757367028248, + "tokens_seen": 1041629184 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034564694082246737, + "loss": 1.3883, + "theoretical_loss": 3.66355413577731, + "tokens_seen": 1041694720 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003456369107321966, + "loss": 1.5111, + "theoretical_loss": 3.6635325365912146, + "tokens_seen": 1041760256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003456268806419258, + "loss": 1.3871, + "theoretical_loss": 3.663510939144289, + "tokens_seen": 1041825792 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034561685055165496, + "loss": 1.4343, + "theoretical_loss": 3.6634893434362845, + "tokens_seen": 1041891328 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034560682046138415, + "loss": 1.3699, + "theoretical_loss": 3.6634677494669505, + "tokens_seen": 1041956864 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 382133, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4042109251022339, + "objective/train/theoretical_loss": 3.663446157236039, + "objective/train/tokens_used": 1062482400, + "theoretical_loss": 3.663446157236039, + "tokens_seen": 1042022400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003455967903711134, + "loss": 1.3851, + "theoretical_loss": 3.663446157236039, + "tokens_seen": 1042022400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003455867602808425, + "loss": 1.4971, + "theoretical_loss": 3.6634245667432994, + "tokens_seen": 1042087936 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034557673019057174, + "loss": 1.4504, + "theoretical_loss": 3.6634029779884836, + "tokens_seen": 1042153472 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034556670010030087, + "loss": 1.5036, + "theoretical_loss": 3.6633813909713417, + "tokens_seen": 1042219008 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003455566700100301, + "loss": 1.3989, + "theoretical_loss": 3.6633598056916243, + "tokens_seen": 1042284544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003455466399197593, + "loss": 1.4396, + "theoretical_loss": 3.6633382221490836, + "tokens_seen": 1042350080 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034553660982948847, + "loss": 1.4231, + "theoretical_loss": 3.66331664034347, + "tokens_seen": 1042415616 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034552657973921765, + "loss": 1.3441, + "theoretical_loss": 3.6632950602745336, + "tokens_seen": 1042481152 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034551654964894683, + "loss": 1.474, + "theoretical_loss": 3.663273481942026, + "tokens_seen": 1042546688 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345506519558676, + "loss": 1.4425, + "theoretical_loss": 3.6632519053456996, + "tokens_seen": 1042612224 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034549648946840525, + "loss": 1.4178, + "theoretical_loss": 3.6632303304853044, + "tokens_seen": 1042677760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003454864593781344, + "loss": 1.4744, + "theoretical_loss": 3.663208757360591, + "tokens_seen": 1042743296 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003454764292878636, + "loss": 1.3984, + "theoretical_loss": 3.6631871859713123, + "tokens_seen": 1042808832 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034546639919759274, + "loss": 1.3597, + "theoretical_loss": 3.663165616317219, + "tokens_seen": 1042874368 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034545636910732197, + "loss": 1.3341, + "theoretical_loss": 3.6631440483980615, + "tokens_seen": 1042939904 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034544633901705115, + "loss": 1.3546, + "theoretical_loss": 3.663122482213593, + "tokens_seen": 1043005440 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034543630892678033, + "loss": 1.416, + "theoretical_loss": 3.6631009177635634, + "tokens_seen": 1043070976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003454262788365095, + "loss": 1.3953, + "theoretical_loss": 3.663079355047725, + "tokens_seen": 1043136512 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034541624874623875, + "loss": 1.4406, + "theoretical_loss": 3.6630577940658298, + "tokens_seen": 1043202048 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003454062186559679, + "loss": 1.3804, + "theoretical_loss": 3.6630362348176284, + "tokens_seen": 1043267584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003453961885656971, + "loss": 1.5529, + "theoretical_loss": 3.663014677302874, + "tokens_seen": 1043333120 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034538615847542624, + "loss": 1.5203, + "theoretical_loss": 3.6629931215213167, + "tokens_seen": 1043398656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003453761283851555, + "loss": 1.3293, + "theoretical_loss": 3.6629715674727095, + "tokens_seen": 1043464192 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034536609829488466, + "loss": 1.5306, + "theoretical_loss": 3.6629500151568033, + "tokens_seen": 1043529728 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034535606820461384, + "loss": 1.3278, + "theoretical_loss": 3.6629284645733504, + "tokens_seen": 1043595264 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345346038114343, + "loss": 1.4744, + "theoretical_loss": 3.662906915722103, + "tokens_seen": 1043660800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003453360080240722, + "loss": 1.5606, + "theoretical_loss": 3.6628853686028133, + "tokens_seen": 1043726336 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034532597793380143, + "loss": 1.4019, + "theoretical_loss": 3.662863823215233, + "tokens_seen": 1043791872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003453159478435306, + "loss": 1.4387, + "theoretical_loss": 3.662842279559114, + "tokens_seen": 1043857408 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003453059177532598, + "loss": 1.4306, + "theoretical_loss": 3.6628207376342092, + "tokens_seen": 1043922944 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345295887662989, + "loss": 1.483, + "theoretical_loss": 3.66279919744027, + "tokens_seen": 1043988480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452858575727182, + "loss": 1.4548, + "theoretical_loss": 3.6627776589770487, + "tokens_seen": 1044054016 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034527582748244734, + "loss": 1.394, + "theoretical_loss": 3.662756122244298, + "tokens_seen": 1044119552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452657973921766, + "loss": 1.5332, + "theoretical_loss": 3.6627345872417703, + "tokens_seen": 1044185088 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452557673019057, + "loss": 1.5864, + "theoretical_loss": 3.662713053969218, + "tokens_seen": 1044250624 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034524573721163494, + "loss": 1.3968, + "theoretical_loss": 3.6626915224263934, + "tokens_seen": 1044316160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452357071213641, + "loss": 1.4855, + "theoretical_loss": 3.662669992613049, + "tokens_seen": 1044381696 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452256770310933, + "loss": 1.4821, + "theoretical_loss": 3.6626484645289374, + "tokens_seen": 1044447232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003452156469408225, + "loss": 1.4276, + "theoretical_loss": 3.662626938173811, + "tokens_seen": 1044512768 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034520561685055166, + "loss": 1.4139, + "theoretical_loss": 3.662605413547423, + "tokens_seen": 1044578304 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034519558676028084, + "loss": 1.2742, + "theoretical_loss": 3.6625838906495263, + "tokens_seen": 1044643840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003451855566700101, + "loss": 1.3665, + "theoretical_loss": 3.6625623694798723, + "tokens_seen": 1044709376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003451755265797392, + "loss": 1.6004, + "theoretical_loss": 3.6625408500382153, + "tokens_seen": 1044774912 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034516549648946844, + "loss": 1.4969, + "theoretical_loss": 3.6625193323243073, + "tokens_seen": 1044840448 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034515546639919757, + "loss": 1.4526, + "theoretical_loss": 3.6624978163379014, + "tokens_seen": 1044905984 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003451454363089268, + "loss": 1.349, + "theoretical_loss": 3.662476302078751, + "tokens_seen": 1044971520 + }, + { + "epoch": 0.32, + "learning_rate": 0.000345135406218656, + "loss": 1.3516, + "theoretical_loss": 3.6624547895466093, + "tokens_seen": 1045037056 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034512537612838517, + "loss": 1.4518, + "theoretical_loss": 3.6624332787412284, + "tokens_seen": 1045102592 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034511534603811435, + "loss": 1.4197, + "theoretical_loss": 3.662411769662362, + "tokens_seen": 1045168128 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003451053159478436, + "loss": 1.4913, + "theoretical_loss": 3.6623902623097635, + "tokens_seen": 1045233664 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 383342, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4441049098968506, + "objective/train/theoretical_loss": 3.6623687566831857, + "objective/train/tokens_used": 1065759200, + "theoretical_loss": 3.6623687566831857, + "tokens_seen": 1045299200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003450952858575727, + "loss": 1.4771, + "theoretical_loss": 3.6623687566831857, + "tokens_seen": 1045299200 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034508525576730194, + "loss": 1.4645, + "theoretical_loss": 3.662347252782382, + "tokens_seen": 1045364736 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034507522567703107, + "loss": 1.3817, + "theoretical_loss": 3.6623257506071054, + "tokens_seen": 1045430272 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003450651955867603, + "loss": 1.4438, + "theoretical_loss": 3.66230425015711, + "tokens_seen": 1045495808 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003450551654964895, + "loss": 1.3566, + "theoretical_loss": 3.662282751432149, + "tokens_seen": 1045561344 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034504513540621867, + "loss": 1.4901, + "theoretical_loss": 3.662261254431976, + "tokens_seen": 1045626880 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034503510531594785, + "loss": 1.465, + "theoretical_loss": 3.6622397591563436, + "tokens_seen": 1045692416 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034502507522567703, + "loss": 1.398, + "theoretical_loss": 3.6622182656050066, + "tokens_seen": 1045757952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003450150451354062, + "loss": 1.4561, + "theoretical_loss": 3.662196773777718, + "tokens_seen": 1045823488 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034500501504513545, + "loss": 1.4178, + "theoretical_loss": 3.6621752836742316, + "tokens_seen": 1045889024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003449949849548646, + "loss": 1.4513, + "theoretical_loss": 3.662153795294301, + "tokens_seen": 1045954560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003449849548645938, + "loss": 1.4099, + "theoretical_loss": 3.6621323086376805, + "tokens_seen": 1046020096 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034497492477432294, + "loss": 1.4419, + "theoretical_loss": 3.6621108237041233, + "tokens_seen": 1046085632 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034496489468405217, + "loss": 1.3125, + "theoretical_loss": 3.6620893404933836, + "tokens_seen": 1046151168 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034495486459378135, + "loss": 1.5077, + "theoretical_loss": 3.6620678590052154, + "tokens_seen": 1046216704 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034494483450351053, + "loss": 1.3783, + "theoretical_loss": 3.6620463792393725, + "tokens_seen": 1046282240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003449348044132397, + "loss": 1.4203, + "theoretical_loss": 3.6620249011956094, + "tokens_seen": 1046347776 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034492477432296895, + "loss": 1.4408, + "theoretical_loss": 3.662003424873679, + "tokens_seen": 1046413312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003449147442326981, + "loss": 1.4872, + "theoretical_loss": 3.6619819502733373, + "tokens_seen": 1046478848 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003449047141424273, + "loss": 1.3363, + "theoretical_loss": 3.6619604773943366, + "tokens_seen": 1046544384 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034489468405215644, + "loss": 1.348, + "theoretical_loss": 3.661939006236432, + "tokens_seen": 1046609920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448846539618857, + "loss": 1.4919, + "theoretical_loss": 3.661917536799378, + "tokens_seen": 1046675456 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034487462387161486, + "loss": 1.4722, + "theoretical_loss": 3.661896069082929, + "tokens_seen": 1046740992 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034486459378134404, + "loss": 1.5347, + "theoretical_loss": 3.661874603086839, + "tokens_seen": 1046806528 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448545636910732, + "loss": 1.4131, + "theoretical_loss": 3.661853138810863, + "tokens_seen": 1046872064 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448445336008024, + "loss": 1.3818, + "theoretical_loss": 3.6618316762547543, + "tokens_seen": 1046937600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448345035105316, + "loss": 1.5068, + "theoretical_loss": 3.6618102154182686, + "tokens_seen": 1047003136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448244734202608, + "loss": 1.3943, + "theoretical_loss": 3.66178875630116, + "tokens_seen": 1047068672 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034481444332998994, + "loss": 1.4068, + "theoretical_loss": 3.6617672989031833, + "tokens_seen": 1047134208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003448044132397192, + "loss": 1.5136, + "theoretical_loss": 3.661745843224093, + "tokens_seen": 1047199744 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034479438314944836, + "loss": 1.3288, + "theoretical_loss": 3.6617243892636444, + "tokens_seen": 1047265280 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034478435305917754, + "loss": 1.3672, + "theoretical_loss": 3.6617029370215914, + "tokens_seen": 1047330816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447743229689067, + "loss": 1.3664, + "theoretical_loss": 3.66168148649769, + "tokens_seen": 1047396352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447642928786359, + "loss": 1.4558, + "theoretical_loss": 3.6616600376916937, + "tokens_seen": 1047461888 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447542627883651, + "loss": 1.4151, + "theoretical_loss": 3.6616385906033586, + "tokens_seen": 1047527424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447442326980943, + "loss": 1.3846, + "theoretical_loss": 3.661617145232439, + "tokens_seen": 1047592960 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034473420260782345, + "loss": 1.5665, + "theoretical_loss": 3.6615957015786904, + "tokens_seen": 1047658496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447241725175527, + "loss": 1.4411, + "theoretical_loss": 3.661574259641868, + "tokens_seen": 1047724032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003447141424272818, + "loss": 1.4061, + "theoretical_loss": 3.6615528194217264, + "tokens_seen": 1047789568 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034470411233701104, + "loss": 1.3655, + "theoretical_loss": 3.661531380918021, + "tokens_seen": 1047855104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446940822467402, + "loss": 1.3914, + "theoretical_loss": 3.6615099441305077, + "tokens_seen": 1047920640 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446840521564694, + "loss": 1.2793, + "theoretical_loss": 3.6614885090589406, + "tokens_seen": 1047986176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446740220661986, + "loss": 1.5174, + "theoretical_loss": 3.661467075703076, + "tokens_seen": 1048051712 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034466399197592777, + "loss": 1.4102, + "theoretical_loss": 3.6614456440626686, + "tokens_seen": 1048117248 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034465396188565695, + "loss": 1.4187, + "theoretical_loss": 3.6614242141374747, + "tokens_seen": 1048182784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446439317953862, + "loss": 1.4986, + "theoretical_loss": 3.6614027859272493, + "tokens_seen": 1048248320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446339017051153, + "loss": 1.3471, + "theoretical_loss": 3.6613813594317475, + "tokens_seen": 1048313856 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034462387161484455, + "loss": 1.4119, + "theoretical_loss": 3.6613599346507257, + "tokens_seen": 1048379392 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034461384152457373, + "loss": 1.3835, + "theoretical_loss": 3.6613385115839394, + "tokens_seen": 1048444928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003446038114343029, + "loss": 1.4206, + "theoretical_loss": 3.661317090231144, + "tokens_seen": 1048510464 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 384505, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5295982360839844, + "objective/train/theoretical_loss": 3.661295670592095, + "objective/train/tokens_used": 1069036000, + "theoretical_loss": 3.661295670592095, + "tokens_seen": 1048576000 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034459378134403214, + "loss": 1.418, + "theoretical_loss": 3.661295670592095, + "tokens_seen": 1048576000 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034458375125376127, + "loss": 1.3328, + "theoretical_loss": 3.661274252666549, + "tokens_seen": 1048641536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003445737211634905, + "loss": 1.411, + "theoretical_loss": 3.6612528364542616, + "tokens_seen": 1048707072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003445636910732197, + "loss": 1.4973, + "theoretical_loss": 3.661231421954988, + "tokens_seen": 1048772608 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034455366098294887, + "loss": 1.4881, + "theoretical_loss": 3.661210009168485, + "tokens_seen": 1048838144 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034454363089267805, + "loss": 1.3173, + "theoretical_loss": 3.6611885980945083, + "tokens_seen": 1048903680 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034453360080240723, + "loss": 1.4759, + "theoretical_loss": 3.661167188732814, + "tokens_seen": 1048969216 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003445235707121364, + "loss": 1.38, + "theoretical_loss": 3.661145781083158, + "tokens_seen": 1049034752 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034451354062186565, + "loss": 1.4175, + "theoretical_loss": 3.6611243751452966, + "tokens_seen": 1049100288 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003445035105315948, + "loss": 1.3999, + "theoretical_loss": 3.661102970918986, + "tokens_seen": 1049165824 + }, + { + "epoch": 0.32, + "learning_rate": 0.000344493480441324, + "loss": 1.3994, + "theoretical_loss": 3.661081568403983, + "tokens_seen": 1049231360 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034448345035105314, + "loss": 1.4193, + "theoretical_loss": 3.6610601676000423, + "tokens_seen": 1049296896 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034447342026078237, + "loss": 1.5142, + "theoretical_loss": 3.6610387685069217, + "tokens_seen": 1049362432 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034446339017051155, + "loss": 1.4514, + "theoretical_loss": 3.6610173711243776, + "tokens_seen": 1049427968 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034445336008024073, + "loss": 1.4721, + "theoretical_loss": 3.660995975452166, + "tokens_seen": 1049493504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003444433299899699, + "loss": 1.5066, + "theoretical_loss": 3.660974581490043, + "tokens_seen": 1049559040 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034443329989969915, + "loss": 1.4259, + "theoretical_loss": 3.660953189237766, + "tokens_seen": 1049624576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003444232698094283, + "loss": 1.3698, + "theoretical_loss": 3.660931798695091, + "tokens_seen": 1049690112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003444132397191575, + "loss": 1.4315, + "theoretical_loss": 3.6609104098617746, + "tokens_seen": 1049755648 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034440320962888664, + "loss": 1.2695, + "theoretical_loss": 3.6608890227375737, + "tokens_seen": 1049821184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003443931795386159, + "loss": 1.4549, + "theoretical_loss": 3.6608676373222457, + "tokens_seen": 1049886720 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034438314944834506, + "loss": 1.4513, + "theoretical_loss": 3.6608462536155466, + "tokens_seen": 1049952256 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034437311935807424, + "loss": 1.4674, + "theoretical_loss": 3.6608248716172334, + "tokens_seen": 1050017792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003443630892678034, + "loss": 1.3996, + "theoretical_loss": 3.660803491327063, + "tokens_seen": 1050083328 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003443530591775326, + "loss": 1.441, + "theoretical_loss": 3.6607821127447924, + "tokens_seen": 1050148864 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003443430290872618, + "loss": 1.4166, + "theoretical_loss": 3.6607607358701784, + "tokens_seen": 1050214400 + }, + { + "epoch": 0.32, + "learning_rate": 0.000344332998996991, + "loss": 1.5652, + "theoretical_loss": 3.660739360702978, + "tokens_seen": 1050279936 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034432296890672014, + "loss": 1.5007, + "theoretical_loss": 3.660717987242949, + "tokens_seen": 1050345472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003443129388164494, + "loss": 1.3855, + "theoretical_loss": 3.660696615489848, + "tokens_seen": 1050411008 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034430290872617856, + "loss": 1.4021, + "theoretical_loss": 3.6606752454434326, + "tokens_seen": 1050476544 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034429287863590774, + "loss": 1.438, + "theoretical_loss": 3.660653877103459, + "tokens_seen": 1050542080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442828485456369, + "loss": 1.4038, + "theoretical_loss": 3.660632510469686, + "tokens_seen": 1050607616 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442728184553661, + "loss": 1.4136, + "theoretical_loss": 3.66061114554187, + "tokens_seen": 1050673152 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442627883650953, + "loss": 1.3386, + "theoretical_loss": 3.6605897823197684, + "tokens_seen": 1050738688 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442527582748245, + "loss": 1.5029, + "theoretical_loss": 3.6605684208031386, + "tokens_seen": 1050804224 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034424272818455365, + "loss": 1.5489, + "theoretical_loss": 3.660547060991739, + "tokens_seen": 1050869760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442326980942829, + "loss": 1.5231, + "theoretical_loss": 3.6605257028853257, + "tokens_seen": 1050935296 + }, + { + "epoch": 0.32, + "learning_rate": 0.000344222668004012, + "loss": 1.4631, + "theoretical_loss": 3.6605043464836573, + "tokens_seen": 1051000832 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034421263791374124, + "loss": 1.3766, + "theoretical_loss": 3.660482991786491, + "tokens_seen": 1051066368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003442026078234704, + "loss": 1.4543, + "theoretical_loss": 3.660461638793585, + "tokens_seen": 1051131904 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441925777331996, + "loss": 1.5175, + "theoretical_loss": 3.660440287504697, + "tokens_seen": 1051197440 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441825476429288, + "loss": 1.4321, + "theoretical_loss": 3.6604189379195837, + "tokens_seen": 1051262976 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034417251755265797, + "loss": 1.443, + "theoretical_loss": 3.6603975900380044, + "tokens_seen": 1051328512 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034416248746238715, + "loss": 1.5078, + "theoretical_loss": 3.660376243859716, + "tokens_seen": 1051394048 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441524573721164, + "loss": 1.534, + "theoretical_loss": 3.660354899384477, + "tokens_seen": 1051459584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441424272818455, + "loss": 1.472, + "theoretical_loss": 3.6603335566120454, + "tokens_seen": 1051525120 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034413239719157475, + "loss": 1.5227, + "theoretical_loss": 3.660312215542179, + "tokens_seen": 1051590656 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034412236710130393, + "loss": 1.3936, + "theoretical_loss": 3.6602908761746358, + "tokens_seen": 1051656192 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441123370110331, + "loss": 1.5337, + "theoretical_loss": 3.6602695385091737, + "tokens_seen": 1051721728 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003441023069207623, + "loss": 1.4985, + "theoretical_loss": 3.660248202545551, + "tokens_seen": 1051787264 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 385698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4051916599273682, + "objective/train/theoretical_loss": 3.660226868283527, + "objective/train/tokens_used": 1072312800, + "theoretical_loss": 3.660226868283527, + "tokens_seen": 1051852800 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034409227683049147, + "loss": 1.3933, + "theoretical_loss": 3.660226868283527, + "tokens_seen": 1051852800 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034408224674022065, + "loss": 1.4988, + "theoretical_loss": 3.660205535722859, + "tokens_seen": 1051918336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003440722166499499, + "loss": 1.4607, + "theoretical_loss": 3.660184204863305, + "tokens_seen": 1051983872 + }, + { + "epoch": 0.32, + "learning_rate": 0.000344062186559679, + "loss": 1.3357, + "theoretical_loss": 3.6601628757046245, + "tokens_seen": 1052049408 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034405215646940825, + "loss": 1.4035, + "theoretical_loss": 3.660141548246575, + "tokens_seen": 1052114944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003440421263791374, + "loss": 1.3284, + "theoretical_loss": 3.6601202224889153, + "tokens_seen": 1052180480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003440320962888666, + "loss": 1.3376, + "theoretical_loss": 3.6600988984314045, + "tokens_seen": 1052246016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003440220661985958, + "loss": 1.405, + "theoretical_loss": 3.6600775760738, + "tokens_seen": 1052311552 + }, + { + "epoch": 0.32, + "learning_rate": 0.000344012036108325, + "loss": 1.2767, + "theoretical_loss": 3.660056255415861, + "tokens_seen": 1052377088 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034400200601805416, + "loss": 1.4854, + "theoretical_loss": 3.660034936457347, + "tokens_seen": 1052442624 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034399197592778334, + "loss": 1.3801, + "theoretical_loss": 3.660013619198015, + "tokens_seen": 1052508160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003439819458375125, + "loss": 1.4045, + "theoretical_loss": 3.659992303637625, + "tokens_seen": 1052573696 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034397191574724175, + "loss": 1.331, + "theoretical_loss": 3.6599709897759363, + "tokens_seen": 1052639232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003439618856569709, + "loss": 1.3891, + "theoretical_loss": 3.6599496776127065, + "tokens_seen": 1052704768 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003439518555667001, + "loss": 1.4033, + "theoretical_loss": 3.659928367147695, + "tokens_seen": 1052770304 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003439418254764293, + "loss": 1.3574, + "theoretical_loss": 3.6599070583806617, + "tokens_seen": 1052835840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003439317953861585, + "loss": 1.3759, + "theoretical_loss": 3.6598857513113634, + "tokens_seen": 1052901376 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034392176529588766, + "loss": 1.3112, + "theoretical_loss": 3.6598644459395615, + "tokens_seen": 1052966912 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034391173520561684, + "loss": 1.3703, + "theoretical_loss": 3.6598431422650144, + "tokens_seen": 1053032448 + }, + { + "epoch": 0.32, + "learning_rate": 0.000343901705115346, + "loss": 1.4473, + "theoretical_loss": 3.6598218402874805, + "tokens_seen": 1053097984 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034389167502507526, + "loss": 1.3557, + "theoretical_loss": 3.65980054000672, + "tokens_seen": 1053163520 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003438816449348044, + "loss": 1.3433, + "theoretical_loss": 3.6597792414224912, + "tokens_seen": 1053229056 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003438716148445336, + "loss": 1.5037, + "theoretical_loss": 3.6597579445345545, + "tokens_seen": 1053294592 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034386158475426275, + "loss": 1.4849, + "theoretical_loss": 3.6597366493426686, + "tokens_seen": 1053360128 + }, + { + "epoch": 0.32, + "learning_rate": 0.000343851554663992, + "loss": 1.6497, + "theoretical_loss": 3.6597153558465934, + "tokens_seen": 1053425664 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003438415245737212, + "loss": 1.4614, + "theoretical_loss": 3.6596940640460875, + "tokens_seen": 1053491200 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034383149448345034, + "loss": 1.4072, + "theoretical_loss": 3.6596727739409114, + "tokens_seen": 1053556736 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003438214643931796, + "loss": 1.4538, + "theoretical_loss": 3.6596514855308238, + "tokens_seen": 1053622272 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034381143430290876, + "loss": 1.4195, + "theoretical_loss": 3.6596301988155853, + "tokens_seen": 1053687808 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034380140421263794, + "loss": 1.4635, + "theoretical_loss": 3.659608913794955, + "tokens_seen": 1053753344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437913741223671, + "loss": 1.5024, + "theoretical_loss": 3.659587630468692, + "tokens_seen": 1053818880 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437813440320963, + "loss": 1.4726, + "theoretical_loss": 3.6595663488365573, + "tokens_seen": 1053884416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437713139418255, + "loss": 1.4755, + "theoretical_loss": 3.6595450688983098, + "tokens_seen": 1053949952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437612838515547, + "loss": 1.478, + "theoretical_loss": 3.6595237906537097, + "tokens_seen": 1054015488 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034375125376128385, + "loss": 1.3913, + "theoretical_loss": 3.659502514102517, + "tokens_seen": 1054081024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437412236710131, + "loss": 1.3154, + "theoretical_loss": 3.6594812392444913, + "tokens_seen": 1054146560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437311935807422, + "loss": 1.4508, + "theoretical_loss": 3.6594599660793934, + "tokens_seen": 1054212096 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034372116349047144, + "loss": 1.4835, + "theoretical_loss": 3.6594386946069823, + "tokens_seen": 1054277632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437111334002006, + "loss": 1.4451, + "theoretical_loss": 3.659417424827019, + "tokens_seen": 1054343168 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003437011033099298, + "loss": 1.4037, + "theoretical_loss": 3.6593961567392626, + "tokens_seen": 1054408704 + }, + { + "epoch": 0.32, + "learning_rate": 0.000343691073219659, + "loss": 1.3983, + "theoretical_loss": 3.6593748903434746, + "tokens_seen": 1054474240 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034368104312938817, + "loss": 1.3693, + "theoretical_loss": 3.6593536256394144, + "tokens_seen": 1054539776 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034367101303911735, + "loss": 1.4339, + "theoretical_loss": 3.6593323626268424, + "tokens_seen": 1054605312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003436609829488466, + "loss": 1.4531, + "theoretical_loss": 3.6593111013055193, + "tokens_seen": 1054670848 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003436509528585757, + "loss": 1.4552, + "theoretical_loss": 3.6592898416752053, + "tokens_seen": 1054736384 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034364092276830495, + "loss": 1.4057, + "theoretical_loss": 3.6592685837356607, + "tokens_seen": 1054801920 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034363089267803413, + "loss": 1.4836, + "theoretical_loss": 3.6592473274866464, + "tokens_seen": 1054867456 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003436208625877633, + "loss": 1.2539, + "theoretical_loss": 3.6592260729279222, + "tokens_seen": 1054932992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003436108324974925, + "loss": 1.4376, + "theoretical_loss": 3.6592048200592497, + "tokens_seen": 1054998528 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034360080240722167, + "loss": 1.4329, + "theoretical_loss": 3.6591835688803886, + "tokens_seen": 1055064064 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 386814, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3760838508605957, + "objective/train/theoretical_loss": 3.6591623193911, + "objective/train/tokens_used": 1075589600, + "theoretical_loss": 3.6591623193911, + "tokens_seen": 1055129600 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034359077231695085, + "loss": 1.2889, + "theoretical_loss": 3.6591623193911, + "tokens_seen": 1055129600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003435807422266801, + "loss": 1.3536, + "theoretical_loss": 3.6591410715911445, + "tokens_seen": 1055195136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003435707121364092, + "loss": 1.4453, + "theoretical_loss": 3.6591198254802837, + "tokens_seen": 1055260672 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034356068204613845, + "loss": 1.3592, + "theoretical_loss": 3.6590985810582772, + "tokens_seen": 1055326208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003435506519558676, + "loss": 1.394, + "theoretical_loss": 3.6590773383248862, + "tokens_seen": 1055391744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003435406218655968, + "loss": 1.4429, + "theoretical_loss": 3.6590560972798727, + "tokens_seen": 1055457280 + }, + { + "epoch": 0.32, + "learning_rate": 0.000343530591775326, + "loss": 1.4061, + "theoretical_loss": 3.6590348579229963, + "tokens_seen": 1055522816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003435205616850552, + "loss": 1.4638, + "theoretical_loss": 3.6590136202540187, + "tokens_seen": 1055588352 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034351053159478436, + "loss": 1.3613, + "theoretical_loss": 3.6589923842727012, + "tokens_seen": 1055653888 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034350050150451354, + "loss": 1.3665, + "theoretical_loss": 3.6589711499788042, + "tokens_seen": 1055719424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434904714142427, + "loss": 1.4496, + "theoretical_loss": 3.658949917372089, + "tokens_seen": 1055784960 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034348044132397195, + "loss": 1.4152, + "theoretical_loss": 3.6589286864523176, + "tokens_seen": 1055850496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434704112337011, + "loss": 1.4122, + "theoretical_loss": 3.658907457219251, + "tokens_seen": 1055916032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434603811434303, + "loss": 1.3926, + "theoretical_loss": 3.65888622967265, + "tokens_seen": 1055981568 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434503510531595, + "loss": 1.3879, + "theoretical_loss": 3.6588650038122763, + "tokens_seen": 1056047104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434403209628887, + "loss": 1.4266, + "theoretical_loss": 3.6588437796378916, + "tokens_seen": 1056112640 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034343029087261786, + "loss": 1.5467, + "theoretical_loss": 3.658822557149257, + "tokens_seen": 1056178176 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034342026078234704, + "loss": 1.4569, + "theoretical_loss": 3.658801336346134, + "tokens_seen": 1056243712 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003434102306920762, + "loss": 1.384, + "theoretical_loss": 3.658780117228284, + "tokens_seen": 1056309248 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034340020060180546, + "loss": 1.4192, + "theoretical_loss": 3.65875889979547, + "tokens_seen": 1056374784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433901705115346, + "loss": 1.5536, + "theoretical_loss": 3.658737684047451, + "tokens_seen": 1056440320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433801404212638, + "loss": 1.3701, + "theoretical_loss": 3.6587164699839914, + "tokens_seen": 1056505856 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034337011033099295, + "loss": 1.3835, + "theoretical_loss": 3.6586952576048515, + "tokens_seen": 1056571392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433600802407222, + "loss": 1.3764, + "theoretical_loss": 3.658674046909793, + "tokens_seen": 1056636928 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034335005015045136, + "loss": 1.3476, + "theoretical_loss": 3.658652837898579, + "tokens_seen": 1056702464 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034334002006018054, + "loss": 1.3559, + "theoretical_loss": 3.6586316305709703, + "tokens_seen": 1056768000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433299899699097, + "loss": 1.484, + "theoretical_loss": 3.658610424926729, + "tokens_seen": 1056833536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433199598796389, + "loss": 1.4052, + "theoretical_loss": 3.6585892209656175, + "tokens_seen": 1056899072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003433099297893681, + "loss": 1.4769, + "theoretical_loss": 3.6585680186873972, + "tokens_seen": 1056964608 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003432998996990973, + "loss": 1.3847, + "theoretical_loss": 3.658546818091831, + "tokens_seen": 1057030144 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034328986960882645, + "loss": 1.362, + "theoretical_loss": 3.658525619178681, + "tokens_seen": 1057095680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003432798395185557, + "loss": 1.4416, + "theoretical_loss": 3.658504421947708, + "tokens_seen": 1057161216 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034326980942828487, + "loss": 1.5128, + "theoretical_loss": 3.658483226398676, + "tokens_seen": 1057226752 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034325977933801405, + "loss": 1.4772, + "theoretical_loss": 3.6584620325313466, + "tokens_seen": 1057292288 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034324974924774323, + "loss": 1.3688, + "theoretical_loss": 3.6584408403454822, + "tokens_seen": 1057357824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003432397191574724, + "loss": 1.3144, + "theoretical_loss": 3.6584196498408454, + "tokens_seen": 1057423360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003432296890672016, + "loss": 1.3786, + "theoretical_loss": 3.658398461017198, + "tokens_seen": 1057488896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003432196589769308, + "loss": 1.474, + "theoretical_loss": 3.6583772738743026, + "tokens_seen": 1057554432 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034320962888665995, + "loss": 1.4727, + "theoretical_loss": 3.6583560884119226, + "tokens_seen": 1057619968 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431995987963892, + "loss": 1.438, + "theoretical_loss": 3.6583349046298195, + "tokens_seen": 1057685504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431895687061183, + "loss": 1.4195, + "theoretical_loss": 3.6583137225277564, + "tokens_seen": 1057751040 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034317953861584755, + "loss": 1.4462, + "theoretical_loss": 3.6582925421054964, + "tokens_seen": 1057816576 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034316950852557673, + "loss": 1.3321, + "theoretical_loss": 3.6582713633628017, + "tokens_seen": 1057882112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431594784353059, + "loss": 1.3444, + "theoretical_loss": 3.658250186299435, + "tokens_seen": 1057947648 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431494483450351, + "loss": 1.3966, + "theoretical_loss": 3.65822901091516, + "tokens_seen": 1058013184 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034313941825476433, + "loss": 1.3146, + "theoretical_loss": 3.658207837209738, + "tokens_seen": 1058078720 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034312938816449345, + "loss": 1.4197, + "theoretical_loss": 3.6581866651829333, + "tokens_seen": 1058144256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431193580742227, + "loss": 1.3354, + "theoretical_loss": 3.658165494834508, + "tokens_seen": 1058209792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003431093279839518, + "loss": 1.3986, + "theoretical_loss": 3.6581443261642264, + "tokens_seen": 1058275328 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034309929789368105, + "loss": 1.3771, + "theoretical_loss": 3.6581231591718497, + "tokens_seen": 1058340864 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 388023, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3644042015075684, + "objective/train/theoretical_loss": 3.6581019938571426, + "objective/train/tokens_used": 1078866400, + "theoretical_loss": 3.6581019938571426, + "tokens_seen": 1058406400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003430892678034103, + "loss": 1.3334, + "theoretical_loss": 3.6581019938571426, + "tokens_seen": 1058406400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003430792377131394, + "loss": 1.391, + "theoretical_loss": 3.658080830219867, + "tokens_seen": 1058471936 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034306920762286865, + "loss": 1.4906, + "theoretical_loss": 3.6580596682597877, + "tokens_seen": 1058537472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003430591775325978, + "loss": 1.4329, + "theoretical_loss": 3.6580385079766664, + "tokens_seen": 1058603008 + }, + { + "epoch": 0.32, + "learning_rate": 0.000343049147442327, + "loss": 1.4951, + "theoretical_loss": 3.658017349370267, + "tokens_seen": 1058668544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003430391173520562, + "loss": 1.3488, + "theoretical_loss": 3.6579961924403532, + "tokens_seen": 1058734080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003430290872617854, + "loss": 1.4284, + "theoretical_loss": 3.657975037186688, + "tokens_seen": 1058799616 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034301905717151456, + "loss": 1.4113, + "theoretical_loss": 3.657953883609035, + "tokens_seen": 1058865152 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034300902708124374, + "loss": 1.3699, + "theoretical_loss": 3.6579327317071573, + "tokens_seen": 1058930688 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429989969909729, + "loss": 1.4788, + "theoretical_loss": 3.6579115814808185, + "tokens_seen": 1058996224 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034298896690070215, + "loss": 1.4335, + "theoretical_loss": 3.6578904329297837, + "tokens_seen": 1059061760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429789368104313, + "loss": 1.6038, + "theoretical_loss": 3.6578692860538142, + "tokens_seen": 1059127296 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429689067201605, + "loss": 1.3737, + "theoretical_loss": 3.6578481408526753, + "tokens_seen": 1059192832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429588766298897, + "loss": 1.4862, + "theoretical_loss": 3.65782699732613, + "tokens_seen": 1059258368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429488465396189, + "loss": 1.5136, + "theoretical_loss": 3.657805855473942, + "tokens_seen": 1059323904 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034293881644934806, + "loss": 1.4215, + "theoretical_loss": 3.6577847152958762, + "tokens_seen": 1059389440 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034292878635907724, + "loss": 1.4215, + "theoretical_loss": 3.6577635767916954, + "tokens_seen": 1059454976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003429187562688064, + "loss": 1.3405, + "theoretical_loss": 3.657742439961164, + "tokens_seen": 1059520512 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034290872617853566, + "loss": 1.2649, + "theoretical_loss": 3.657721304804046, + "tokens_seen": 1059586048 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428986960882648, + "loss": 1.2961, + "theoretical_loss": 3.657700171320105, + "tokens_seen": 1059651584 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342888665997994, + "loss": 1.4577, + "theoretical_loss": 3.6576790395091052, + "tokens_seen": 1059717120 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034287863590772315, + "loss": 1.3754, + "theoretical_loss": 3.6576579093708115, + "tokens_seen": 1059782656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428686058174524, + "loss": 1.387, + "theoretical_loss": 3.657636780904987, + "tokens_seen": 1059848192 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034285857572718156, + "loss": 1.4783, + "theoretical_loss": 3.6576156541113964, + "tokens_seen": 1059913728 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034284854563691074, + "loss": 1.392, + "theoretical_loss": 3.657594528989804, + "tokens_seen": 1059979264 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428385155466399, + "loss": 1.2477, + "theoretical_loss": 3.6575734055399742, + "tokens_seen": 1060044800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428284854563691, + "loss": 1.4264, + "theoretical_loss": 3.657552283761671, + "tokens_seen": 1060110336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428184553660983, + "loss": 1.3341, + "theoretical_loss": 3.6575311636546584, + "tokens_seen": 1060175872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003428084252758275, + "loss": 1.3514, + "theoretical_loss": 3.6575100452187024, + "tokens_seen": 1060241408 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034279839518555665, + "loss": 1.4196, + "theoretical_loss": 3.6574889284535663, + "tokens_seen": 1060306944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003427883650952859, + "loss": 1.3857, + "theoretical_loss": 3.657467813359015, + "tokens_seen": 1060372480 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034277833500501507, + "loss": 1.4707, + "theoretical_loss": 3.6574466999348125, + "tokens_seen": 1060438016 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034276830491474425, + "loss": 1.3628, + "theoretical_loss": 3.6574255881807245, + "tokens_seen": 1060503552 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034275827482447343, + "loss": 1.454, + "theoretical_loss": 3.657404478096515, + "tokens_seen": 1060569088 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003427482447342026, + "loss": 1.3188, + "theoretical_loss": 3.657383369681949, + "tokens_seen": 1060634624 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003427382146439318, + "loss": 1.419, + "theoretical_loss": 3.657362262936791, + "tokens_seen": 1060700160 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342728184553661, + "loss": 1.3814, + "theoretical_loss": 3.6573411578608064, + "tokens_seen": 1060765696 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034271815446339015, + "loss": 1.4509, + "theoretical_loss": 3.657320054453759, + "tokens_seen": 1060831232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003427081243731194, + "loss": 1.4288, + "theoretical_loss": 3.6572989527154154, + "tokens_seen": 1060896768 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003426980942828485, + "loss": 1.4651, + "theoretical_loss": 3.6572778526455387, + "tokens_seen": 1060962304 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034268806419257775, + "loss": 1.5421, + "theoretical_loss": 3.6572567542438956, + "tokens_seen": 1061027840 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034267803410230693, + "loss": 1.3705, + "theoretical_loss": 3.6572356575102503, + "tokens_seen": 1061093376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003426680040120361, + "loss": 1.5513, + "theoretical_loss": 3.6572145624443677, + "tokens_seen": 1061158912 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003426579739217653, + "loss": 1.4667, + "theoretical_loss": 3.6571934690460135, + "tokens_seen": 1061224448 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034264794383149453, + "loss": 1.3565, + "theoretical_loss": 3.6571723773149527, + "tokens_seen": 1061289984 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034263791374122366, + "loss": 1.5154, + "theoretical_loss": 3.657151287250951, + "tokens_seen": 1061355520 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003426278836509529, + "loss": 1.2885, + "theoretical_loss": 3.657130198853773, + "tokens_seen": 1061421056 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342617853560682, + "loss": 1.3919, + "theoretical_loss": 3.6571091121231847, + "tokens_seen": 1061486592 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034260782347041125, + "loss": 1.4327, + "theoretical_loss": 3.6570880270589505, + "tokens_seen": 1061552128 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034259779338014043, + "loss": 1.3682, + "theoretical_loss": 3.657066943660837, + "tokens_seen": 1061617664 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 389244, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4898905754089355, + "objective/train/theoretical_loss": 3.6570458619286095, + "objective/train/tokens_used": 1082143200, + "theoretical_loss": 3.6570458619286095, + "tokens_seen": 1061683200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425877632898696, + "loss": 1.4793, + "theoretical_loss": 3.6570458619286095, + "tokens_seen": 1061683200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425777331995988, + "loss": 1.4554, + "theoretical_loss": 3.6570247818620327, + "tokens_seen": 1061748736 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342567703109328, + "loss": 1.4491, + "theoretical_loss": 3.657003703460873, + "tokens_seen": 1061814272 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034255767301905716, + "loss": 1.3844, + "theoretical_loss": 3.656982626724896, + "tokens_seen": 1061879808 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425476429287864, + "loss": 1.4103, + "theoretical_loss": 3.6569615516538674, + "tokens_seen": 1061945344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425376128385155, + "loss": 1.5171, + "theoretical_loss": 3.656940478247553, + "tokens_seen": 1062010880 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034252758274824476, + "loss": 1.4498, + "theoretical_loss": 3.6569194065057182, + "tokens_seen": 1062076416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425175526579739, + "loss": 1.4962, + "theoretical_loss": 3.656898336428129, + "tokens_seen": 1062141952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003425075225677031, + "loss": 1.4, + "theoretical_loss": 3.6568772680145507, + "tokens_seen": 1062207488 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424974924774323, + "loss": 1.4265, + "theoretical_loss": 3.65685620126475, + "tokens_seen": 1062273024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424874623871615, + "loss": 1.3339, + "theoretical_loss": 3.6568351361784934, + "tokens_seen": 1062338560 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034247743229689066, + "loss": 1.4604, + "theoretical_loss": 3.656814072755546, + "tokens_seen": 1062404096 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424674022066199, + "loss": 1.4479, + "theoretical_loss": 3.656793010995674, + "tokens_seen": 1062469632 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342457372116349, + "loss": 1.3664, + "theoretical_loss": 3.6567719508986434, + "tokens_seen": 1062535168 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034244734202607826, + "loss": 1.3407, + "theoretical_loss": 3.6567508924642205, + "tokens_seen": 1062600704 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424373119358074, + "loss": 1.4931, + "theoretical_loss": 3.656729835692172, + "tokens_seen": 1062666240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424272818455366, + "loss": 1.5176, + "theoretical_loss": 3.656708780582264, + "tokens_seen": 1062731776 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003424172517552658, + "loss": 1.3411, + "theoretical_loss": 3.656687727134262, + "tokens_seen": 1062797312 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342407221664995, + "loss": 1.361, + "theoretical_loss": 3.656666675347933, + "tokens_seen": 1062862848 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034239719157472416, + "loss": 1.5328, + "theoretical_loss": 3.6566456252230437, + "tokens_seen": 1062928384 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034238716148445335, + "loss": 1.4155, + "theoretical_loss": 3.65662457675936, + "tokens_seen": 1062993920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423771313941825, + "loss": 1.3559, + "theoretical_loss": 3.6566035299566484, + "tokens_seen": 1063059456 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034236710130391176, + "loss": 1.4122, + "theoretical_loss": 3.6565824848146757, + "tokens_seen": 1063124992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423570712136409, + "loss": 1.4291, + "theoretical_loss": 3.656561441333209, + "tokens_seen": 1063190528 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423470411233701, + "loss": 1.3063, + "theoretical_loss": 3.6565403995120134, + "tokens_seen": 1063256064 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423370110330993, + "loss": 1.4183, + "theoretical_loss": 3.656519359350857, + "tokens_seen": 1063321600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423269809428285, + "loss": 1.278, + "theoretical_loss": 3.656498320849506, + "tokens_seen": 1063387136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003423169508525577, + "loss": 1.439, + "theoretical_loss": 3.6564772840077278, + "tokens_seen": 1063452672 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034230692076228685, + "loss": 1.4662, + "theoretical_loss": 3.6564562488252883, + "tokens_seen": 1063518208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003422968906720161, + "loss": 1.4601, + "theoretical_loss": 3.6564352153019546, + "tokens_seen": 1063583744 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034228686058174527, + "loss": 1.4701, + "theoretical_loss": 3.656414183437494, + "tokens_seen": 1063649280 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034227683049147445, + "loss": 1.2584, + "theoretical_loss": 3.656393153231673, + "tokens_seen": 1063714816 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034226680040120363, + "loss": 1.4302, + "theoretical_loss": 3.6563721246842595, + "tokens_seen": 1063780352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003422567703109328, + "loss": 1.3956, + "theoretical_loss": 3.656351097795019, + "tokens_seen": 1063845888 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342246740220662, + "loss": 1.2859, + "theoretical_loss": 3.6563300725637204, + "tokens_seen": 1063911424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003422367101303912, + "loss": 1.3523, + "theoretical_loss": 3.6563090489901295, + "tokens_seen": 1063976960 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034222668004012035, + "loss": 1.473, + "theoretical_loss": 3.656288027074014, + "tokens_seen": 1064042496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003422166499498496, + "loss": 1.5474, + "theoretical_loss": 3.656267006815142, + "tokens_seen": 1064108032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003422066198595787, + "loss": 1.2468, + "theoretical_loss": 3.6562459882132794, + "tokens_seen": 1064173568 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034219658976930795, + "loss": 1.5213, + "theoretical_loss": 3.656224971268194, + "tokens_seen": 1064239104 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034218655967903713, + "loss": 1.3759, + "theoretical_loss": 3.6562039559796533, + "tokens_seen": 1064304640 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003421765295887663, + "loss": 1.4632, + "theoretical_loss": 3.6561829423474252, + "tokens_seen": 1064370176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003421664994984955, + "loss": 1.558, + "theoretical_loss": 3.656161930371276, + "tokens_seen": 1064435712 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034215646940822473, + "loss": 1.4399, + "theoretical_loss": 3.6561409200509747, + "tokens_seen": 1064501248 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034214643931795386, + "loss": 1.3869, + "theoretical_loss": 3.6561199113862877, + "tokens_seen": 1064566784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003421364092276831, + "loss": 1.3648, + "theoretical_loss": 3.656098904376983, + "tokens_seen": 1064632320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003421263791374122, + "loss": 1.3738, + "theoretical_loss": 3.6560778990228284, + "tokens_seen": 1064697856 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034211634904714145, + "loss": 1.5178, + "theoretical_loss": 3.6560568953235917, + "tokens_seen": 1064763392 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034210631895687063, + "loss": 1.4164, + "theoretical_loss": 3.6560358932790407, + "tokens_seen": 1064828928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420962888665998, + "loss": 1.4265, + "theoretical_loss": 3.656014892888943, + "tokens_seen": 1064894464 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 390462, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3349173069000244, + "objective/train/theoretical_loss": 3.6559938941530667, + "objective/train/tokens_used": 1085420000, + "theoretical_loss": 3.6559938941530667, + "tokens_seen": 1064960000 + }, + { + "epoch": 0.32, + "learning_rate": 0.000342086258776329, + "loss": 1.4129, + "theoretical_loss": 3.6559938941530667, + "tokens_seen": 1064960000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420762286860582, + "loss": 1.4261, + "theoretical_loss": 3.655972897071179, + "tokens_seen": 1065025536 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034206619859578736, + "loss": 1.3984, + "theoretical_loss": 3.655951901643049, + "tokens_seen": 1065091072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420561685055166, + "loss": 1.4909, + "theoretical_loss": 3.655930907868443, + "tokens_seen": 1065156608 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420461384152457, + "loss": 1.4921, + "theoretical_loss": 3.6559099157471313, + "tokens_seen": 1065222144 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034203610832497496, + "loss": 1.368, + "theoretical_loss": 3.65588892527888, + "tokens_seen": 1065287680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420260782347041, + "loss": 1.365, + "theoretical_loss": 3.655867936463459, + "tokens_seen": 1065353216 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420160481444333, + "loss": 1.4729, + "theoretical_loss": 3.655846949300635, + "tokens_seen": 1065418752 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003420060180541625, + "loss": 1.3958, + "theoretical_loss": 3.6558259637901767, + "tokens_seen": 1065484288 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419959879638917, + "loss": 1.3928, + "theoretical_loss": 3.655804979931853, + "tokens_seen": 1065549824 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034198595787362086, + "loss": 1.4069, + "theoretical_loss": 3.655783997725431, + "tokens_seen": 1065615360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419759277833501, + "loss": 1.4655, + "theoretical_loss": 3.6557630171706803, + "tokens_seen": 1065680896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419658976930792, + "loss": 1.5796, + "theoretical_loss": 3.655742038267369, + "tokens_seen": 1065746432 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034195586760280846, + "loss": 1.4714, + "theoretical_loss": 3.655721061015265, + "tokens_seen": 1065811968 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419458375125376, + "loss": 1.4659, + "theoretical_loss": 3.6557000854141375, + "tokens_seen": 1065877504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419358074222668, + "loss": 1.3881, + "theoretical_loss": 3.6556791114637543, + "tokens_seen": 1065943040 + }, + { + "epoch": 0.32, + "learning_rate": 0.000341925777331996, + "loss": 1.4792, + "theoretical_loss": 3.6556581391638847, + "tokens_seen": 1066008576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003419157472417252, + "loss": 1.3802, + "theoretical_loss": 3.655637168514297, + "tokens_seen": 1066074112 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034190571715145436, + "loss": 1.4258, + "theoretical_loss": 3.6556161995147605, + "tokens_seen": 1066139648 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034189568706118355, + "loss": 1.4415, + "theoretical_loss": 3.6555952321650436, + "tokens_seen": 1066205184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003418856569709127, + "loss": 1.2702, + "theoretical_loss": 3.6555742664649147, + "tokens_seen": 1066270720 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034187562688064196, + "loss": 1.4669, + "theoretical_loss": 3.6555533024141424, + "tokens_seen": 1066336256 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034187562688064196, + "loss": 1.5254, + "theoretical_loss": 3.655532340012497, + "tokens_seen": 1066401792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003418655967903711, + "loss": 1.3303, + "theoretical_loss": 3.655511379259746, + "tokens_seen": 1066467328 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003418555667001003, + "loss": 1.404, + "theoretical_loss": 3.655490420155659, + "tokens_seen": 1066532864 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034184553660982945, + "loss": 1.4621, + "theoretical_loss": 3.655469462700005, + "tokens_seen": 1066598400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003418355065195587, + "loss": 1.3777, + "theoretical_loss": 3.6554485068925535, + "tokens_seen": 1066663936 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034182547642928787, + "loss": 1.463, + "theoretical_loss": 3.655427552733072, + "tokens_seen": 1066729472 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034181544633901705, + "loss": 1.4062, + "theoretical_loss": 3.6554066002213323, + "tokens_seen": 1066795008 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034180541624874623, + "loss": 1.444, + "theoretical_loss": 3.6553856493571013, + "tokens_seen": 1066860544 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034179538615847547, + "loss": 1.4877, + "theoretical_loss": 3.6553647001401495, + "tokens_seen": 1066926080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003417853560682046, + "loss": 1.4366, + "theoretical_loss": 3.6553437525702455, + "tokens_seen": 1066991616 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034177532597793383, + "loss": 1.4297, + "theoretical_loss": 3.6553228066471593, + "tokens_seen": 1067057152 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034176529588766295, + "loss": 1.4602, + "theoretical_loss": 3.65530186237066, + "tokens_seen": 1067122688 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003417552657973922, + "loss": 1.4453, + "theoretical_loss": 3.655280919740517, + "tokens_seen": 1067188224 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034174523570712137, + "loss": 1.3504, + "theoretical_loss": 3.6552599787564994, + "tokens_seen": 1067253760 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034173520561685055, + "loss": 1.3945, + "theoretical_loss": 3.655239039418378, + "tokens_seen": 1067319296 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034172517552657973, + "loss": 1.276, + "theoretical_loss": 3.6552181017259207, + "tokens_seen": 1067384832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003417151454363089, + "loss": 1.5072, + "theoretical_loss": 3.6551971656788984, + "tokens_seen": 1067450368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003417051153460381, + "loss": 1.3856, + "theoretical_loss": 3.65517623127708, + "tokens_seen": 1067515904 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034169508525576733, + "loss": 1.3741, + "theoretical_loss": 3.6551552985202362, + "tokens_seen": 1067581440 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034168505516549646, + "loss": 1.4492, + "theoretical_loss": 3.6551343674081362, + "tokens_seen": 1067646976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003416750250752257, + "loss": 1.4392, + "theoretical_loss": 3.655113437940549, + "tokens_seen": 1067712512 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003416649949849548, + "loss": 1.3807, + "theoretical_loss": 3.655092510117246, + "tokens_seen": 1067778048 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034165496489468406, + "loss": 1.4809, + "theoretical_loss": 3.655071583937996, + "tokens_seen": 1067843584 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034164493480441324, + "loss": 1.4234, + "theoretical_loss": 3.65505065940257, + "tokens_seen": 1067909120 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003416349047141424, + "loss": 1.4435, + "theoretical_loss": 3.6550297365107367, + "tokens_seen": 1067974656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003416248746238716, + "loss": 1.365, + "theoretical_loss": 3.655008815262267, + "tokens_seen": 1068040192 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034161484453360083, + "loss": 1.4873, + "theoretical_loss": 3.65498789565693, + "tokens_seen": 1068105728 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034160481444332996, + "loss": 1.4334, + "theoretical_loss": 3.6549669776944977, + "tokens_seen": 1068171264 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 391774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5020546913146973, + "objective/train/theoretical_loss": 3.654946061374739, + "objective/train/tokens_used": 1088696800, + "theoretical_loss": 3.654946061374739, + "tokens_seen": 1068236800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415947843530592, + "loss": 1.4252, + "theoretical_loss": 3.654946061374739, + "tokens_seen": 1068236800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415847542627884, + "loss": 1.4341, + "theoretical_loss": 3.6549251466974244, + "tokens_seen": 1068302336 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034157472417251756, + "loss": 1.3277, + "theoretical_loss": 3.6549042336623234, + "tokens_seen": 1068367872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415646940822468, + "loss": 1.4702, + "theoretical_loss": 3.6548833222692076, + "tokens_seen": 1068433408 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415546639919759, + "loss": 1.4465, + "theoretical_loss": 3.654862412517847, + "tokens_seen": 1068498944 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034154463390170516, + "loss": 1.4178, + "theoretical_loss": 3.654841504408012, + "tokens_seen": 1068564480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415346038114343, + "loss": 1.405, + "theoretical_loss": 3.6548205979394726, + "tokens_seen": 1068630016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415245737211635, + "loss": 1.4874, + "theoretical_loss": 3.6547996931119995, + "tokens_seen": 1068695552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415145436308927, + "loss": 1.454, + "theoretical_loss": 3.6547787899253636, + "tokens_seen": 1068761088 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003415045135406219, + "loss": 1.3292, + "theoretical_loss": 3.6547578883793355, + "tokens_seen": 1068826624 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034149448345035106, + "loss": 1.3143, + "theoretical_loss": 3.6547369884736858, + "tokens_seen": 1068892160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003414844533600803, + "loss": 1.4552, + "theoretical_loss": 3.6547160902081846, + "tokens_seen": 1068957696 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003414744232698094, + "loss": 1.4394, + "theoretical_loss": 3.6546951935826035, + "tokens_seen": 1069023232 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034146439317953866, + "loss": 1.4829, + "theoretical_loss": 3.6546742985967127, + "tokens_seen": 1069088768 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003414543630892678, + "loss": 1.4634, + "theoretical_loss": 3.654653405250283, + "tokens_seen": 1069154304 + }, + { + "epoch": 0.32, + "learning_rate": 0.000341444332998997, + "loss": 1.3465, + "theoretical_loss": 3.6546325135430857, + "tokens_seen": 1069219840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003414343029087262, + "loss": 1.4755, + "theoretical_loss": 3.6546116234748913, + "tokens_seen": 1069285376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003414242728184554, + "loss": 1.5537, + "theoretical_loss": 3.6545907350454714, + "tokens_seen": 1069350912 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034141424272818456, + "loss": 1.3582, + "theoretical_loss": 3.6545698482545963, + "tokens_seen": 1069416448 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034140421263791375, + "loss": 1.4562, + "theoretical_loss": 3.6545489631020374, + "tokens_seen": 1069481984 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034139418254764293, + "loss": 1.4073, + "theoretical_loss": 3.654528079587566, + "tokens_seen": 1069547520 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034138415245737216, + "loss": 1.399, + "theoretical_loss": 3.654507197710953, + "tokens_seen": 1069613056 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003413741223671013, + "loss": 1.4668, + "theoretical_loss": 3.65448631747197, + "tokens_seen": 1069678592 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003413640922768305, + "loss": 1.469, + "theoretical_loss": 3.6544654388703877, + "tokens_seen": 1069744128 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034135406218655965, + "loss": 1.2987, + "theoretical_loss": 3.6544445619059776, + "tokens_seen": 1069809664 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003413440320962889, + "loss": 1.3653, + "theoretical_loss": 3.654423686578511, + "tokens_seen": 1069875200 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034133400200601807, + "loss": 1.3882, + "theoretical_loss": 3.65440281288776, + "tokens_seen": 1069940736 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034132397191574725, + "loss": 1.4191, + "theoretical_loss": 3.6543819408334945, + "tokens_seen": 1070006272 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034131394182547643, + "loss": 1.3935, + "theoretical_loss": 3.6543610704154874, + "tokens_seen": 1070071808 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034130391173520567, + "loss": 1.4276, + "theoretical_loss": 3.6543402016335094, + "tokens_seen": 1070137344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003412938816449348, + "loss": 1.4863, + "theoretical_loss": 3.6543193344873326, + "tokens_seen": 1070202880 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034128385155466403, + "loss": 1.4398, + "theoretical_loss": 3.6542984689767284, + "tokens_seen": 1070268416 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034127382146439315, + "loss": 1.408, + "theoretical_loss": 3.654277605101468, + "tokens_seen": 1070333952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003412637913741224, + "loss": 1.4896, + "theoretical_loss": 3.6542567428613237, + "tokens_seen": 1070399488 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034125376128385157, + "loss": 1.4188, + "theoretical_loss": 3.654235882256067, + "tokens_seen": 1070465024 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034124373119358075, + "loss": 1.525, + "theoretical_loss": 3.6542150232854707, + "tokens_seen": 1070530560 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034123370110330993, + "loss": 1.5011, + "theoretical_loss": 3.6541941659493054, + "tokens_seen": 1070596096 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003412236710130391, + "loss": 1.5189, + "theoretical_loss": 3.6541733102473426, + "tokens_seen": 1070661632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003412136409227683, + "loss": 1.4917, + "theoretical_loss": 3.654152456179356, + "tokens_seen": 1070727168 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034120361083249753, + "loss": 1.3301, + "theoretical_loss": 3.6541316037451157, + "tokens_seen": 1070792704 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034119358074222666, + "loss": 1.4112, + "theoretical_loss": 3.654110752944395, + "tokens_seen": 1070858240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003411835506519559, + "loss": 1.4944, + "theoretical_loss": 3.6540899037769656, + "tokens_seen": 1070923776 + }, + { + "epoch": 0.32, + "learning_rate": 0.000341173520561685, + "loss": 1.4398, + "theoretical_loss": 3.6540690562425997, + "tokens_seen": 1070989312 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034116349047141426, + "loss": 1.4791, + "theoretical_loss": 3.6540482103410694, + "tokens_seen": 1071054848 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034115346038114344, + "loss": 1.4183, + "theoretical_loss": 3.654027366072147, + "tokens_seen": 1071120384 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003411434302908726, + "loss": 1.4133, + "theoretical_loss": 3.654006523435604, + "tokens_seen": 1071185920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003411334002006018, + "loss": 1.5477, + "theoretical_loss": 3.6539856824312142, + "tokens_seen": 1071251456 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034112337011033103, + "loss": 1.368, + "theoretical_loss": 3.6539648430587484, + "tokens_seen": 1071316992 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034111334002006016, + "loss": 1.4389, + "theoretical_loss": 3.65394400531798, + "tokens_seen": 1071382528 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003411033099297894, + "loss": 1.3623, + "theoretical_loss": 3.653923169208681, + "tokens_seen": 1071448064 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 393037, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.487640380859375, + "objective/train/theoretical_loss": 3.6539023347306245, + "objective/train/tokens_used": 1091973600, + "theoretical_loss": 3.6539023347306245, + "tokens_seen": 1071513600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003410932798395185, + "loss": 1.4346, + "theoretical_loss": 3.6539023347306245, + "tokens_seen": 1071513600 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034108324974924776, + "loss": 1.4325, + "theoretical_loss": 3.653881501883582, + "tokens_seen": 1071579136 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034107321965897694, + "loss": 1.4216, + "theoretical_loss": 3.653860670667327, + "tokens_seen": 1071644672 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003410631895687061, + "loss": 1.4874, + "theoretical_loss": 3.6538398410816315, + "tokens_seen": 1071710208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003410531594784353, + "loss": 1.5054, + "theoretical_loss": 3.6538190131262684, + "tokens_seen": 1071775744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003410431293881645, + "loss": 1.3637, + "theoretical_loss": 3.653798186801011, + "tokens_seen": 1071841280 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034103309929789366, + "loss": 1.3647, + "theoretical_loss": 3.6537773621056306, + "tokens_seen": 1071906816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003410230692076229, + "loss": 1.4875, + "theoretical_loss": 3.653756539039902, + "tokens_seen": 1071972352 + }, + { + "epoch": 0.32, + "learning_rate": 0.000341013039117352, + "loss": 1.3497, + "theoretical_loss": 3.6537357176035963, + "tokens_seen": 1072037888 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034100300902708126, + "loss": 1.4754, + "theoretical_loss": 3.653714897796487, + "tokens_seen": 1072103424 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034099297893681044, + "loss": 1.4222, + "theoretical_loss": 3.6536940796183472, + "tokens_seen": 1072168960 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003409829488465396, + "loss": 1.4694, + "theoretical_loss": 3.65367326306895, + "tokens_seen": 1072234496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003409729187562688, + "loss": 1.4689, + "theoretical_loss": 3.653652448148068, + "tokens_seen": 1072300032 + }, + { + "epoch": 0.32, + "learning_rate": 0.000340962888665998, + "loss": 1.4198, + "theoretical_loss": 3.6536316348554747, + "tokens_seen": 1072365568 + }, + { + "epoch": 0.32, + "learning_rate": 0.00034095285857572717, + "loss": 1.4149, + "theoretical_loss": 3.653610823190943, + "tokens_seen": 1072431104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0003409428284854564, + "loss": 1.398, + "theoretical_loss": 3.653590013154246, + "tokens_seen": 1072496640 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034093279839518553, + "loss": 1.455, + "theoretical_loss": 3.653569204745157, + "tokens_seen": 1072562176 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034092276830491477, + "loss": 1.3973, + "theoretical_loss": 3.6535483979634495, + "tokens_seen": 1072627712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003409127382146439, + "loss": 1.3683, + "theoretical_loss": 3.6535275928088966, + "tokens_seen": 1072693248 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034090270812437313, + "loss": 1.4785, + "theoretical_loss": 3.6535067892812716, + "tokens_seen": 1072758784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003408926780341023, + "loss": 1.3974, + "theoretical_loss": 3.6534859873803476, + "tokens_seen": 1072824320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003408826479438315, + "loss": 1.4983, + "theoretical_loss": 3.6534651871058985, + "tokens_seen": 1072889856 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034087261785356067, + "loss": 1.4363, + "theoretical_loss": 3.653444388457698, + "tokens_seen": 1072955392 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034086258776328985, + "loss": 1.4794, + "theoretical_loss": 3.6534235914355193, + "tokens_seen": 1073020928 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034085255767301903, + "loss": 1.4444, + "theoretical_loss": 3.6534027960391358, + "tokens_seen": 1073086464 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034084252758274827, + "loss": 1.4759, + "theoretical_loss": 3.653382002268321, + "tokens_seen": 1073152000 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034083249749247745, + "loss": 1.3586, + "theoretical_loss": 3.653361210122849, + "tokens_seen": 1073217536 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034082246740220663, + "loss": 1.3972, + "theoretical_loss": 3.653340419602494, + "tokens_seen": 1073283072 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034081243731193587, + "loss": 1.4237, + "theoretical_loss": 3.6533196307070286, + "tokens_seen": 1073348608 + }, + { + "epoch": 0.33, + "learning_rate": 0.000340802407221665, + "loss": 1.4875, + "theoretical_loss": 3.6532988434362275, + "tokens_seen": 1073414144 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034079237713139423, + "loss": 1.3929, + "theoretical_loss": 3.653278057789864, + "tokens_seen": 1073479680 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034078234704112335, + "loss": 1.4343, + "theoretical_loss": 3.6532572737677125, + "tokens_seen": 1073545216 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003407723169508526, + "loss": 1.4094, + "theoretical_loss": 3.6532364913695465, + "tokens_seen": 1073610752 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034076228686058177, + "loss": 1.4893, + "theoretical_loss": 3.6532157105951395, + "tokens_seen": 1073676288 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034075225677031095, + "loss": 1.3759, + "theoretical_loss": 3.6531949314442667, + "tokens_seen": 1073741824 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034074222668004013, + "loss": 1.3437, + "theoretical_loss": 3.6531741539167024, + "tokens_seen": 1073807360 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003407321965897693, + "loss": 1.3803, + "theoretical_loss": 3.653153378012219, + "tokens_seen": 1073872896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003407221664994985, + "loss": 1.4539, + "theoretical_loss": 3.653132603730592, + "tokens_seen": 1073938432 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034071213640922773, + "loss": 1.4554, + "theoretical_loss": 3.6531118310715947, + "tokens_seen": 1074003968 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034070210631895686, + "loss": 1.4369, + "theoretical_loss": 3.6530910600350026, + "tokens_seen": 1074069504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003406920762286861, + "loss": 1.4763, + "theoretical_loss": 3.6530702906205894, + "tokens_seen": 1074135040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003406820461384152, + "loss": 1.38, + "theoretical_loss": 3.6530495228281286, + "tokens_seen": 1074200576 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034067201604814446, + "loss": 1.5096, + "theoretical_loss": 3.6530287566573962, + "tokens_seen": 1074266112 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034066198595787364, + "loss": 1.5679, + "theoretical_loss": 3.653007992108165, + "tokens_seen": 1074331648 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003406519558676028, + "loss": 1.3115, + "theoretical_loss": 3.6529872291802112, + "tokens_seen": 1074397184 + }, + { + "epoch": 0.33, + "learning_rate": 0.000340641925777332, + "loss": 1.4214, + "theoretical_loss": 3.6529664678733074, + "tokens_seen": 1074462720 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034063189568706123, + "loss": 1.3845, + "theoretical_loss": 3.65294570818723, + "tokens_seen": 1074528256 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034062186559679036, + "loss": 1.4081, + "theoretical_loss": 3.6529249501217524, + "tokens_seen": 1074593792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003406118355065196, + "loss": 1.4219, + "theoretical_loss": 3.65290419367665, + "tokens_seen": 1074659328 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003406018054162487, + "loss": 1.463, + "theoretical_loss": 3.652883438851697, + "tokens_seen": 1074724864 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 394219, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.356762170791626, + "objective/train/theoretical_loss": 3.652862685646668, + "objective/train/tokens_used": 1095250400, + "theoretical_loss": 3.652862685646668, + "tokens_seen": 1074790400 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034059177532597796, + "loss": 1.4107, + "theoretical_loss": 3.652862685646668, + "tokens_seen": 1074790400 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034058174523570714, + "loss": 1.4244, + "theoretical_loss": 3.652841934061339, + "tokens_seen": 1074855936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003405717151454363, + "loss": 1.435, + "theoretical_loss": 3.652821184095483, + "tokens_seen": 1074921472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003405616850551655, + "loss": 1.389, + "theoretical_loss": 3.652800435748876, + "tokens_seen": 1074987008 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003405516549648947, + "loss": 1.4222, + "theoretical_loss": 3.6527796890212936, + "tokens_seen": 1075052544 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034054162487462386, + "loss": 1.4142, + "theoretical_loss": 3.6527589439125094, + "tokens_seen": 1075118080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003405315947843531, + "loss": 1.5291, + "theoretical_loss": 3.6527382004222995, + "tokens_seen": 1075183616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003405215646940822, + "loss": 1.4169, + "theoretical_loss": 3.6527174585504376, + "tokens_seen": 1075249152 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034051153460381146, + "loss": 1.416, + "theoretical_loss": 3.6526967182967, + "tokens_seen": 1075314688 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034050150451354064, + "loss": 1.4799, + "theoretical_loss": 3.652675979660862, + "tokens_seen": 1075380224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003404914744232698, + "loss": 1.4895, + "theoretical_loss": 3.652655242642698, + "tokens_seen": 1075445760 + }, + { + "epoch": 0.33, + "learning_rate": 0.000340481444332999, + "loss": 1.3515, + "theoretical_loss": 3.652634507241984, + "tokens_seen": 1075511296 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003404714142427282, + "loss": 1.5173, + "theoretical_loss": 3.6526137734584943, + "tokens_seen": 1075576832 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034046138415245737, + "loss": 1.3506, + "theoretical_loss": 3.6525930412920053, + "tokens_seen": 1075642368 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003404513540621866, + "loss": 1.4048, + "theoretical_loss": 3.6525723107422916, + "tokens_seen": 1075707904 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034044132397191573, + "loss": 1.4684, + "theoretical_loss": 3.6525515818091296, + "tokens_seen": 1075773440 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034043129388164497, + "loss": 1.3796, + "theoretical_loss": 3.6525308544922934, + "tokens_seen": 1075838976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003404212637913741, + "loss": 1.4325, + "theoretical_loss": 3.6525101287915596, + "tokens_seen": 1075904512 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034041123370110333, + "loss": 1.4185, + "theoretical_loss": 3.6524894047067034, + "tokens_seen": 1075970048 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003404012036108325, + "loss": 1.5297, + "theoretical_loss": 3.6524686822375, + "tokens_seen": 1076035584 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003403911735205617, + "loss": 1.3634, + "theoretical_loss": 3.652447961383726, + "tokens_seen": 1076101120 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034038114343029087, + "loss": 1.4428, + "theoretical_loss": 3.6524272421451567, + "tokens_seen": 1076166656 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034037111334002005, + "loss": 1.4039, + "theoretical_loss": 3.6524065245215676, + "tokens_seen": 1076232192 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034036108324974923, + "loss": 1.3355, + "theoretical_loss": 3.652385808512734, + "tokens_seen": 1076297728 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034035105315947847, + "loss": 1.4809, + "theoretical_loss": 3.6523650941184327, + "tokens_seen": 1076363264 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003403410230692076, + "loss": 1.4347, + "theoretical_loss": 3.6523443813384393, + "tokens_seen": 1076428800 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034033099297893683, + "loss": 1.3389, + "theoretical_loss": 3.65232367017253, + "tokens_seen": 1076494336 + }, + { + "epoch": 0.33, + "learning_rate": 0.000340320962888666, + "loss": 1.4523, + "theoretical_loss": 3.6523029606204798, + "tokens_seen": 1076559872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003403109327983952, + "loss": 1.4982, + "theoretical_loss": 3.6522822526820655, + "tokens_seen": 1076625408 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003403009027081244, + "loss": 1.3907, + "theoretical_loss": 3.652261546357063, + "tokens_seen": 1076690944 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034029087261785356, + "loss": 1.415, + "theoretical_loss": 3.6522408416452485, + "tokens_seen": 1076756480 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034028084252758274, + "loss": 1.4285, + "theoretical_loss": 3.6522201385463977, + "tokens_seen": 1076822016 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034027081243731197, + "loss": 1.425, + "theoretical_loss": 3.6521994370602875, + "tokens_seen": 1076887552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003402607823470411, + "loss": 1.4191, + "theoretical_loss": 3.6521787371866936, + "tokens_seen": 1076953088 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034025075225677033, + "loss": 1.356, + "theoretical_loss": 3.6521580389253927, + "tokens_seen": 1077018624 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034024072216649946, + "loss": 1.4617, + "theoretical_loss": 3.65213734227616, + "tokens_seen": 1077084160 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003402306920762287, + "loss": 1.4055, + "theoretical_loss": 3.652116647238773, + "tokens_seen": 1077149696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003402206619859579, + "loss": 1.5246, + "theoretical_loss": 3.6520959538130087, + "tokens_seen": 1077215232 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034021063189568706, + "loss": 1.5098, + "theoretical_loss": 3.652075261998642, + "tokens_seen": 1077280768 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034020060180541624, + "loss": 1.4867, + "theoretical_loss": 3.65205457179545, + "tokens_seen": 1077346304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401905717151454, + "loss": 1.3923, + "theoretical_loss": 3.6520338832032095, + "tokens_seen": 1077411840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401805416248746, + "loss": 1.3765, + "theoretical_loss": 3.652013196221697, + "tokens_seen": 1077477376 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034017051153460384, + "loss": 1.4456, + "theoretical_loss": 3.6519925108506888, + "tokens_seen": 1077542912 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034016048144433296, + "loss": 1.4409, + "theoretical_loss": 3.6519718270899624, + "tokens_seen": 1077608448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401504513540622, + "loss": 1.3347, + "theoretical_loss": 3.651951144939294, + "tokens_seen": 1077673984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401404212637914, + "loss": 1.4161, + "theoretical_loss": 3.6519304643984594, + "tokens_seen": 1077739520 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034013039117352056, + "loss": 1.3519, + "theoretical_loss": 3.651909785467237, + "tokens_seen": 1077805056 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034012036108324974, + "loss": 1.4309, + "theoretical_loss": 3.651889108145403, + "tokens_seen": 1077870592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401103309929789, + "loss": 1.4862, + "theoretical_loss": 3.6518684324327344, + "tokens_seen": 1077936128 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003401003009027081, + "loss": 1.5014, + "theoretical_loss": 3.6518477583290077, + "tokens_seen": 1078001664 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 395298, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3404419422149658, + "objective/train/theoretical_loss": 3.6518270858340007, + "objective/train/tokens_used": 1098527200, + "theoretical_loss": 3.6518270858340007, + "tokens_seen": 1078067200 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034009027081243734, + "loss": 1.3793, + "theoretical_loss": 3.6518270858340007, + "tokens_seen": 1078067200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003400802407221665, + "loss": 1.4232, + "theoretical_loss": 3.6518064149474903, + "tokens_seen": 1078132736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003400702106318957, + "loss": 1.3811, + "theoretical_loss": 3.651785745669253, + "tokens_seen": 1078198272 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003400601805416249, + "loss": 1.3991, + "theoretical_loss": 3.6517650779990656, + "tokens_seen": 1078263808 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034005015045135406, + "loss": 1.4961, + "theoretical_loss": 3.6517444119367064, + "tokens_seen": 1078329344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003400401203610833, + "loss": 1.4805, + "theoretical_loss": 3.651723747481952, + "tokens_seen": 1078394880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003400300902708124, + "loss": 1.3342, + "theoretical_loss": 3.6517030846345797, + "tokens_seen": 1078460416 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034002006018054166, + "loss": 1.5075, + "theoretical_loss": 3.6516824233943668, + "tokens_seen": 1078525952 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034001003009027084, + "loss": 1.5009, + "theoretical_loss": 3.651661763761091, + "tokens_seen": 1078591488 + }, + { + "epoch": 0.33, + "learning_rate": 0.00034, + "loss": 1.4072, + "theoretical_loss": 3.6516411057345293, + "tokens_seen": 1078657024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003399899699097292, + "loss": 1.3007, + "theoretical_loss": 3.651620449314459, + "tokens_seen": 1078722560 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003399799398194584, + "loss": 1.4581, + "theoretical_loss": 3.651599794500658, + "tokens_seen": 1078788096 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033996990972918757, + "loss": 1.4621, + "theoretical_loss": 3.6515791412929035, + "tokens_seen": 1078853632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003399598796389168, + "loss": 1.3646, + "theoretical_loss": 3.6515584896909736, + "tokens_seen": 1078919168 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033994984954864593, + "loss": 1.5386, + "theoretical_loss": 3.651537839694645, + "tokens_seen": 1078984704 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033993981945837517, + "loss": 1.4948, + "theoretical_loss": 3.6515171913036966, + "tokens_seen": 1079050240 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003399297893681043, + "loss": 1.4573, + "theoretical_loss": 3.651496544517905, + "tokens_seen": 1079115776 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033991975927783353, + "loss": 1.4084, + "theoretical_loss": 3.651475899337048, + "tokens_seen": 1079181312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003399097291875627, + "loss": 1.5287, + "theoretical_loss": 3.651455255760904, + "tokens_seen": 1079246848 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003398996990972919, + "loss": 1.4708, + "theoretical_loss": 3.6514346137892506, + "tokens_seen": 1079312384 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033988966900702107, + "loss": 1.4315, + "theoretical_loss": 3.6514139734218656, + "tokens_seen": 1079377920 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033987963891675025, + "loss": 1.4169, + "theoretical_loss": 3.6513933346585272, + "tokens_seen": 1079443456 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033986960882647943, + "loss": 1.4868, + "theoretical_loss": 3.6513726974990126, + "tokens_seen": 1079508992 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033985957873620867, + "loss": 1.4944, + "theoretical_loss": 3.6513520619431006, + "tokens_seen": 1079574528 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003398495486459378, + "loss": 1.2521, + "theoretical_loss": 3.651331427990569, + "tokens_seen": 1079640064 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033983951855566703, + "loss": 1.423, + "theoretical_loss": 3.651310795641196, + "tokens_seen": 1079705600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003398294884653962, + "loss": 1.3744, + "theoretical_loss": 3.65129016489476, + "tokens_seen": 1079771136 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003398194583751254, + "loss": 1.3748, + "theoretical_loss": 3.6512695357510383, + "tokens_seen": 1079836672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003398094282848546, + "loss": 1.3373, + "theoretical_loss": 3.6512489082098103, + "tokens_seen": 1079902208 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033979939819458376, + "loss": 1.5631, + "theoretical_loss": 3.6512282822708526, + "tokens_seen": 1079967744 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033978936810431294, + "loss": 1.4651, + "theoretical_loss": 3.6512076579339454, + "tokens_seen": 1080033280 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033977933801404217, + "loss": 1.399, + "theoretical_loss": 3.6511870351988662, + "tokens_seen": 1080098816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003397693079237713, + "loss": 1.5121, + "theoretical_loss": 3.6511664140653934, + "tokens_seen": 1080164352 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033975927783350053, + "loss": 1.4115, + "theoretical_loss": 3.6511457945333055, + "tokens_seen": 1080229888 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033974924774322966, + "loss": 1.366, + "theoretical_loss": 3.6511251766023807, + "tokens_seen": 1080295424 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003397392176529589, + "loss": 1.4852, + "theoretical_loss": 3.651104560272398, + "tokens_seen": 1080360960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003397291875626881, + "loss": 1.4397, + "theoretical_loss": 3.6510839455431356, + "tokens_seen": 1080426496 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033971915747241726, + "loss": 1.4378, + "theoretical_loss": 3.6510633324143726, + "tokens_seen": 1080492032 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033970912738214644, + "loss": 1.418, + "theoretical_loss": 3.6510427208858873, + "tokens_seen": 1080557568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396990972918756, + "loss": 1.416, + "theoretical_loss": 3.651022110957459, + "tokens_seen": 1080623104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396890672016048, + "loss": 1.3758, + "theoretical_loss": 3.6510015026288656, + "tokens_seen": 1080688640 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033967903711133404, + "loss": 1.5103, + "theoretical_loss": 3.6509808958998864, + "tokens_seen": 1080754176 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033966900702106316, + "loss": 1.3962, + "theoretical_loss": 3.6509602907703003, + "tokens_seen": 1080819712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396589769307924, + "loss": 1.4488, + "theoretical_loss": 3.650939687239886, + "tokens_seen": 1080885248 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396489468405216, + "loss": 1.3271, + "theoretical_loss": 3.6509190853084226, + "tokens_seen": 1080950784 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033963891675025076, + "loss": 1.3837, + "theoretical_loss": 3.6508984849756887, + "tokens_seen": 1081016320 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033962888665997994, + "loss": 1.5208, + "theoretical_loss": 3.6508778862414637, + "tokens_seen": 1081081856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396188565697091, + "loss": 1.494, + "theoretical_loss": 3.650857289105527, + "tokens_seen": 1081147392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003396088264794383, + "loss": 1.4178, + "theoretical_loss": 3.6508366935676566, + "tokens_seen": 1081212928 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033959879638916754, + "loss": 1.3471, + "theoretical_loss": 3.650816099627633, + "tokens_seen": 1081278464 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 396489, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5100921392440796, + "objective/train/theoretical_loss": 3.6507955072852343, + "objective/train/tokens_used": 1101804000, + "theoretical_loss": 3.6507955072852343, + "tokens_seen": 1081344000 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033958876629889667, + "loss": 1.4727, + "theoretical_loss": 3.6507955072852343, + "tokens_seen": 1081344000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003395787362086259, + "loss": 1.4398, + "theoretical_loss": 3.6507749165402403, + "tokens_seen": 1081409536 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033956870611835503, + "loss": 1.3605, + "theoretical_loss": 3.6507543273924306, + "tokens_seen": 1081475072 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033955867602808426, + "loss": 1.3687, + "theoretical_loss": 3.6507337398415833, + "tokens_seen": 1081540608 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033954864593781345, + "loss": 1.3757, + "theoretical_loss": 3.6507131538874793, + "tokens_seen": 1081606144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003395386158475426, + "loss": 1.4886, + "theoretical_loss": 3.6506925695298973, + "tokens_seen": 1081671680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003395285857572718, + "loss": 1.4027, + "theoretical_loss": 3.650671986768616, + "tokens_seen": 1081737216 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033951855566700104, + "loss": 1.4898, + "theoretical_loss": 3.6506514056034165, + "tokens_seen": 1081802752 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033950852557673017, + "loss": 1.6232, + "theoretical_loss": 3.6506308260340776, + "tokens_seen": 1081868288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003394984954864594, + "loss": 1.4516, + "theoretical_loss": 3.6506102480603784, + "tokens_seen": 1081933824 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033948846539618853, + "loss": 1.3635, + "theoretical_loss": 3.6505896716820994, + "tokens_seen": 1081999360 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033947843530591777, + "loss": 1.4037, + "theoretical_loss": 3.6505690968990194, + "tokens_seen": 1082064896 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033946840521564695, + "loss": 1.5267, + "theoretical_loss": 3.650548523710919, + "tokens_seen": 1082130432 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033945837512537613, + "loss": 1.4459, + "theoretical_loss": 3.6505279521175775, + "tokens_seen": 1082195968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003394483450351053, + "loss": 1.4237, + "theoretical_loss": 3.650507382118775, + "tokens_seen": 1082261504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003394383149448345, + "loss": 1.4832, + "theoretical_loss": 3.6504868137142905, + "tokens_seen": 1082327040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003394282848545637, + "loss": 1.3307, + "theoretical_loss": 3.650466246903905, + "tokens_seen": 1082392576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003394182547642929, + "loss": 1.4739, + "theoretical_loss": 3.6504456816873976, + "tokens_seen": 1082458112 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033940822467402204, + "loss": 1.397, + "theoretical_loss": 3.6504251180645495, + "tokens_seen": 1082523648 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033939819458375127, + "loss": 1.4938, + "theoretical_loss": 3.6504045560351392, + "tokens_seen": 1082589184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393881644934804, + "loss": 1.456, + "theoretical_loss": 3.650383995598948, + "tokens_seen": 1082654720 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033937813440320963, + "loss": 1.5304, + "theoretical_loss": 3.6503634367557547, + "tokens_seen": 1082720256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393681043129388, + "loss": 1.4684, + "theoretical_loss": 3.6503428795053408, + "tokens_seen": 1082785792 + }, + { + "epoch": 0.33, + "learning_rate": 0.000339358074222668, + "loss": 1.3803, + "theoretical_loss": 3.6503223238474862, + "tokens_seen": 1082851328 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393480441323972, + "loss": 1.5452, + "theoretical_loss": 3.650301769781971, + "tokens_seen": 1082916864 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393380140421264, + "loss": 1.4415, + "theoretical_loss": 3.6502812173085752, + "tokens_seen": 1082982400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393279839518556, + "loss": 1.4577, + "theoretical_loss": 3.650260666427079, + "tokens_seen": 1083047936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003393179538615848, + "loss": 1.4156, + "theoretical_loss": 3.6502401171372636, + "tokens_seen": 1083113472 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033930792377131396, + "loss": 1.397, + "theoretical_loss": 3.650219569438909, + "tokens_seen": 1083179008 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033929789368104314, + "loss": 1.4018, + "theoretical_loss": 3.650199023331796, + "tokens_seen": 1083244544 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033928786359077237, + "loss": 1.3499, + "theoretical_loss": 3.650178478815704, + "tokens_seen": 1083310080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003392778335005015, + "loss": 1.3664, + "theoretical_loss": 3.6501579358904146, + "tokens_seen": 1083375616 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033926780341023073, + "loss": 1.5431, + "theoretical_loss": 3.650137394555708, + "tokens_seen": 1083441152 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033925777331995986, + "loss": 1.3689, + "theoretical_loss": 3.6501168548113654, + "tokens_seen": 1083506688 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003392477432296891, + "loss": 1.4531, + "theoretical_loss": 3.6500963166571667, + "tokens_seen": 1083572224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003392377131394183, + "loss": 1.372, + "theoretical_loss": 3.650075780092893, + "tokens_seen": 1083637760 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033922768304914746, + "loss": 1.5138, + "theoretical_loss": 3.6500552451183252, + "tokens_seen": 1083703296 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033921765295887664, + "loss": 1.425, + "theoretical_loss": 3.650034711733244, + "tokens_seen": 1083768832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003392076228686058, + "loss": 1.3796, + "theoretical_loss": 3.6500141799374304, + "tokens_seen": 1083834368 + }, + { + "epoch": 0.33, + "learning_rate": 0.000339197592778335, + "loss": 1.4335, + "theoretical_loss": 3.649993649730665, + "tokens_seen": 1083899904 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033918756268806424, + "loss": 1.4864, + "theoretical_loss": 3.649973121112729, + "tokens_seen": 1083965440 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033917753259779336, + "loss": 1.4576, + "theoretical_loss": 3.649952594083403, + "tokens_seen": 1084030976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003391675025075226, + "loss": 1.4475, + "theoretical_loss": 3.6499320686424683, + "tokens_seen": 1084096512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003391574724172518, + "loss": 1.4368, + "theoretical_loss": 3.649911544789706, + "tokens_seen": 1084162048 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033914744232698096, + "loss": 1.4076, + "theoretical_loss": 3.6498910225248977, + "tokens_seen": 1084227584 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033913741223671014, + "loss": 1.4237, + "theoretical_loss": 3.649870501847824, + "tokens_seen": 1084293120 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003391273821464393, + "loss": 1.4746, + "theoretical_loss": 3.6498499827582664, + "tokens_seen": 1084358656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003391173520561685, + "loss": 1.3923, + "theoretical_loss": 3.6498294652560053, + "tokens_seen": 1084424192 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033910732196589774, + "loss": 1.3485, + "theoretical_loss": 3.649808949340823, + "tokens_seen": 1084489728 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033909729187562687, + "loss": 1.39, + "theoretical_loss": 3.6497884350125007, + "tokens_seen": 1084555264 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 397685, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3117731809616089, + "objective/train/theoretical_loss": 3.649767922270819, + "objective/train/tokens_used": 1105080800, + "theoretical_loss": 3.649767922270819, + "tokens_seen": 1084620800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003390872617853561, + "loss": 1.4269, + "theoretical_loss": 3.649767922270819, + "tokens_seen": 1084620800 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033907723169508523, + "loss": 1.498, + "theoretical_loss": 3.64974741111556, + "tokens_seen": 1084686336 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033906720160481446, + "loss": 1.4484, + "theoretical_loss": 3.6497269015465053, + "tokens_seen": 1084751872 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033905717151454365, + "loss": 1.4617, + "theoretical_loss": 3.649706393563436, + "tokens_seen": 1084817408 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033904714142427283, + "loss": 1.5528, + "theoretical_loss": 3.649685887166134, + "tokens_seen": 1084882944 + }, + { + "epoch": 0.33, + "learning_rate": 0.000339037111334002, + "loss": 1.5113, + "theoretical_loss": 3.6496653823543816, + "tokens_seen": 1084948480 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033902708124373124, + "loss": 1.5194, + "theoretical_loss": 3.6496448791279583, + "tokens_seen": 1085014016 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033901705115346037, + "loss": 1.3892, + "theoretical_loss": 3.6496243774866475, + "tokens_seen": 1085079552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003390070210631896, + "loss": 1.3752, + "theoretical_loss": 3.6496038774302306, + "tokens_seen": 1085145088 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033899699097291873, + "loss": 1.4313, + "theoretical_loss": 3.6495833789584893, + "tokens_seen": 1085210624 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033898696088264797, + "loss": 1.3512, + "theoretical_loss": 3.649562882071205, + "tokens_seen": 1085276160 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033897693079237715, + "loss": 1.4906, + "theoretical_loss": 3.64954238676816, + "tokens_seen": 1085341696 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033896690070210633, + "loss": 1.4352, + "theoretical_loss": 3.649521893049137, + "tokens_seen": 1085407232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003389568706118355, + "loss": 1.5189, + "theoretical_loss": 3.649501400913916, + "tokens_seen": 1085472768 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003389468405215647, + "loss": 1.4146, + "theoretical_loss": 3.6494809103622803, + "tokens_seen": 1085538304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003389368104312939, + "loss": 1.4842, + "theoretical_loss": 3.6494604213940125, + "tokens_seen": 1085603840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003389267803410231, + "loss": 1.4476, + "theoretical_loss": 3.649439934008893, + "tokens_seen": 1085669376 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033891675025075224, + "loss": 1.4127, + "theoretical_loss": 3.649419448206705, + "tokens_seen": 1085734912 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033890672016048147, + "loss": 1.3991, + "theoretical_loss": 3.6493989639872306, + "tokens_seen": 1085800448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003388966900702106, + "loss": 1.4336, + "theoretical_loss": 3.6493784813502517, + "tokens_seen": 1085865984 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033888665997993983, + "loss": 1.4359, + "theoretical_loss": 3.649358000295551, + "tokens_seen": 1085931520 + }, + { + "epoch": 0.33, + "learning_rate": 0.000338876629889669, + "loss": 1.4351, + "theoretical_loss": 3.64933752082291, + "tokens_seen": 1085997056 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003388665997993982, + "loss": 1.401, + "theoretical_loss": 3.649317042932112, + "tokens_seen": 1086062592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003388565697091274, + "loss": 1.3873, + "theoretical_loss": 3.649296566622939, + "tokens_seen": 1086128128 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003388465396188566, + "loss": 1.3216, + "theoretical_loss": 3.6492760918951728, + "tokens_seen": 1086193664 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033883650952858574, + "loss": 1.4911, + "theoretical_loss": 3.6492556187485965, + "tokens_seen": 1086259200 + }, + { + "epoch": 0.33, + "learning_rate": 0.000338826479438315, + "loss": 1.4006, + "theoretical_loss": 3.649235147182993, + "tokens_seen": 1086324736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003388164493480441, + "loss": 1.4714, + "theoretical_loss": 3.649214677198144, + "tokens_seen": 1086390272 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033880641925777334, + "loss": 1.4179, + "theoretical_loss": 3.6491942087938325, + "tokens_seen": 1086455808 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387963891675025, + "loss": 1.4032, + "theoretical_loss": 3.649173741969841, + "tokens_seen": 1086521344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387863590772317, + "loss": 1.4086, + "theoretical_loss": 3.6491532767259525, + "tokens_seen": 1086586880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387763289869609, + "loss": 1.3841, + "theoretical_loss": 3.649132813061949, + "tokens_seen": 1086652416 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033876629889669006, + "loss": 1.2725, + "theoretical_loss": 3.6491123509776138, + "tokens_seen": 1086717952 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033875626880641924, + "loss": 1.5204, + "theoretical_loss": 3.64909189047273, + "tokens_seen": 1086783488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387462387161485, + "loss": 1.3431, + "theoretical_loss": 3.64907143154708, + "tokens_seen": 1086849024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387362086258776, + "loss": 1.3591, + "theoretical_loss": 3.6490509742004464, + "tokens_seen": 1086914560 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033872617853560684, + "loss": 1.4487, + "theoretical_loss": 3.649030518432613, + "tokens_seen": 1086980096 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033871614844533597, + "loss": 1.4151, + "theoretical_loss": 3.6490100642433623, + "tokens_seen": 1087045632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003387061183550652, + "loss": 1.446, + "theoretical_loss": 3.648989611632477, + "tokens_seen": 1087111168 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003386960882647944, + "loss": 1.388, + "theoretical_loss": 3.6489691605997403, + "tokens_seen": 1087176704 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033868605817452356, + "loss": 1.318, + "theoretical_loss": 3.6489487111449357, + "tokens_seen": 1087242240 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033867602808425275, + "loss": 1.3678, + "theoretical_loss": 3.6489282632678464, + "tokens_seen": 1087307776 + }, + { + "epoch": 0.33, + "learning_rate": 0.000338665997993982, + "loss": 1.4689, + "theoretical_loss": 3.648907816968255, + "tokens_seen": 1087373312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003386559679037111, + "loss": 1.5379, + "theoretical_loss": 3.6488873722459454, + "tokens_seen": 1087438848 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033864593781344034, + "loss": 1.2891, + "theoretical_loss": 3.6488669291007003, + "tokens_seen": 1087504384 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033863590772316947, + "loss": 1.3815, + "theoretical_loss": 3.6488464875323032, + "tokens_seen": 1087569920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003386258776328987, + "loss": 1.5066, + "theoretical_loss": 3.6488260475405374, + "tokens_seen": 1087635456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003386158475426279, + "loss": 1.4601, + "theoretical_loss": 3.648805609125187, + "tokens_seen": 1087700992 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033860581745235707, + "loss": 1.3672, + "theoretical_loss": 3.6487851722860345, + "tokens_seen": 1087766528 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033859578736208625, + "loss": 1.4183, + "theoretical_loss": 3.648764737022864, + "tokens_seen": 1087832064 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 398862, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3425993919372559, + "objective/train/theoretical_loss": 3.6487443033354583, + "objective/train/tokens_used": 1108357600, + "theoretical_loss": 3.6487443033354583, + "tokens_seen": 1087897600 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033858575727181543, + "loss": 1.3625, + "theoretical_loss": 3.6487443033354583, + "tokens_seen": 1087897600 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033857572718154467, + "loss": 1.364, + "theoretical_loss": 3.648723871223602, + "tokens_seen": 1087963136 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033856569709127385, + "loss": 1.3796, + "theoretical_loss": 3.6487034406870777, + "tokens_seen": 1088028672 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033855566700100303, + "loss": 1.3983, + "theoretical_loss": 3.6486830117256703, + "tokens_seen": 1088094208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003385456369107322, + "loss": 1.3824, + "theoretical_loss": 3.648662584339162, + "tokens_seen": 1088159744 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033853560682046144, + "loss": 1.4056, + "theoretical_loss": 3.6486421585273376, + "tokens_seen": 1088225280 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033852557673019057, + "loss": 1.3272, + "theoretical_loss": 3.648621734289981, + "tokens_seen": 1088290816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003385155466399198, + "loss": 1.4444, + "theoretical_loss": 3.648601311626875, + "tokens_seen": 1088356352 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033850551654964893, + "loss": 1.5469, + "theoretical_loss": 3.6485808905378048, + "tokens_seen": 1088421888 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033849548645937817, + "loss": 1.3876, + "theoretical_loss": 3.648560471022553, + "tokens_seen": 1088487424 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033848545636910735, + "loss": 1.4371, + "theoretical_loss": 3.6485400530809047, + "tokens_seen": 1088552960 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033847542627883653, + "loss": 1.4593, + "theoretical_loss": 3.6485196367126433, + "tokens_seen": 1088618496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003384653961885657, + "loss": 1.3845, + "theoretical_loss": 3.648499221917553, + "tokens_seen": 1088684032 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003384553660982949, + "loss": 1.3397, + "theoretical_loss": 3.6484788086954176, + "tokens_seen": 1088749568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003384453360080241, + "loss": 1.4627, + "theoretical_loss": 3.6484583970460216, + "tokens_seen": 1088815104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003384353059177533, + "loss": 1.4149, + "theoretical_loss": 3.6484379869691494, + "tokens_seen": 1088880640 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033842527582748244, + "loss": 1.4513, + "theoretical_loss": 3.6484175784645845, + "tokens_seen": 1088946176 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033841524573721167, + "loss": 1.5618, + "theoretical_loss": 3.6483971715321117, + "tokens_seen": 1089011712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003384052156469408, + "loss": 1.5173, + "theoretical_loss": 3.6483767661715154, + "tokens_seen": 1089077248 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033839518555667003, + "loss": 1.3639, + "theoretical_loss": 3.648356362382579, + "tokens_seen": 1089142784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383851554663992, + "loss": 1.3657, + "theoretical_loss": 3.648335960165088, + "tokens_seen": 1089208320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383751253761284, + "loss": 1.639, + "theoretical_loss": 3.6483155595188266, + "tokens_seen": 1089273856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383650952858576, + "loss": 1.4245, + "theoretical_loss": 3.648295160443579, + "tokens_seen": 1089339392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383550651955868, + "loss": 1.4128, + "theoretical_loss": 3.6482747629391294, + "tokens_seen": 1089404928 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033834503510531594, + "loss": 1.4684, + "theoretical_loss": 3.648254367005263, + "tokens_seen": 1089470464 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383350050150452, + "loss": 1.4872, + "theoretical_loss": 3.6482339726417643, + "tokens_seen": 1089536000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383249749247743, + "loss": 1.4072, + "theoretical_loss": 3.6482135798484174, + "tokens_seen": 1089601536 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033831494483450354, + "loss": 1.5069, + "theoretical_loss": 3.648193188625008, + "tokens_seen": 1089667072 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003383049147442327, + "loss": 1.3622, + "theoretical_loss": 3.64817279897132, + "tokens_seen": 1089732608 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382948846539619, + "loss": 1.4307, + "theoretical_loss": 3.648152410887138, + "tokens_seen": 1089798144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382848545636911, + "loss": 1.5331, + "theoretical_loss": 3.6481320243722477, + "tokens_seen": 1089863680 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033827482447342026, + "loss": 1.4875, + "theoretical_loss": 3.6481116394264337, + "tokens_seen": 1089929216 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033826479438314944, + "loss": 1.4953, + "theoretical_loss": 3.64809125604948, + "tokens_seen": 1089994752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382547642928787, + "loss": 1.385, + "theoretical_loss": 3.6480708742411725, + "tokens_seen": 1090060288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382447342026078, + "loss": 1.427, + "theoretical_loss": 3.6480504940012954, + "tokens_seen": 1090125824 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033823470411233704, + "loss": 1.5518, + "theoretical_loss": 3.648030115329635, + "tokens_seen": 1090191360 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033822467402206617, + "loss": 1.4915, + "theoretical_loss": 3.648009738225975, + "tokens_seen": 1090256896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382146439317954, + "loss": 1.4034, + "theoretical_loss": 3.647989362690101, + "tokens_seen": 1090322432 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003382046138415246, + "loss": 1.3443, + "theoretical_loss": 3.6479689887217988, + "tokens_seen": 1090387968 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033819458375125376, + "loss": 1.3953, + "theoretical_loss": 3.6479486163208525, + "tokens_seen": 1090453504 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033818455366098295, + "loss": 1.4941, + "theoretical_loss": 3.6479282454870487, + "tokens_seen": 1090519040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003381745235707122, + "loss": 1.433, + "theoretical_loss": 3.647907876220171, + "tokens_seen": 1090584576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003381644934804413, + "loss": 1.3961, + "theoretical_loss": 3.6478875085200055, + "tokens_seen": 1090650112 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033815446339017054, + "loss": 1.4941, + "theoretical_loss": 3.647867142386338, + "tokens_seen": 1090715648 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033814443329989967, + "loss": 1.3372, + "theoretical_loss": 3.6478467778189536, + "tokens_seen": 1090781184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003381344032096289, + "loss": 1.4795, + "theoretical_loss": 3.647826414817638, + "tokens_seen": 1090846720 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003381243731193581, + "loss": 1.4762, + "theoretical_loss": 3.6478060533821757, + "tokens_seen": 1090912256 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033811434302908727, + "loss": 1.5604, + "theoretical_loss": 3.647785693512353, + "tokens_seen": 1090977792 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033810431293881645, + "loss": 1.3522, + "theoretical_loss": 3.6477653352079558, + "tokens_seen": 1091043328 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033809428284854563, + "loss": 1.4464, + "theoretical_loss": 3.647744978468769, + "tokens_seen": 1091108864 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 400125, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5291483402252197, + "objective/train/theoretical_loss": 3.647724623294579, + "objective/train/tokens_used": 1111634400, + "theoretical_loss": 3.647724623294579, + "tokens_seen": 1091174400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003380842527582748, + "loss": 1.4222, + "theoretical_loss": 3.647724623294579, + "tokens_seen": 1091174400 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033807422266800405, + "loss": 1.5153, + "theoretical_loss": 3.6477042696851707, + "tokens_seen": 1091239936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003380641925777332, + "loss": 1.4372, + "theoretical_loss": 3.64768391764033, + "tokens_seen": 1091305472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003380541624874624, + "loss": 1.3353, + "theoretical_loss": 3.647663567159843, + "tokens_seen": 1091371008 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033804413239719154, + "loss": 1.4404, + "theoretical_loss": 3.6476432182434957, + "tokens_seen": 1091436544 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033803410230692077, + "loss": 1.4184, + "theoretical_loss": 3.647622870891074, + "tokens_seen": 1091502080 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033802407221664995, + "loss": 1.5676, + "theoretical_loss": 3.647602525102363, + "tokens_seen": 1091567616 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033801404212637913, + "loss": 1.3239, + "theoretical_loss": 3.6475821808771496, + "tokens_seen": 1091633152 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003380040120361083, + "loss": 1.548, + "theoretical_loss": 3.6475618382152186, + "tokens_seen": 1091698688 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033799398194583755, + "loss": 1.5771, + "theoretical_loss": 3.6475414971163573, + "tokens_seen": 1091764224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003379839518555667, + "loss": 1.4456, + "theoretical_loss": 3.6475211575803517, + "tokens_seen": 1091829760 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003379739217652959, + "loss": 1.3567, + "theoretical_loss": 3.6475008196069867, + "tokens_seen": 1091895296 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033796389167502504, + "loss": 1.4483, + "theoretical_loss": 3.6474804831960497, + "tokens_seen": 1091960832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003379538615847543, + "loss": 1.4613, + "theoretical_loss": 3.6474601483473266, + "tokens_seen": 1092026368 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033794383149448346, + "loss": 1.4709, + "theoretical_loss": 3.6474398150606033, + "tokens_seen": 1092091904 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033793380140421264, + "loss": 1.4899, + "theoretical_loss": 3.647419483335667, + "tokens_seen": 1092157440 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003379237713139418, + "loss": 1.4668, + "theoretical_loss": 3.6473991531723025, + "tokens_seen": 1092222976 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337913741223671, + "loss": 1.449, + "theoretical_loss": 3.647378824570297, + "tokens_seen": 1092288512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003379037111334002, + "loss": 1.4207, + "theoretical_loss": 3.6473584975294377, + "tokens_seen": 1092354048 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378936810431294, + "loss": 1.4392, + "theoretical_loss": 3.64733817204951, + "tokens_seen": 1092419584 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033788365095285854, + "loss": 1.3537, + "theoretical_loss": 3.6473178481303004, + "tokens_seen": 1092485120 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378736208625878, + "loss": 1.3672, + "theoretical_loss": 3.647297525771596, + "tokens_seen": 1092550656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378635907723169, + "loss": 1.4993, + "theoretical_loss": 3.647277204973183, + "tokens_seen": 1092616192 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033785356068204614, + "loss": 1.4054, + "theoretical_loss": 3.6472568857348486, + "tokens_seen": 1092681728 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378435305917753, + "loss": 1.4405, + "theoretical_loss": 3.647236568056379, + "tokens_seen": 1092747264 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378335005015045, + "loss": 1.4832, + "theoretical_loss": 3.647216251937561, + "tokens_seen": 1092812800 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033782347041123374, + "loss": 1.4274, + "theoretical_loss": 3.647195937378181, + "tokens_seen": 1092878336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378134403209629, + "loss": 1.4876, + "theoretical_loss": 3.6471756243780264, + "tokens_seen": 1092943872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003378034102306921, + "loss": 1.4844, + "theoretical_loss": 3.647155312936883, + "tokens_seen": 1093009408 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003377933801404213, + "loss": 1.4217, + "theoretical_loss": 3.6471350030545393, + "tokens_seen": 1093074944 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033778335005015046, + "loss": 1.4908, + "theoretical_loss": 3.6471146947307815, + "tokens_seen": 1093140480 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033777331995987964, + "loss": 1.3302, + "theoretical_loss": 3.647094387965396, + "tokens_seen": 1093206016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003377632898696089, + "loss": 1.4431, + "theoretical_loss": 3.64707408275817, + "tokens_seen": 1093271552 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337753259779338, + "loss": 1.4455, + "theoretical_loss": 3.6470537791088913, + "tokens_seen": 1093337088 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033774322968906724, + "loss": 1.3532, + "theoretical_loss": 3.647033477017346, + "tokens_seen": 1093402624 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033773319959879637, + "loss": 1.443, + "theoretical_loss": 3.6470131764833225, + "tokens_seen": 1093468160 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003377231695085256, + "loss": 1.3357, + "theoretical_loss": 3.646992877506606, + "tokens_seen": 1093533696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003377131394182548, + "loss": 1.4408, + "theoretical_loss": 3.646972580086986, + "tokens_seen": 1093599232 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033770310932798396, + "loss": 1.4614, + "theoretical_loss": 3.646952284224248, + "tokens_seen": 1093664768 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033769307923771315, + "loss": 1.3463, + "theoretical_loss": 3.64693198991818, + "tokens_seen": 1093730304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003376830491474424, + "loss": 1.4183, + "theoretical_loss": 3.6469116971685693, + "tokens_seen": 1093795840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003376730190571715, + "loss": 1.5421, + "theoretical_loss": 3.6468914059752033, + "tokens_seen": 1093861376 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033766298896690074, + "loss": 1.4539, + "theoretical_loss": 3.6468711163378695, + "tokens_seen": 1093926912 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033765295887662987, + "loss": 1.4538, + "theoretical_loss": 3.646850828256355, + "tokens_seen": 1093992448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003376429287863591, + "loss": 1.554, + "theoretical_loss": 3.646830541730448, + "tokens_seen": 1094057984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003376328986960883, + "loss": 1.4377, + "theoretical_loss": 3.6468102567599354, + "tokens_seen": 1094123520 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033762286860581747, + "loss": 1.4466, + "theoretical_loss": 3.646789973344605, + "tokens_seen": 1094189056 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033761283851554665, + "loss": 1.3502, + "theoretical_loss": 3.6467696914842445, + "tokens_seen": 1094254592 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033760280842527583, + "loss": 1.4938, + "theoretical_loss": 3.646749411178641, + "tokens_seen": 1094320128 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337592778335005, + "loss": 1.4072, + "theoretical_loss": 3.6467291324275832, + "tokens_seen": 1094385664 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 401268, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3652352094650269, + "objective/train/theoretical_loss": 3.646708855230859, + "objective/train/tokens_used": 1114911200, + "theoretical_loss": 3.646708855230859, + "tokens_seen": 1094451200 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033758274824473425, + "loss": 1.4644, + "theoretical_loss": 3.646708855230859, + "tokens_seen": 1094451200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003375727181544634, + "loss": 1.4835, + "theoretical_loss": 3.646688579588255, + "tokens_seen": 1094516736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003375626880641926, + "loss": 1.4294, + "theoretical_loss": 3.646668305499559, + "tokens_seen": 1094582272 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033755265797392174, + "loss": 1.3021, + "theoretical_loss": 3.6466480329645603, + "tokens_seen": 1094647808 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033754262788365097, + "loss": 1.434, + "theoretical_loss": 3.646627761983046, + "tokens_seen": 1094713344 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033753259779338015, + "loss": 1.4998, + "theoretical_loss": 3.6466074925548035, + "tokens_seen": 1094778880 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033752256770310933, + "loss": 1.476, + "theoretical_loss": 3.646587224679622, + "tokens_seen": 1094844416 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003375125376128385, + "loss": 1.413, + "theoretical_loss": 3.646566958357289, + "tokens_seen": 1094909952 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033750250752256775, + "loss": 1.5016, + "theoretical_loss": 3.6465466935875925, + "tokens_seen": 1094975488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374924774322969, + "loss": 1.3662, + "theoretical_loss": 3.6465264303703204, + "tokens_seen": 1095041024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374824473420261, + "loss": 1.4419, + "theoretical_loss": 3.6465061687052613, + "tokens_seen": 1095106560 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033747241725175524, + "loss": 1.372, + "theoretical_loss": 3.6464859085922035, + "tokens_seen": 1095172096 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374623871614845, + "loss": 1.3455, + "theoretical_loss": 3.646465650030935, + "tokens_seen": 1095237632 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033745235707121366, + "loss": 1.3816, + "theoretical_loss": 3.6464453930212435, + "tokens_seen": 1095303168 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033744232698094284, + "loss": 1.4454, + "theoretical_loss": 3.6464251375629186, + "tokens_seen": 1095368704 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337432296890672, + "loss": 1.492, + "theoretical_loss": 3.646404883655748, + "tokens_seen": 1095434240 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374222668004012, + "loss": 1.4086, + "theoretical_loss": 3.6463846312995196, + "tokens_seen": 1095499776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374122367101304, + "loss": 1.3994, + "theoretical_loss": 3.646364380494023, + "tokens_seen": 1095565312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003374022066198596, + "loss": 1.4061, + "theoretical_loss": 3.6463441312390463, + "tokens_seen": 1095630848 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033739217652958874, + "loss": 1.4448, + "theoretical_loss": 3.646323883534378, + "tokens_seen": 1095696384 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337382146439318, + "loss": 1.5574, + "theoretical_loss": 3.6463036373798063, + "tokens_seen": 1095761920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373721163490471, + "loss": 1.3099, + "theoretical_loss": 3.64628339277512, + "tokens_seen": 1095827456 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033736208625877634, + "loss": 1.3062, + "theoretical_loss": 3.6462631497201077, + "tokens_seen": 1095892992 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373520561685055, + "loss": 1.4531, + "theoretical_loss": 3.6462429082145587, + "tokens_seen": 1095958528 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373420260782347, + "loss": 1.4637, + "theoretical_loss": 3.646222668258261, + "tokens_seen": 1096024064 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373319959879639, + "loss": 1.411, + "theoretical_loss": 3.6462024298510043, + "tokens_seen": 1096089600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373219658976931, + "loss": 1.3712, + "theoretical_loss": 3.6461821929925766, + "tokens_seen": 1096155136 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033731193580742225, + "loss": 1.3921, + "theoretical_loss": 3.646161957682767, + "tokens_seen": 1096220672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003373019057171515, + "loss": 1.4203, + "theoretical_loss": 3.6461417239213647, + "tokens_seen": 1096286208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003372918756268806, + "loss": 1.4897, + "theoretical_loss": 3.6461214917081586, + "tokens_seen": 1096351744 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033728184553660984, + "loss": 1.3773, + "theoretical_loss": 3.6461012610429373, + "tokens_seen": 1096417280 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337271815446339, + "loss": 1.4612, + "theoretical_loss": 3.64608103192549, + "tokens_seen": 1096482816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003372617853560682, + "loss": 1.404, + "theoretical_loss": 3.6460608043556064, + "tokens_seen": 1096548352 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003372517552657974, + "loss": 1.4833, + "theoretical_loss": 3.6460405783330745, + "tokens_seen": 1096613888 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033724172517552657, + "loss": 1.3163, + "theoretical_loss": 3.6460203538576845, + "tokens_seen": 1096679424 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033723169508525575, + "loss": 1.3815, + "theoretical_loss": 3.6460001309292247, + "tokens_seen": 1096744960 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337221664994985, + "loss": 1.4745, + "theoretical_loss": 3.6459799095474854, + "tokens_seen": 1096810496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003372116349047141, + "loss": 1.4524, + "theoretical_loss": 3.6459596897122553, + "tokens_seen": 1096876032 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033720160481444335, + "loss": 1.5143, + "theoretical_loss": 3.6459394714233238, + "tokens_seen": 1096941568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371915747241725, + "loss": 1.49, + "theoretical_loss": 3.64591925468048, + "tokens_seen": 1097007104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371815446339017, + "loss": 1.3666, + "theoretical_loss": 3.645899039483514, + "tokens_seen": 1097072640 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371715145436309, + "loss": 1.4321, + "theoretical_loss": 3.6458788258322143, + "tokens_seen": 1097138176 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033716148445336007, + "loss": 1.4744, + "theoretical_loss": 3.6458586137263715, + "tokens_seen": 1097203712 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033715145436308925, + "loss": 1.3195, + "theoretical_loss": 3.645838403165774, + "tokens_seen": 1097269248 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371414242728185, + "loss": 1.4014, + "theoretical_loss": 3.6458181941502126, + "tokens_seen": 1097334784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371313941825476, + "loss": 1.4093, + "theoretical_loss": 3.645797986679476, + "tokens_seen": 1097400320 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033712136409227685, + "loss": 1.3972, + "theoretical_loss": 3.6457777807533542, + "tokens_seen": 1097465856 + }, + { + "epoch": 0.33, + "learning_rate": 0.000337111334002006, + "loss": 1.3977, + "theoretical_loss": 3.6457575763716363, + "tokens_seen": 1097531392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003371013039117352, + "loss": 1.3955, + "theoretical_loss": 3.645737373534113, + "tokens_seen": 1097596928 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003370912738214644, + "loss": 1.5346, + "theoretical_loss": 3.645717172240574, + "tokens_seen": 1097662464 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 402429, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3249115943908691, + "objective/train/theoretical_loss": 3.645696972490809, + "objective/train/tokens_used": 1118188000, + "theoretical_loss": 3.645696972490809, + "tokens_seen": 1097728000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003370812437311936, + "loss": 1.4545, + "theoretical_loss": 3.645696972490809, + "tokens_seen": 1097728000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003370712136409228, + "loss": 1.4312, + "theoretical_loss": 3.645676774284607, + "tokens_seen": 1097793536 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033706118355065194, + "loss": 1.4186, + "theoretical_loss": 3.645656577621758, + "tokens_seen": 1097859072 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033705115346038117, + "loss": 1.4648, + "theoretical_loss": 3.6456363825020537, + "tokens_seen": 1097924608 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033704112337011035, + "loss": 1.5138, + "theoretical_loss": 3.645616188925283, + "tokens_seen": 1097990144 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033703109327983953, + "loss": 1.3543, + "theoretical_loss": 3.6455959968912355, + "tokens_seen": 1098055680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003370210631895687, + "loss": 1.3676, + "theoretical_loss": 3.6455758063997017, + "tokens_seen": 1098121216 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033701103309929795, + "loss": 1.3994, + "theoretical_loss": 3.645555617450472, + "tokens_seen": 1098186752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003370010030090271, + "loss": 1.4074, + "theoretical_loss": 3.645535430043336, + "tokens_seen": 1098252288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369909729187563, + "loss": 1.3582, + "theoretical_loss": 3.6455152441780836, + "tokens_seen": 1098317824 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033698094282848544, + "loss": 1.3779, + "theoretical_loss": 3.645495059854506, + "tokens_seen": 1098383360 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369709127382147, + "loss": 1.3872, + "theoretical_loss": 3.6454748770723935, + "tokens_seen": 1098448896 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033696088264794386, + "loss": 1.5205, + "theoretical_loss": 3.6454546958315355, + "tokens_seen": 1098514432 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033695085255767304, + "loss": 1.4776, + "theoretical_loss": 3.645434516131723, + "tokens_seen": 1098579968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369408224674022, + "loss": 1.4488, + "theoretical_loss": 3.6454143379727464, + "tokens_seen": 1098645504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369307923771314, + "loss": 1.4865, + "theoretical_loss": 3.645394161354395, + "tokens_seen": 1098711040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369207622868606, + "loss": 1.434, + "theoretical_loss": 3.6453739862764616, + "tokens_seen": 1098776576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003369107321965898, + "loss": 1.3934, + "theoretical_loss": 3.6453538127387346, + "tokens_seen": 1098842112 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033690070210631894, + "loss": 1.4019, + "theoretical_loss": 3.645333640741006, + "tokens_seen": 1098907648 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368906720160482, + "loss": 1.415, + "theoretical_loss": 3.645313470283065, + "tokens_seen": 1098973184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368806419257773, + "loss": 1.4491, + "theoretical_loss": 3.6452933013647035, + "tokens_seen": 1099038720 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033687061183550654, + "loss": 1.4066, + "theoretical_loss": 3.645273133985712, + "tokens_seen": 1099104256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368605817452357, + "loss": 1.4007, + "theoretical_loss": 3.6452529681458805, + "tokens_seen": 1099169792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368505516549649, + "loss": 1.2647, + "theoretical_loss": 3.6452328038450004, + "tokens_seen": 1099235328 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368405215646941, + "loss": 1.3584, + "theoretical_loss": 3.6452126410828622, + "tokens_seen": 1099300864 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368304914744233, + "loss": 1.5061, + "theoretical_loss": 3.645192479859257, + "tokens_seen": 1099366400 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033682046138415245, + "loss": 1.3136, + "theoretical_loss": 3.645172320173976, + "tokens_seen": 1099431936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368104312938817, + "loss": 1.5077, + "theoretical_loss": 3.6451521620268093, + "tokens_seen": 1099497472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003368004012036108, + "loss": 1.5082, + "theoretical_loss": 3.645132005417548, + "tokens_seen": 1099563008 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033679037111334004, + "loss": 1.4081, + "theoretical_loss": 3.645111850345984, + "tokens_seen": 1099628544 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003367803410230692, + "loss": 1.3485, + "theoretical_loss": 3.6450916968119076, + "tokens_seen": 1099694080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003367703109327984, + "loss": 1.462, + "theoretical_loss": 3.64507154481511, + "tokens_seen": 1099759616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003367602808425276, + "loss": 1.3907, + "theoretical_loss": 3.645051394355382, + "tokens_seen": 1099825152 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033675025075225677, + "loss": 1.5251, + "theoretical_loss": 3.6450312454325164, + "tokens_seen": 1099890688 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033674022066198595, + "loss": 1.4299, + "theoretical_loss": 3.645011098046302, + "tokens_seen": 1099956224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003367301905717152, + "loss": 1.3782, + "theoretical_loss": 3.6449909521965322, + "tokens_seen": 1100021760 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003367201604814443, + "loss": 1.4771, + "theoretical_loss": 3.6449708078829968, + "tokens_seen": 1100087296 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033671013039117355, + "loss": 1.3875, + "theoretical_loss": 3.644950665105488, + "tokens_seen": 1100152832 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033670010030090273, + "loss": 1.4784, + "theoretical_loss": 3.6449305238637972, + "tokens_seen": 1100218368 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366900702106319, + "loss": 1.4402, + "theoretical_loss": 3.644910384157715, + "tokens_seen": 1100283904 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366800401203611, + "loss": 1.3804, + "theoretical_loss": 3.6448902459870336, + "tokens_seen": 1100349440 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033667001003009027, + "loss": 1.3273, + "theoretical_loss": 3.6448701093515448, + "tokens_seen": 1100414976 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033665997993981945, + "loss": 1.4678, + "theoretical_loss": 3.644849974251039, + "tokens_seen": 1100480512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366499498495487, + "loss": 1.442, + "theoretical_loss": 3.6448298406853095, + "tokens_seen": 1100546048 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366399197592778, + "loss": 1.5557, + "theoretical_loss": 3.6448097086541456, + "tokens_seen": 1100611584 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033662988966900705, + "loss": 1.4851, + "theoretical_loss": 3.644789578157341, + "tokens_seen": 1100677120 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366198595787362, + "loss": 1.3544, + "theoretical_loss": 3.6447694491946865, + "tokens_seen": 1100742656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003366098294884654, + "loss": 1.495, + "theoretical_loss": 3.6447493217659743, + "tokens_seen": 1100808192 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365997993981946, + "loss": 1.5098, + "theoretical_loss": 3.6447291958709958, + "tokens_seen": 1100873728 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365897693079238, + "loss": 1.3547, + "theoretical_loss": 3.644709071509543, + "tokens_seen": 1100939264 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 403530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6016714572906494, + "objective/train/theoretical_loss": 3.6446889486814076, + "objective/train/tokens_used": 1121464800, + "theoretical_loss": 3.6446889486814076, + "tokens_seen": 1101004800 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033657973921765295, + "loss": 1.5421, + "theoretical_loss": 3.6446889486814076, + "tokens_seen": 1101004800 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033656970912738214, + "loss": 1.4082, + "theoretical_loss": 3.6446688273863814, + "tokens_seen": 1101070336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365596790371113, + "loss": 1.3949, + "theoretical_loss": 3.644648707624257, + "tokens_seen": 1101135872 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033654964894684055, + "loss": 1.4555, + "theoretical_loss": 3.6446285893948263, + "tokens_seen": 1101201408 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365396188565697, + "loss": 1.3378, + "theoretical_loss": 3.6446084726978807, + "tokens_seen": 1101266944 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365295887662989, + "loss": 1.3616, + "theoretical_loss": 3.6445883575332125, + "tokens_seen": 1101332480 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365195586760281, + "loss": 1.2864, + "theoretical_loss": 3.644568243900614, + "tokens_seen": 1101398016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003365095285857573, + "loss": 1.4239, + "theoretical_loss": 3.644548131799877, + "tokens_seen": 1101463552 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033649949849548646, + "loss": 1.4652, + "theoretical_loss": 3.6445280212307942, + "tokens_seen": 1101529088 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033648946840521564, + "loss": 1.4426, + "theoretical_loss": 3.644507912193158, + "tokens_seen": 1101594624 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003364794383149448, + "loss": 1.4796, + "theoretical_loss": 3.64448780468676, + "tokens_seen": 1101660160 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033646940822467406, + "loss": 1.29, + "theoretical_loss": 3.644467698711393, + "tokens_seen": 1101725696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003364593781344032, + "loss": 1.618, + "theoretical_loss": 3.644447594266849, + "tokens_seen": 1101791232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003364493480441324, + "loss": 1.4113, + "theoretical_loss": 3.644427491352921, + "tokens_seen": 1101856768 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033643931795386154, + "loss": 1.3298, + "theoretical_loss": 3.6444073899694005, + "tokens_seen": 1101922304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003364292878635908, + "loss": 1.2323, + "theoretical_loss": 3.6443872901160805, + "tokens_seen": 1101987840 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033641925777331996, + "loss": 1.4594, + "theoretical_loss": 3.644367191792754, + "tokens_seen": 1102053376 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033640922768304914, + "loss": 1.304, + "theoretical_loss": 3.6443470949992127, + "tokens_seen": 1102118912 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363991975927783, + "loss": 1.4138, + "theoretical_loss": 3.64432699973525, + "tokens_seen": 1102184448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363891675025075, + "loss": 1.3939, + "theoretical_loss": 3.644306906000658, + "tokens_seen": 1102249984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363791374122367, + "loss": 1.4778, + "theoretical_loss": 3.6442868137952296, + "tokens_seen": 1102315520 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363691073219659, + "loss": 1.5859, + "theoretical_loss": 3.6442667231187573, + "tokens_seen": 1102381056 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033635907723169505, + "loss": 1.3648, + "theoretical_loss": 3.644246633971034, + "tokens_seen": 1102446592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363490471414243, + "loss": 1.4557, + "theoretical_loss": 3.6442265463518533, + "tokens_seen": 1102512128 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033633901705115346, + "loss": 1.5547, + "theoretical_loss": 3.6442064602610063, + "tokens_seen": 1102577664 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033632898696088265, + "loss": 1.4319, + "theoretical_loss": 3.644186375698287, + "tokens_seen": 1102643200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003363189568706119, + "loss": 1.5002, + "theoretical_loss": 3.6441662926634883, + "tokens_seen": 1102708736 + }, + { + "epoch": 0.33, + "learning_rate": 0.000336308926780341, + "loss": 1.48, + "theoretical_loss": 3.6441462111564036, + "tokens_seen": 1102774272 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033629889669007024, + "loss": 1.487, + "theoretical_loss": 3.644126131176825, + "tokens_seen": 1102839808 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003362888665997994, + "loss": 1.3152, + "theoretical_loss": 3.6441060527245455, + "tokens_seen": 1102905344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003362788365095286, + "loss": 1.5754, + "theoretical_loss": 3.6440859757993596, + "tokens_seen": 1102970880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003362688064192578, + "loss": 1.5017, + "theoretical_loss": 3.644065900401059, + "tokens_seen": 1103036416 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033625877632898697, + "loss": 1.3998, + "theoretical_loss": 3.644045826529437, + "tokens_seen": 1103101952 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033624874623871615, + "loss": 1.3485, + "theoretical_loss": 3.644025754184287, + "tokens_seen": 1103167488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003362387161484454, + "loss": 1.4205, + "theoretical_loss": 3.644005683365403, + "tokens_seen": 1103233024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003362286860581745, + "loss": 1.4669, + "theoretical_loss": 3.643985614072577, + "tokens_seen": 1103298560 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033621865596790375, + "loss": 1.4122, + "theoretical_loss": 3.643965546305603, + "tokens_seen": 1103364096 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033620862587763293, + "loss": 1.3908, + "theoretical_loss": 3.643945480064275, + "tokens_seen": 1103429632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361985957873621, + "loss": 1.4876, + "theoretical_loss": 3.6439254153483853, + "tokens_seen": 1103495168 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361885656970913, + "loss": 1.3626, + "theoretical_loss": 3.643905352157728, + "tokens_seen": 1103560704 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033617853560682047, + "loss": 1.3928, + "theoretical_loss": 3.6438852904920966, + "tokens_seen": 1103626240 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033616850551654965, + "loss": 1.3807, + "theoretical_loss": 3.6438652303512846, + "tokens_seen": 1103691776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361584754262789, + "loss": 1.4076, + "theoretical_loss": 3.6438451717350846, + "tokens_seen": 1103757312 + }, + { + "epoch": 0.33, + "learning_rate": 0.000336148445336008, + "loss": 1.4306, + "theoretical_loss": 3.643825114643292, + "tokens_seen": 1103822848 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033613841524573725, + "loss": 1.4449, + "theoretical_loss": 3.6438050590756985, + "tokens_seen": 1103888384 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361283851554664, + "loss": 1.4286, + "theoretical_loss": 3.6437850050320995, + "tokens_seen": 1103953920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361183550651956, + "loss": 1.4228, + "theoretical_loss": 3.6437649525122877, + "tokens_seen": 1104019456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003361083249749248, + "loss": 1.4531, + "theoretical_loss": 3.6437449015160572, + "tokens_seen": 1104084992 + }, + { + "epoch": 0.33, + "learning_rate": 0.000336098294884654, + "loss": 1.4845, + "theoretical_loss": 3.643724852043202, + "tokens_seen": 1104150528 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033608826479438316, + "loss": 1.4902, + "theoretical_loss": 3.6437048040935154, + "tokens_seen": 1104216064 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 404699, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5483827590942383, + "objective/train/theoretical_loss": 3.643684757666792, + "objective/train/tokens_used": 1124741600, + "theoretical_loss": 3.643684757666792, + "tokens_seen": 1104281600 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033607823470411234, + "loss": 1.4711, + "theoretical_loss": 3.643684757666792, + "tokens_seen": 1104281600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003360682046138415, + "loss": 1.5178, + "theoretical_loss": 3.643664712762825, + "tokens_seen": 1104347136 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033605817452357075, + "loss": 1.469, + "theoretical_loss": 3.643644669381409, + "tokens_seen": 1104412672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003360481444332999, + "loss": 1.2671, + "theoretical_loss": 3.6436246275223376, + "tokens_seen": 1104478208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003360381143430291, + "loss": 1.4791, + "theoretical_loss": 3.643604587185405, + "tokens_seen": 1104543744 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003360280842527583, + "loss": 1.5099, + "theoretical_loss": 3.6435845483704057, + "tokens_seen": 1104609280 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003360180541624875, + "loss": 1.4733, + "theoretical_loss": 3.643564511077133, + "tokens_seen": 1104674816 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033600802407221666, + "loss": 1.4019, + "theoretical_loss": 3.643544475305382, + "tokens_seen": 1104740352 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033599799398194584, + "loss": 1.4387, + "theoretical_loss": 3.6435244410549457, + "tokens_seen": 1104805888 + }, + { + "epoch": 0.33, + "learning_rate": 0.000335987963891675, + "loss": 1.4183, + "theoretical_loss": 3.6435044083256196, + "tokens_seen": 1104871424 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033597793380140426, + "loss": 1.4472, + "theoretical_loss": 3.6434843771171974, + "tokens_seen": 1104936960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003359679037111334, + "loss": 1.4422, + "theoretical_loss": 3.6434643474294734, + "tokens_seen": 1105002496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003359578736208626, + "loss": 1.3394, + "theoretical_loss": 3.643444319262242, + "tokens_seen": 1105068032 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033594784353059174, + "loss": 1.4791, + "theoretical_loss": 3.643424292615298, + "tokens_seen": 1105133568 + }, + { + "epoch": 0.33, + "learning_rate": 0.000335937813440321, + "loss": 1.3844, + "theoretical_loss": 3.6434042674884357, + "tokens_seen": 1105199104 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033592778335005016, + "loss": 1.4654, + "theoretical_loss": 3.6433842438814494, + "tokens_seen": 1105264640 + }, + { + "epoch": 0.33, + "learning_rate": 0.00033591775325977934, + "loss": 1.4155, + "theoretical_loss": 3.6433642217941333, + "tokens_seen": 1105330176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003359077231695085, + "loss": 1.417, + "theoretical_loss": 3.6433442012262827, + "tokens_seen": 1105395712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0003358976930792377, + "loss": 1.6433, + "theoretical_loss": 3.6433241821776914, + "tokens_seen": 1105461248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003358876629889669, + "loss": 1.4894, + "theoretical_loss": 3.643304164648155, + "tokens_seen": 1105526784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003358776328986961, + "loss": 1.448, + "theoretical_loss": 3.643284148637467, + "tokens_seen": 1105592320 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033586760280842525, + "loss": 1.5375, + "theoretical_loss": 3.643264134145424, + "tokens_seen": 1105657856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003358575727181545, + "loss": 1.3565, + "theoretical_loss": 3.643244121171819, + "tokens_seen": 1105723392 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033584754262788366, + "loss": 1.4499, + "theoretical_loss": 3.6432241097164475, + "tokens_seen": 1105788928 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033583751253761285, + "loss": 1.3821, + "theoretical_loss": 3.643204099779104, + "tokens_seen": 1105854464 + }, + { + "epoch": 0.34, + "learning_rate": 0.000335827482447342, + "loss": 1.4107, + "theoretical_loss": 3.643184091359584, + "tokens_seen": 1105920000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003358174523570712, + "loss": 1.5181, + "theoretical_loss": 3.6431640844576823, + "tokens_seen": 1105985536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003358074222668004, + "loss": 1.354, + "theoretical_loss": 3.6431440790731933, + "tokens_seen": 1106051072 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003357973921765296, + "loss": 1.464, + "theoretical_loss": 3.6431240752059124, + "tokens_seen": 1106116608 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033578736208625875, + "loss": 1.294, + "theoretical_loss": 3.643104072855635, + "tokens_seen": 1106182144 + }, + { + "epoch": 0.34, + "learning_rate": 0.000335777331995988, + "loss": 1.4002, + "theoretical_loss": 3.6430840720221553, + "tokens_seen": 1106247680 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003357673019057171, + "loss": 1.4169, + "theoretical_loss": 3.6430640727052697, + "tokens_seen": 1106313216 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033575727181544635, + "loss": 1.3972, + "theoretical_loss": 3.643044074904772, + "tokens_seen": 1106378752 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033574724172517553, + "loss": 1.3707, + "theoretical_loss": 3.6430240786204586, + "tokens_seen": 1106444288 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003357372116349047, + "loss": 1.4651, + "theoretical_loss": 3.643004083852124, + "tokens_seen": 1106509824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003357271815446339, + "loss": 1.4125, + "theoretical_loss": 3.6429840905995636, + "tokens_seen": 1106575360 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033571715145436313, + "loss": 1.3949, + "theoretical_loss": 3.6429640988625724, + "tokens_seen": 1106640896 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033570712136409225, + "loss": 1.4079, + "theoretical_loss": 3.642944108640947, + "tokens_seen": 1106706432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356970912738215, + "loss": 1.4356, + "theoretical_loss": 3.6429241199344817, + "tokens_seen": 1106771968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356870611835506, + "loss": 1.5484, + "theoretical_loss": 3.6429041327429723, + "tokens_seen": 1106837504 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033567703109327985, + "loss": 1.4126, + "theoretical_loss": 3.642884147066214, + "tokens_seen": 1106903040 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033566700100300903, + "loss": 1.3996, + "theoretical_loss": 3.642864162904003, + "tokens_seen": 1106968576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356569709127382, + "loss": 1.4604, + "theoretical_loss": 3.6428441802561347, + "tokens_seen": 1107034112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356469408224674, + "loss": 1.4326, + "theoretical_loss": 3.642824199122404, + "tokens_seen": 1107099648 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356369107321966, + "loss": 1.415, + "theoretical_loss": 3.642804219502607, + "tokens_seen": 1107165184 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033562688064192576, + "loss": 1.4354, + "theoretical_loss": 3.6427842413965394, + "tokens_seen": 1107230720 + }, + { + "epoch": 0.34, + "learning_rate": 0.000335616850551655, + "loss": 1.4701, + "theoretical_loss": 3.642764264803997, + "tokens_seen": 1107296256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003356068204613841, + "loss": 1.3266, + "theoretical_loss": 3.6427442897247753, + "tokens_seen": 1107361792 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033559679037111336, + "loss": 1.3803, + "theoretical_loss": 3.6427243161586706, + "tokens_seen": 1107427328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355867602808425, + "loss": 1.348, + "theoretical_loss": 3.6427043441054785, + "tokens_seen": 1107492864 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 405946, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4452531337738037, + "objective/train/theoretical_loss": 3.642684373564995, + "objective/train/tokens_used": 1128018400, + "theoretical_loss": 3.642684373564995, + "tokens_seen": 1107558400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355767301905717, + "loss": 1.5025, + "theoretical_loss": 3.642684373564995, + "tokens_seen": 1107558400 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033556670010030095, + "loss": 1.499, + "theoretical_loss": 3.6426644045370153, + "tokens_seen": 1107623936 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355566700100301, + "loss": 1.4405, + "theoretical_loss": 3.6426444370213362, + "tokens_seen": 1107689472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355466399197593, + "loss": 1.5155, + "theoretical_loss": 3.6426244710177533, + "tokens_seen": 1107755008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355366098294885, + "loss": 1.4407, + "theoretical_loss": 3.6426045065260633, + "tokens_seen": 1107820544 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003355265797392177, + "loss": 1.4374, + "theoretical_loss": 3.642584543546061, + "tokens_seen": 1107886080 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033551654964894686, + "loss": 1.4683, + "theoretical_loss": 3.642564582077544, + "tokens_seen": 1107951616 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033550651955867604, + "loss": 1.3999, + "theoretical_loss": 3.6425446221203073, + "tokens_seen": 1108017152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354964894684052, + "loss": 1.3897, + "theoretical_loss": 3.6425246636741475, + "tokens_seen": 1108082688 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033548645937813446, + "loss": 1.4066, + "theoretical_loss": 3.6425047067388614, + "tokens_seen": 1108148224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354764292878636, + "loss": 1.3616, + "theoretical_loss": 3.6424847513142447, + "tokens_seen": 1108213760 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354663991975928, + "loss": 1.4609, + "theoretical_loss": 3.6424647974000934, + "tokens_seen": 1108279296 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033545636910732195, + "loss": 1.3947, + "theoretical_loss": 3.642444844996205, + "tokens_seen": 1108344832 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354463390170512, + "loss": 1.4156, + "theoretical_loss": 3.6424248941023745, + "tokens_seen": 1108410368 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033543630892678036, + "loss": 1.4417, + "theoretical_loss": 3.642404944718399, + "tokens_seen": 1108475904 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033542627883650954, + "loss": 1.367, + "theoretical_loss": 3.6423849968440756, + "tokens_seen": 1108541440 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354162487462387, + "loss": 1.3663, + "theoretical_loss": 3.6423650504791993, + "tokens_seen": 1108606976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003354062186559679, + "loss": 1.3607, + "theoretical_loss": 3.6423451056235683, + "tokens_seen": 1108672512 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353961885656971, + "loss": 1.4105, + "theoretical_loss": 3.642325162276978, + "tokens_seen": 1108738048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353861584754263, + "loss": 1.4063, + "theoretical_loss": 3.642305220439226, + "tokens_seen": 1108803584 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033537612838515545, + "loss": 1.491, + "theoretical_loss": 3.6422852801101078, + "tokens_seen": 1108869120 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353660982948847, + "loss": 1.4882, + "theoretical_loss": 3.6422653412894217, + "tokens_seen": 1108934656 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033535606820461386, + "loss": 1.4018, + "theoretical_loss": 3.642245403976963, + "tokens_seen": 1109000192 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033534603811434305, + "loss": 1.4638, + "theoretical_loss": 3.642225468172529, + "tokens_seen": 1109065728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353360080240722, + "loss": 1.3134, + "theoretical_loss": 3.6422055338759165, + "tokens_seen": 1109131264 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353259779338014, + "loss": 1.4748, + "theoretical_loss": 3.642185601086923, + "tokens_seen": 1109196800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353159478435306, + "loss": 1.3503, + "theoretical_loss": 3.642165669805344, + "tokens_seen": 1109262336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003353059177532598, + "loss": 1.4979, + "theoretical_loss": 3.642145740030978, + "tokens_seen": 1109327872 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033529588766298895, + "loss": 1.4591, + "theoretical_loss": 3.6421258117636217, + "tokens_seen": 1109393408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003352858575727182, + "loss": 1.5485, + "theoretical_loss": 3.6421058850030708, + "tokens_seen": 1109458944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003352758274824473, + "loss": 1.4874, + "theoretical_loss": 3.642085959749124, + "tokens_seen": 1109524480 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033526579739217655, + "loss": 1.4655, + "theoretical_loss": 3.6420660360015775, + "tokens_seen": 1109590016 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033525576730190573, + "loss": 1.5234, + "theoretical_loss": 3.6420461137602285, + "tokens_seen": 1109655552 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003352457372116349, + "loss": 1.4302, + "theoretical_loss": 3.6420261930248747, + "tokens_seen": 1109721088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003352357071213641, + "loss": 1.4867, + "theoretical_loss": 3.642006273795313, + "tokens_seen": 1109786624 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033522567703109333, + "loss": 1.393, + "theoretical_loss": 3.64198635607134, + "tokens_seen": 1109852160 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033521564694082245, + "loss": 1.4364, + "theoretical_loss": 3.6419664398527543, + "tokens_seen": 1109917696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003352056168505517, + "loss": 1.4109, + "theoretical_loss": 3.6419465251393524, + "tokens_seen": 1109983232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351955867602808, + "loss": 1.5234, + "theoretical_loss": 3.641926611930932, + "tokens_seen": 1110048768 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033518555667001005, + "loss": 1.5082, + "theoretical_loss": 3.64190670022729, + "tokens_seen": 1110114304 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033517552657973923, + "loss": 1.4702, + "theoretical_loss": 3.641886790028225, + "tokens_seen": 1110179840 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351654964894684, + "loss": 1.3644, + "theoretical_loss": 3.6418668813335335, + "tokens_seen": 1110245376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351554663991976, + "loss": 1.5181, + "theoretical_loss": 3.6418469741430135, + "tokens_seen": 1110310912 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351454363089268, + "loss": 1.498, + "theoretical_loss": 3.641827068456462, + "tokens_seen": 1110376448 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033513540621865596, + "loss": 1.4561, + "theoretical_loss": 3.6418071642736773, + "tokens_seen": 1110441984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351253761283852, + "loss": 1.4628, + "theoretical_loss": 3.6417872615944566, + "tokens_seen": 1110507520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003351153460381143, + "loss": 1.2958, + "theoretical_loss": 3.641767360418598, + "tokens_seen": 1110573056 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033510531594784356, + "loss": 1.496, + "theoretical_loss": 3.641747460745899, + "tokens_seen": 1110638592 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350952858575727, + "loss": 1.4212, + "theoretical_loss": 3.6417275625761576, + "tokens_seen": 1110704128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350852557673019, + "loss": 1.459, + "theoretical_loss": 3.6417076659091707, + "tokens_seen": 1110769664 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 406993, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4353649616241455, + "objective/train/theoretical_loss": 3.6416877707447375, + "objective/train/tokens_used": 1131295200, + "theoretical_loss": 3.6416877707447375, + "tokens_seen": 1110835200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350752256770311, + "loss": 1.397, + "theoretical_loss": 3.6416877707447375, + "tokens_seen": 1110835200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350651955867603, + "loss": 1.4053, + "theoretical_loss": 3.6416678770826554, + "tokens_seen": 1110900736 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033505516549648946, + "loss": 1.5355, + "theoretical_loss": 3.641647984922722, + "tokens_seen": 1110966272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350451354062187, + "loss": 1.5963, + "theoretical_loss": 3.6416280942647354, + "tokens_seen": 1111031808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350351053159478, + "loss": 1.3895, + "theoretical_loss": 3.6416082051084935, + "tokens_seen": 1111097344 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033502507522567706, + "loss": 1.413, + "theoretical_loss": 3.641588317453795, + "tokens_seen": 1111162880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350150451354062, + "loss": 1.5219, + "theoretical_loss": 3.641568431300437, + "tokens_seen": 1111228416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003350050150451354, + "loss": 1.4829, + "theoretical_loss": 3.6415485466482185, + "tokens_seen": 1111293952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349949849548646, + "loss": 1.6084, + "theoretical_loss": 3.6415286634969375, + "tokens_seen": 1111359488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349849548645938, + "loss": 1.4698, + "theoretical_loss": 3.6415087818463916, + "tokens_seen": 1111425024 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033497492477432296, + "loss": 1.4886, + "theoretical_loss": 3.6414889016963796, + "tokens_seen": 1111490560 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033496489468405215, + "loss": 1.4815, + "theoretical_loss": 3.6414690230467, + "tokens_seen": 1111556096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349548645937813, + "loss": 1.4862, + "theoretical_loss": 3.6414491458971505, + "tokens_seen": 1111621632 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033494483450351056, + "loss": 1.4457, + "theoretical_loss": 3.6414292702475297, + "tokens_seen": 1111687168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349348044132397, + "loss": 1.4288, + "theoretical_loss": 3.641409396097636, + "tokens_seen": 1111752704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349247743229689, + "loss": 1.5252, + "theoretical_loss": 3.641389523447268, + "tokens_seen": 1111818240 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033491474423269805, + "loss": 1.4985, + "theoretical_loss": 3.6413696522962242, + "tokens_seen": 1111883776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003349047141424273, + "loss": 1.4214, + "theoretical_loss": 3.641349782644303, + "tokens_seen": 1111949312 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033489468405215647, + "loss": 1.4239, + "theoretical_loss": 3.641329914491303, + "tokens_seen": 1112014848 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033488465396188565, + "loss": 1.4677, + "theoretical_loss": 3.6413100478370226, + "tokens_seen": 1112080384 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033487462387161483, + "loss": 1.3964, + "theoretical_loss": 3.6412901826812605, + "tokens_seen": 1112145920 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033486459378134406, + "loss": 1.5062, + "theoretical_loss": 3.641270319023816, + "tokens_seen": 1112211456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003348545636910732, + "loss": 1.522, + "theoretical_loss": 3.6412504568644866, + "tokens_seen": 1112276992 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003348445336008024, + "loss": 1.3992, + "theoretical_loss": 3.641230596203072, + "tokens_seen": 1112342528 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003348345035105316, + "loss": 1.4342, + "theoretical_loss": 3.6412107370393705, + "tokens_seen": 1112408064 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003348244734202608, + "loss": 1.3256, + "theoretical_loss": 3.6411908793731813, + "tokens_seen": 1112473600 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033481444332999, + "loss": 1.4951, + "theoretical_loss": 3.6411710232043033, + "tokens_seen": 1112539136 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033480441323971915, + "loss": 1.4246, + "theoretical_loss": 3.641151168532535, + "tokens_seen": 1112604672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003347943831494484, + "loss": 1.4539, + "theoretical_loss": 3.6411313153576756, + "tokens_seen": 1112670208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003347843530591775, + "loss": 1.4997, + "theoretical_loss": 3.6411114636795237, + "tokens_seen": 1112735744 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033477432296890675, + "loss": 1.4478, + "theoretical_loss": 3.641091613497879, + "tokens_seen": 1112801280 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033476429287863593, + "loss": 1.481, + "theoretical_loss": 3.6410717648125406, + "tokens_seen": 1112866816 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003347542627883651, + "loss": 1.4623, + "theoretical_loss": 3.6410519176233067, + "tokens_seen": 1112932352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003347442326980943, + "loss": 1.3353, + "theoretical_loss": 3.641032071929977, + "tokens_seen": 1112997888 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033473420260782353, + "loss": 1.3703, + "theoretical_loss": 3.641012227732351, + "tokens_seen": 1113063424 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033472417251755265, + "loss": 1.4477, + "theoretical_loss": 3.6409923850302275, + "tokens_seen": 1113128960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003347141424272819, + "loss": 1.4272, + "theoretical_loss": 3.640972543823406, + "tokens_seen": 1113194496 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334704112337011, + "loss": 1.4133, + "theoretical_loss": 3.6409527041116854, + "tokens_seen": 1113260032 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033469408224674025, + "loss": 1.4609, + "theoretical_loss": 3.640932865894865, + "tokens_seen": 1113325568 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033468405215646943, + "loss": 1.6021, + "theoretical_loss": 3.640913029172745, + "tokens_seen": 1113391104 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003346740220661986, + "loss": 1.3214, + "theoretical_loss": 3.640893193945124, + "tokens_seen": 1113456640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003346639919759278, + "loss": 1.4755, + "theoretical_loss": 3.6408733602118017, + "tokens_seen": 1113522176 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334653961885657, + "loss": 1.5697, + "theoretical_loss": 3.6408535279725776, + "tokens_seen": 1113587712 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033464393179538616, + "loss": 1.473, + "theoretical_loss": 3.640833697227251, + "tokens_seen": 1113653248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003346339017051154, + "loss": 1.4293, + "theoretical_loss": 3.640813867975622, + "tokens_seen": 1113718784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003346238716148445, + "loss": 1.5741, + "theoretical_loss": 3.6407940402174894, + "tokens_seen": 1113784320 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033461384152457376, + "loss": 1.5801, + "theoretical_loss": 3.640774213952654, + "tokens_seen": 1113849856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003346038114343029, + "loss": 1.4788, + "theoretical_loss": 3.6407543891809144, + "tokens_seen": 1113915392 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345937813440321, + "loss": 1.383, + "theoretical_loss": 3.640734565902071, + "tokens_seen": 1113980928 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345837512537613, + "loss": 1.4171, + "theoretical_loss": 3.6407147441159236, + "tokens_seen": 1114046464 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 408181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.395033597946167, + "objective/train/theoretical_loss": 3.640694923822271, + "objective/train/tokens_used": 1134572000, + "theoretical_loss": 3.640694923822271, + "tokens_seen": 1114112000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345737211634905, + "loss": 1.356, + "theoretical_loss": 3.640694923822271, + "tokens_seen": 1114112000 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033456369107321966, + "loss": 1.4403, + "theoretical_loss": 3.6406751050209145, + "tokens_seen": 1114177536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345536609829489, + "loss": 1.4942, + "theoretical_loss": 3.6406552877116525, + "tokens_seen": 1114243072 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334543630892678, + "loss": 1.4293, + "theoretical_loss": 3.640635471894286, + "tokens_seen": 1114308608 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033453360080240726, + "loss": 1.4682, + "theoretical_loss": 3.640615657568614, + "tokens_seen": 1114374144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345235707121364, + "loss": 1.4025, + "theoretical_loss": 3.640595844734438, + "tokens_seen": 1114439680 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345135406218656, + "loss": 1.531, + "theoretical_loss": 3.6405760333915564, + "tokens_seen": 1114505216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003345035105315948, + "loss": 1.4192, + "theoretical_loss": 3.640556223539771, + "tokens_seen": 1114570752 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334493480441324, + "loss": 1.688, + "theoretical_loss": 3.6405364151788797, + "tokens_seen": 1114636288 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033448345035105316, + "loss": 1.5489, + "theoretical_loss": 3.640516608308684, + "tokens_seen": 1114701824 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033447342026078235, + "loss": 1.4458, + "theoretical_loss": 3.6404968029289844, + "tokens_seen": 1114767360 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003344633901705115, + "loss": 1.3917, + "theoretical_loss": 3.6404769990395804, + "tokens_seen": 1114832896 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033445336008024076, + "loss": 1.4752, + "theoretical_loss": 3.6404571966402726, + "tokens_seen": 1114898432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003344433299899699, + "loss": 1.504, + "theoretical_loss": 3.640437395730861, + "tokens_seen": 1114963968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003344332998996991, + "loss": 1.4131, + "theoretical_loss": 3.6404175963111465, + "tokens_seen": 1115029504 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033442326980942825, + "loss": 1.3782, + "theoretical_loss": 3.640397798380929, + "tokens_seen": 1115095040 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003344132397191575, + "loss": 1.3015, + "theoretical_loss": 3.640378001940009, + "tokens_seen": 1115160576 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033440320962888667, + "loss": 1.4146, + "theoretical_loss": 3.6403582069881866, + "tokens_seen": 1115226112 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033439317953861585, + "loss": 1.4828, + "theoretical_loss": 3.640338413525263, + "tokens_seen": 1115291648 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033438314944834503, + "loss": 1.3797, + "theoretical_loss": 3.640318621551039, + "tokens_seen": 1115357184 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033437311935807427, + "loss": 1.3301, + "theoretical_loss": 3.640298831065314, + "tokens_seen": 1115422720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003343630892678034, + "loss": 1.4401, + "theoretical_loss": 3.6402790420678888, + "tokens_seen": 1115488256 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033435305917753263, + "loss": 1.4786, + "theoretical_loss": 3.640259254558565, + "tokens_seen": 1115553792 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033434302908726175, + "loss": 1.5798, + "theoretical_loss": 3.6402394685371426, + "tokens_seen": 1115619328 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334332998996991, + "loss": 1.5425, + "theoretical_loss": 3.640219684003422, + "tokens_seen": 1115684864 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033432296890672017, + "loss": 1.4359, + "theoretical_loss": 3.640199900957205, + "tokens_seen": 1115750400 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033431293881644935, + "loss": 1.4483, + "theoretical_loss": 3.6401801193982917, + "tokens_seen": 1115815936 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033430290872617853, + "loss": 1.4612, + "theoretical_loss": 3.6401603393264828, + "tokens_seen": 1115881472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342928786359077, + "loss": 1.4182, + "theoretical_loss": 3.6401405607415795, + "tokens_seen": 1115947008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342828485456369, + "loss": 1.4231, + "theoretical_loss": 3.6401207836433827, + "tokens_seen": 1116012544 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033427281845536613, + "loss": 1.3895, + "theoretical_loss": 3.640101008031693, + "tokens_seen": 1116078080 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033426278836509526, + "loss": 1.3198, + "theoretical_loss": 3.6400812339063124, + "tokens_seen": 1116143616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342527582748245, + "loss": 1.3646, + "theoretical_loss": 3.6400614612670408, + "tokens_seen": 1116209152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342427281845536, + "loss": 1.4875, + "theoretical_loss": 3.6400416901136796, + "tokens_seen": 1116274688 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033423269809428285, + "loss": 1.522, + "theoretical_loss": 3.6400219204460305, + "tokens_seen": 1116340224 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033422266800401204, + "loss": 1.4348, + "theoretical_loss": 3.6400021522638935, + "tokens_seen": 1116405760 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342126379137412, + "loss": 1.385, + "theoretical_loss": 3.6399823855670705, + "tokens_seen": 1116471296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003342026078234704, + "loss": 1.4989, + "theoretical_loss": 3.6399626203553628, + "tokens_seen": 1116536832 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033419257773319963, + "loss": 1.4209, + "theoretical_loss": 3.639942856628571, + "tokens_seen": 1116602368 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033418254764292876, + "loss": 1.4956, + "theoretical_loss": 3.6399230943864977, + "tokens_seen": 1116667904 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334172517552658, + "loss": 1.497, + "theoretical_loss": 3.639903333628943, + "tokens_seen": 1116733440 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003341624874623871, + "loss": 1.3508, + "theoretical_loss": 3.639883574355709, + "tokens_seen": 1116798976 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033415245737211636, + "loss": 1.5073, + "theoretical_loss": 3.6398638165665966, + "tokens_seen": 1116864512 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033414242728184554, + "loss": 1.4659, + "theoretical_loss": 3.6398440602614075, + "tokens_seen": 1116930048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003341323971915747, + "loss": 1.4243, + "theoretical_loss": 3.6398243054399426, + "tokens_seen": 1116995584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003341223671013039, + "loss": 1.379, + "theoretical_loss": 3.639804552102005, + "tokens_seen": 1117061120 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003341123370110331, + "loss": 1.4066, + "theoretical_loss": 3.6397848002473943, + "tokens_seen": 1117126656 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033410230692076226, + "loss": 1.4238, + "theoretical_loss": 3.6397650498759138, + "tokens_seen": 1117192192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340922768304915, + "loss": 1.4333, + "theoretical_loss": 3.639745300987364, + "tokens_seen": 1117257728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340822467402207, + "loss": 1.3821, + "theoretical_loss": 3.639725553581547, + "tokens_seen": 1117323264 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 409409, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5067861080169678, + "objective/train/theoretical_loss": 3.639705807658265, + "objective/train/tokens_used": 1137848800, + "theoretical_loss": 3.639705807658265, + "tokens_seen": 1117388800 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033407221664994986, + "loss": 1.4006, + "theoretical_loss": 3.639705807658265, + "tokens_seen": 1117388800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340621865596791, + "loss": 1.4908, + "theoretical_loss": 3.6396860632173187, + "tokens_seen": 1117454336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340521564694082, + "loss": 1.4059, + "theoretical_loss": 3.6396663202585104, + "tokens_seen": 1117519872 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033404212637913746, + "loss": 1.4186, + "theoretical_loss": 3.639646578781642, + "tokens_seen": 1117585408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340320962888666, + "loss": 1.4678, + "theoretical_loss": 3.6396268387865156, + "tokens_seen": 1117650944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340220661985958, + "loss": 1.5297, + "theoretical_loss": 3.639607100272933, + "tokens_seen": 1117716480 + }, + { + "epoch": 0.34, + "learning_rate": 0.000334012036108325, + "loss": 1.5176, + "theoretical_loss": 3.639587363240696, + "tokens_seen": 1117782016 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003340020060180542, + "loss": 1.4103, + "theoretical_loss": 3.6395676276896065, + "tokens_seen": 1117847552 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033399197592778336, + "loss": 1.422, + "theoretical_loss": 3.6395478936194663, + "tokens_seen": 1117913088 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033398194583751255, + "loss": 1.5062, + "theoretical_loss": 3.6395281610300776, + "tokens_seen": 1117978624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003339719157472417, + "loss": 1.3991, + "theoretical_loss": 3.6395084299212432, + "tokens_seen": 1118044160 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033396188565697096, + "loss": 1.454, + "theoretical_loss": 3.6394887002927647, + "tokens_seen": 1118109696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003339518555667001, + "loss": 1.4605, + "theoretical_loss": 3.639468972144444, + "tokens_seen": 1118175232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003339418254764293, + "loss": 1.4057, + "theoretical_loss": 3.639449245476084, + "tokens_seen": 1118240768 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033393179538615845, + "loss": 1.4338, + "theoretical_loss": 3.6394295202874862, + "tokens_seen": 1118306304 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003339217652958877, + "loss": 1.4811, + "theoretical_loss": 3.6394097965784535, + "tokens_seen": 1118371840 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033391173520561687, + "loss": 1.4283, + "theoretical_loss": 3.6393900743487877, + "tokens_seen": 1118437376 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033390170511534605, + "loss": 1.5318, + "theoretical_loss": 3.6393703535982915, + "tokens_seen": 1118502912 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033389167502507523, + "loss": 1.4073, + "theoretical_loss": 3.6393506343267674, + "tokens_seen": 1118568448 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033388164493480447, + "loss": 1.4275, + "theoretical_loss": 3.6393309165340177, + "tokens_seen": 1118633984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003338716148445336, + "loss": 1.3934, + "theoretical_loss": 3.6393112002198444, + "tokens_seen": 1118699520 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033386158475426283, + "loss": 1.5045, + "theoretical_loss": 3.6392914853840512, + "tokens_seen": 1118765056 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033385155466399195, + "loss": 1.4286, + "theoretical_loss": 3.639271772026439, + "tokens_seen": 1118830592 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003338415245737212, + "loss": 1.5504, + "theoretical_loss": 3.6392520601468124, + "tokens_seen": 1118896128 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033383149448345037, + "loss": 1.3706, + "theoretical_loss": 3.6392323497449723, + "tokens_seen": 1118961664 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033382146439317955, + "loss": 1.4565, + "theoretical_loss": 3.6392126408207224, + "tokens_seen": 1119027200 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033381143430290873, + "loss": 1.3793, + "theoretical_loss": 3.6391929333738644, + "tokens_seen": 1119092736 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003338014042126379, + "loss": 1.4738, + "theoretical_loss": 3.639173227404202, + "tokens_seen": 1119158272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003337913741223671, + "loss": 1.4974, + "theoretical_loss": 3.639153522911538, + "tokens_seen": 1119223808 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033378134403209633, + "loss": 1.3909, + "theoretical_loss": 3.6391338198956746, + "tokens_seen": 1119289344 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033377131394182546, + "loss": 1.5297, + "theoretical_loss": 3.639114118356415, + "tokens_seen": 1119354880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003337612838515547, + "loss": 1.5214, + "theoretical_loss": 3.6390944182935616, + "tokens_seen": 1119420416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003337512537612838, + "loss": 1.5538, + "theoretical_loss": 3.639074719706919, + "tokens_seen": 1119485952 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033374122367101306, + "loss": 1.413, + "theoretical_loss": 3.6390550225962874, + "tokens_seen": 1119551488 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033373119358074224, + "loss": 1.463, + "theoretical_loss": 3.6390353269614724, + "tokens_seen": 1119617024 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003337211634904714, + "loss": 1.5214, + "theoretical_loss": 3.6390156328022756, + "tokens_seen": 1119682560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003337111334002006, + "loss": 1.4071, + "theoretical_loss": 3.638995940118501, + "tokens_seen": 1119748096 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033370110330992983, + "loss": 1.4428, + "theoretical_loss": 3.6389762489099504, + "tokens_seen": 1119813632 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033369107321965896, + "loss": 1.4902, + "theoretical_loss": 3.638956559176428, + "tokens_seen": 1119879168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336810431293882, + "loss": 1.4682, + "theoretical_loss": 3.638936870917737, + "tokens_seen": 1119944704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336710130391173, + "loss": 1.5261, + "theoretical_loss": 3.63891718413368, + "tokens_seen": 1120010240 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033366098294884656, + "loss": 1.3806, + "theoretical_loss": 3.638897498824061, + "tokens_seen": 1120075776 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033365095285857574, + "loss": 1.541, + "theoretical_loss": 3.6388778149886836, + "tokens_seen": 1120141312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336409227683049, + "loss": 1.4533, + "theoretical_loss": 3.63885813262735, + "tokens_seen": 1120206848 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336308926780341, + "loss": 1.3997, + "theoretical_loss": 3.638838451739864, + "tokens_seen": 1120272384 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336208625877633, + "loss": 1.3703, + "theoretical_loss": 3.6388187723260286, + "tokens_seen": 1120337920 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033361083249749246, + "loss": 1.4672, + "theoretical_loss": 3.6387990943856487, + "tokens_seen": 1120403456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003336008024072217, + "loss": 1.4632, + "theoretical_loss": 3.638779417918527, + "tokens_seen": 1120468992 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003335907723169508, + "loss": 1.3108, + "theoretical_loss": 3.6387597429244662, + "tokens_seen": 1120534528 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033358074222668006, + "loss": 1.4718, + "theoretical_loss": 3.638740069403271, + "tokens_seen": 1120600064 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 410712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4078783988952637, + "objective/train/theoretical_loss": 3.638720397354745, + "objective/train/tokens_used": 1141125600, + "theoretical_loss": 3.638720397354745, + "tokens_seen": 1120665600 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033357071213640924, + "loss": 1.4183, + "theoretical_loss": 3.638720397354745, + "tokens_seen": 1120665600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003335606820461384, + "loss": 1.4991, + "theoretical_loss": 3.638700726778691, + "tokens_seen": 1120731136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003335506519558676, + "loss": 1.4817, + "theoretical_loss": 3.6386810576749133, + "tokens_seen": 1120796672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003335406218655968, + "loss": 1.4066, + "theoretical_loss": 3.638661390043216, + "tokens_seen": 1120862208 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033353059177532597, + "loss": 1.3564, + "theoretical_loss": 3.638641723883402, + "tokens_seen": 1120927744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003335205616850552, + "loss": 1.5172, + "theoretical_loss": 3.638622059195275, + "tokens_seen": 1120993280 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033351053159478433, + "loss": 1.5042, + "theoretical_loss": 3.6386023959786398, + "tokens_seen": 1121058816 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033350050150451356, + "loss": 1.4348, + "theoretical_loss": 3.6385827342333, + "tokens_seen": 1121124352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003334904714142427, + "loss": 1.5063, + "theoretical_loss": 3.638563073959059, + "tokens_seen": 1121189888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003334804413239719, + "loss": 1.3349, + "theoretical_loss": 3.638543415155721, + "tokens_seen": 1121255424 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003334704112337011, + "loss": 1.5121, + "theoretical_loss": 3.6385237578230907, + "tokens_seen": 1121320960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003334603811434303, + "loss": 1.4195, + "theoretical_loss": 3.6385041019609705, + "tokens_seen": 1121386496 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033345035105315947, + "loss": 1.4333, + "theoretical_loss": 3.6384844475691662, + "tokens_seen": 1121452032 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033344032096288865, + "loss": 1.473, + "theoretical_loss": 3.638464794647481, + "tokens_seen": 1121517568 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033343029087261783, + "loss": 1.4094, + "theoretical_loss": 3.638445143195719, + "tokens_seen": 1121583104 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033342026078234707, + "loss": 1.4196, + "theoretical_loss": 3.638425493213685, + "tokens_seen": 1121648640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003334102306920762, + "loss": 1.3312, + "theoretical_loss": 3.6384058447011824, + "tokens_seen": 1121714176 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033340020060180543, + "loss": 1.6346, + "theoretical_loss": 3.638386197658016, + "tokens_seen": 1121779712 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003333901705115346, + "loss": 1.3984, + "theoretical_loss": 3.6383665520839896, + "tokens_seen": 1121845248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003333801404212638, + "loss": 1.4223, + "theoretical_loss": 3.638346907978908, + "tokens_seen": 1121910784 + }, + { + "epoch": 0.34, + "learning_rate": 0.000333370110330993, + "loss": 1.5011, + "theoretical_loss": 3.638327265342576, + "tokens_seen": 1121976320 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033336008024072215, + "loss": 1.2938, + "theoretical_loss": 3.6383076241747965, + "tokens_seen": 1122041856 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033335005015045134, + "loss": 1.5034, + "theoretical_loss": 3.638287984475375, + "tokens_seen": 1122107392 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033334002006018057, + "loss": 1.5406, + "theoretical_loss": 3.6382683462441165, + "tokens_seen": 1122172928 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033332998996990975, + "loss": 1.5099, + "theoretical_loss": 3.6382487094808242, + "tokens_seen": 1122238464 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033331995987963893, + "loss": 1.445, + "theoretical_loss": 3.638229074185304, + "tokens_seen": 1122304000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003333099297893681, + "loss": 1.509, + "theoretical_loss": 3.6382094403573593, + "tokens_seen": 1122369536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003332998996990973, + "loss": 1.3946, + "theoretical_loss": 3.6381898079967954, + "tokens_seen": 1122435072 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033328986960882653, + "loss": 1.4076, + "theoretical_loss": 3.6381701771034165, + "tokens_seen": 1122500608 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033327983951855566, + "loss": 1.4364, + "theoretical_loss": 3.6381505476770277, + "tokens_seen": 1122566144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003332698094282849, + "loss": 1.4163, + "theoretical_loss": 3.638130919717434, + "tokens_seen": 1122631680 + }, + { + "epoch": 0.34, + "learning_rate": 0.000333259779338014, + "loss": 1.3417, + "theoretical_loss": 3.6381112932244397, + "tokens_seen": 1122697216 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033324974924774326, + "loss": 1.4419, + "theoretical_loss": 3.6380916681978492, + "tokens_seen": 1122762752 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033323971915747244, + "loss": 1.4773, + "theoretical_loss": 3.6380720446374686, + "tokens_seen": 1122828288 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003332296890672016, + "loss": 1.4551, + "theoretical_loss": 3.638052422543102, + "tokens_seen": 1122893824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003332196589769308, + "loss": 1.5176, + "theoretical_loss": 3.6380328019145542, + "tokens_seen": 1122959360 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033320962888666003, + "loss": 1.48, + "theoretical_loss": 3.63801318275163, + "tokens_seen": 1123024896 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033319959879638916, + "loss": 1.3358, + "theoretical_loss": 3.637993565054135, + "tokens_seen": 1123090432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331895687061184, + "loss": 1.3324, + "theoretical_loss": 3.637973948821874, + "tokens_seen": 1123155968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331795386158475, + "loss": 1.4526, + "theoretical_loss": 3.637954334054652, + "tokens_seen": 1123221504 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033316950852557676, + "loss": 1.4089, + "theoretical_loss": 3.6379347207522748, + "tokens_seen": 1123287040 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033315947843530594, + "loss": 1.412, + "theoretical_loss": 3.6379151089145463, + "tokens_seen": 1123352576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331494483450351, + "loss": 1.4281, + "theoretical_loss": 3.6378954985412726, + "tokens_seen": 1123418112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331394182547643, + "loss": 1.4119, + "theoretical_loss": 3.6378758896322587, + "tokens_seen": 1123483648 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331293881644935, + "loss": 1.4333, + "theoretical_loss": 3.6378562821873097, + "tokens_seen": 1123549184 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033311935807422266, + "loss": 1.4741, + "theoretical_loss": 3.6378366762062306, + "tokens_seen": 1123614720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003331093279839519, + "loss": 1.4069, + "theoretical_loss": 3.637817071688828, + "tokens_seen": 1123680256 + }, + { + "epoch": 0.34, + "learning_rate": 0.000333099297893681, + "loss": 1.5144, + "theoretical_loss": 3.637797468634906, + "tokens_seen": 1123745792 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033308926780341026, + "loss": 1.4058, + "theoretical_loss": 3.63777786704427, + "tokens_seen": 1123811328 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033307923771313944, + "loss": 1.5352, + "theoretical_loss": 3.6377582669167268, + "tokens_seen": 1123876864 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 411904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2147531509399414, + "objective/train/theoretical_loss": 3.63773866825208, + "objective/train/tokens_used": 1144402400, + "theoretical_loss": 3.63773866825208, + "tokens_seen": 1123942400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003330692076228686, + "loss": 1.3082, + "theoretical_loss": 3.63773866825208, + "tokens_seen": 1123942400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003330591775325978, + "loss": 1.4711, + "theoretical_loss": 3.637719071050137, + "tokens_seen": 1124007936 + }, + { + "epoch": 0.34, + "learning_rate": 0.000333049147442327, + "loss": 1.4354, + "theoretical_loss": 3.6376994753107015, + "tokens_seen": 1124073472 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033303911735205617, + "loss": 1.3784, + "theoretical_loss": 3.6376798810335806, + "tokens_seen": 1124139008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003330290872617854, + "loss": 1.4462, + "theoretical_loss": 3.6376602882185796, + "tokens_seen": 1124204544 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033301905717151453, + "loss": 1.465, + "theoretical_loss": 3.6376406968655037, + "tokens_seen": 1124270080 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033300902708124376, + "loss": 1.3855, + "theoretical_loss": 3.6376211069741586, + "tokens_seen": 1124335616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003329989969909729, + "loss": 1.4451, + "theoretical_loss": 3.6376015185443507, + "tokens_seen": 1124401152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003329889669007021, + "loss": 1.54, + "theoretical_loss": 3.6375819315758853, + "tokens_seen": 1124466688 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003329789368104313, + "loss": 1.4634, + "theoretical_loss": 3.6375623460685684, + "tokens_seen": 1124532224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003329689067201605, + "loss": 1.4088, + "theoretical_loss": 3.6375427620222056, + "tokens_seen": 1124597760 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033295887662988967, + "loss": 1.3699, + "theoretical_loss": 3.637523179436603, + "tokens_seen": 1124663296 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033294884653961885, + "loss": 1.3759, + "theoretical_loss": 3.6375035983115667, + "tokens_seen": 1124728832 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033293881644934803, + "loss": 1.4159, + "theoretical_loss": 3.6374840186469024, + "tokens_seen": 1124794368 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033292878635907727, + "loss": 1.3582, + "theoretical_loss": 3.637464440442416, + "tokens_seen": 1124859904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003329187562688064, + "loss": 1.4112, + "theoretical_loss": 3.6374448636979135, + "tokens_seen": 1124925440 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033290872617853563, + "loss": 1.4347, + "theoretical_loss": 3.6374252884132017, + "tokens_seen": 1124990976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003328986960882648, + "loss": 1.4909, + "theoretical_loss": 3.637405714588086, + "tokens_seen": 1125056512 + }, + { + "epoch": 0.34, + "learning_rate": 0.000332888665997994, + "loss": 1.3764, + "theoretical_loss": 3.6373861422223728, + "tokens_seen": 1125122048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003328786359077232, + "loss": 1.5649, + "theoretical_loss": 3.637366571315868, + "tokens_seen": 1125187584 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033286860581745235, + "loss": 1.413, + "theoretical_loss": 3.6373470018683784, + "tokens_seen": 1125253120 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033285857572718154, + "loss": 1.5058, + "theoretical_loss": 3.63732743387971, + "tokens_seen": 1125318656 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033284854563691077, + "loss": 1.3558, + "theoretical_loss": 3.637307867349669, + "tokens_seen": 1125384192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003328385155466399, + "loss": 1.4278, + "theoretical_loss": 3.637288302278061, + "tokens_seen": 1125449728 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033282848545636913, + "loss": 1.4923, + "theoretical_loss": 3.6372687386646945, + "tokens_seen": 1125515264 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033281845536609826, + "loss": 1.3603, + "theoretical_loss": 3.637249176509374, + "tokens_seen": 1125580800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003328084252758275, + "loss": 1.4356, + "theoretical_loss": 3.6372296158119064, + "tokens_seen": 1125646336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003327983951855567, + "loss": 1.4431, + "theoretical_loss": 3.637210056572098, + "tokens_seen": 1125711872 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033278836509528586, + "loss": 1.4785, + "theoretical_loss": 3.6371904987897565, + "tokens_seen": 1125777408 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033277833500501504, + "loss": 1.3534, + "theoretical_loss": 3.637170942464687, + "tokens_seen": 1125842944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003327683049147442, + "loss": 1.4257, + "theoretical_loss": 3.6371513875966968, + "tokens_seen": 1125908480 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003327582748244734, + "loss": 1.4567, + "theoretical_loss": 3.6371318341855927, + "tokens_seen": 1125974016 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033274824473420264, + "loss": 1.4485, + "theoretical_loss": 3.6371122822311808, + "tokens_seen": 1126039552 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033273821464393176, + "loss": 1.5489, + "theoretical_loss": 3.6370927317332677, + "tokens_seen": 1126105088 + }, + { + "epoch": 0.34, + "learning_rate": 0.000332728184553661, + "loss": 1.3452, + "theoretical_loss": 3.6370731826916614, + "tokens_seen": 1126170624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003327181544633902, + "loss": 1.4287, + "theoretical_loss": 3.6370536351061675, + "tokens_seen": 1126236160 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033270812437311936, + "loss": 1.3797, + "theoretical_loss": 3.637034088976593, + "tokens_seen": 1126301696 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033269809428284854, + "loss": 1.3898, + "theoretical_loss": 3.6370145443027453, + "tokens_seen": 1126367232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003326880641925777, + "loss": 1.5408, + "theoretical_loss": 3.6369950010844305, + "tokens_seen": 1126432768 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003326780341023069, + "loss": 1.4891, + "theoretical_loss": 3.636975459321456, + "tokens_seen": 1126498304 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033266800401203614, + "loss": 1.5426, + "theoretical_loss": 3.636955919013629, + "tokens_seen": 1126563840 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033265797392176527, + "loss": 1.45, + "theoretical_loss": 3.636936380160756, + "tokens_seen": 1126629376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003326479438314945, + "loss": 1.5194, + "theoretical_loss": 3.6369168427626444, + "tokens_seen": 1126694912 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033263791374122363, + "loss": 1.4447, + "theoretical_loss": 3.636897306819101, + "tokens_seen": 1126760448 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033262788365095286, + "loss": 1.3028, + "theoretical_loss": 3.6368777723299326, + "tokens_seen": 1126825984 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033261785356068205, + "loss": 1.568, + "theoretical_loss": 3.6368582392949476, + "tokens_seen": 1126891520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003326078234704112, + "loss": 1.4923, + "theoretical_loss": 3.6368387077139523, + "tokens_seen": 1126957056 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003325977933801404, + "loss": 1.4495, + "theoretical_loss": 3.6368191775867533, + "tokens_seen": 1127022592 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033258776328986964, + "loss": 1.435, + "theoretical_loss": 3.636799648913159, + "tokens_seen": 1127088128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003325777331995988, + "loss": 1.4442, + "theoretical_loss": 3.6367801216929765, + "tokens_seen": 1127153664 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 413086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4969782829284668, + "objective/train/theoretical_loss": 3.6367605959260123, + "objective/train/tokens_used": 1147679200, + "theoretical_loss": 3.6367605959260123, + "tokens_seen": 1127219200 + }, + { + "epoch": 0.34, + "learning_rate": 0.000332567703109328, + "loss": 1.3639, + "theoretical_loss": 3.6367605959260123, + "tokens_seen": 1127219200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003325576730190572, + "loss": 1.5188, + "theoretical_loss": 3.636741071612075, + "tokens_seen": 1127284736 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033254764292878637, + "loss": 1.4034, + "theoretical_loss": 3.636721548750971, + "tokens_seen": 1127350272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003325376128385156, + "loss": 1.4936, + "theoretical_loss": 3.6367020273425084, + "tokens_seen": 1127415808 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033252758274824473, + "loss": 1.472, + "theoretical_loss": 3.636682507386494, + "tokens_seen": 1127481344 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033251755265797396, + "loss": 1.4127, + "theoretical_loss": 3.636662988882736, + "tokens_seen": 1127546880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003325075225677031, + "loss": 1.4309, + "theoretical_loss": 3.6366434718310416, + "tokens_seen": 1127612416 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033249749247743233, + "loss": 1.527, + "theoretical_loss": 3.636623956231219, + "tokens_seen": 1127677952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003324874623871615, + "loss": 1.4144, + "theoretical_loss": 3.6366044420830743, + "tokens_seen": 1127743488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003324774322968907, + "loss": 1.5071, + "theoretical_loss": 3.6365849293864168, + "tokens_seen": 1127809024 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033246740220661987, + "loss": 1.436, + "theoretical_loss": 3.6365654181410534, + "tokens_seen": 1127874560 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033245737211634905, + "loss": 1.4944, + "theoretical_loss": 3.6365459083467915, + "tokens_seen": 1127940096 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033244734202607823, + "loss": 1.4881, + "theoretical_loss": 3.63652640000344, + "tokens_seen": 1128005632 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033243731193580747, + "loss": 1.4225, + "theoretical_loss": 3.636506893110806, + "tokens_seen": 1128071168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003324272818455366, + "loss": 1.4848, + "theoretical_loss": 3.6364873876686974, + "tokens_seen": 1128136704 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033241725175526583, + "loss": 1.4697, + "theoretical_loss": 3.636467883676922, + "tokens_seen": 1128202240 + }, + { + "epoch": 0.34, + "learning_rate": 0.000332407221664995, + "loss": 1.4124, + "theoretical_loss": 3.636448381135288, + "tokens_seen": 1128267776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003323971915747242, + "loss": 1.4942, + "theoretical_loss": 3.636428880043603, + "tokens_seen": 1128333312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003323871614844534, + "loss": 1.4514, + "theoretical_loss": 3.636409380401675, + "tokens_seen": 1128398848 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033237713139418255, + "loss": 1.5336, + "theoretical_loss": 3.636389882209313, + "tokens_seen": 1128464384 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033236710130391174, + "loss": 1.4636, + "theoretical_loss": 3.6363703854663236, + "tokens_seen": 1128529920 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033235707121364097, + "loss": 1.5164, + "theoretical_loss": 3.6363508901725154, + "tokens_seen": 1128595456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003323470411233701, + "loss": 1.2894, + "theoretical_loss": 3.6363313963276975, + "tokens_seen": 1128660992 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033233701103309933, + "loss": 1.474, + "theoretical_loss": 3.636311903931677, + "tokens_seen": 1128726528 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033232698094282846, + "loss": 1.4169, + "theoretical_loss": 3.6362924129842624, + "tokens_seen": 1128792064 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003323169508525577, + "loss": 1.4658, + "theoretical_loss": 3.6362729234852615, + "tokens_seen": 1128857600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003323069207622869, + "loss": 1.5514, + "theoretical_loss": 3.636253435434484, + "tokens_seen": 1128923136 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033229689067201606, + "loss": 1.4143, + "theoretical_loss": 3.6362339488317366, + "tokens_seen": 1128988672 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033228686058174524, + "loss": 1.341, + "theoretical_loss": 3.6362144636768283, + "tokens_seen": 1129054208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003322768304914744, + "loss": 1.3629, + "theoretical_loss": 3.6361949799695683, + "tokens_seen": 1129119744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003322668004012036, + "loss": 1.426, + "theoretical_loss": 3.6361754977097633, + "tokens_seen": 1129185280 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033225677031093284, + "loss": 1.4076, + "theoretical_loss": 3.6361560168972233, + "tokens_seen": 1129250816 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033224674022066196, + "loss": 1.6671, + "theoretical_loss": 3.6361365375317565, + "tokens_seen": 1129316352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003322367101303912, + "loss": 1.4921, + "theoretical_loss": 3.6361170596131704, + "tokens_seen": 1129381888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003322266800401204, + "loss": 1.3754, + "theoretical_loss": 3.6360975831412747, + "tokens_seen": 1129447424 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033221664994984956, + "loss": 1.439, + "theoretical_loss": 3.6360781081158775, + "tokens_seen": 1129512960 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033220661985957874, + "loss": 1.4412, + "theoretical_loss": 3.6360586345367882, + "tokens_seen": 1129578496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003321965897693079, + "loss": 1.4321, + "theoretical_loss": 3.6360391624038146, + "tokens_seen": 1129644032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003321865596790371, + "loss": 1.4338, + "theoretical_loss": 3.6360196917167658, + "tokens_seen": 1129709568 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033217652958876634, + "loss": 1.5123, + "theoretical_loss": 3.6360002224754506, + "tokens_seen": 1129775104 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033216649949849547, + "loss": 1.3678, + "theoretical_loss": 3.6359807546796774, + "tokens_seen": 1129840640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003321564694082247, + "loss": 1.4612, + "theoretical_loss": 3.6359612883292556, + "tokens_seen": 1129906176 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033214643931795383, + "loss": 1.5317, + "theoretical_loss": 3.6359418234239937, + "tokens_seen": 1129971712 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033213640922768306, + "loss": 1.5055, + "theoretical_loss": 3.6359223599637005, + "tokens_seen": 1130037248 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033212637913741225, + "loss": 1.5155, + "theoretical_loss": 3.6359028979481853, + "tokens_seen": 1130102784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003321163490471414, + "loss": 1.3672, + "theoretical_loss": 3.635883437377257, + "tokens_seen": 1130168320 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003321063189568706, + "loss": 1.3318, + "theoretical_loss": 3.6358639782507245, + "tokens_seen": 1130233856 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033209628886659984, + "loss": 1.5724, + "theoretical_loss": 3.635844520568397, + "tokens_seen": 1130299392 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033208625877632897, + "loss": 1.549, + "theoretical_loss": 3.6358250643300836, + "tokens_seen": 1130364928 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003320762286860582, + "loss": 1.4176, + "theoretical_loss": 3.635805609535593, + "tokens_seen": 1130430464 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 414307, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5938630104064941, + "objective/train/theoretical_loss": 3.635786156184735, + "objective/train/tokens_used": 1150956000, + "theoretical_loss": 3.635786156184735, + "tokens_seen": 1130496000 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033206619859578733, + "loss": 1.4876, + "theoretical_loss": 3.635786156184735, + "tokens_seen": 1130496000 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033205616850551657, + "loss": 1.3289, + "theoretical_loss": 3.635766704277318, + "tokens_seen": 1130561536 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033204613841524575, + "loss": 1.3468, + "theoretical_loss": 3.635747253813152, + "tokens_seen": 1130627072 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033203610832497493, + "loss": 1.3688, + "theoretical_loss": 3.635727804792046, + "tokens_seen": 1130692608 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003320260782347041, + "loss": 1.368, + "theoretical_loss": 3.635708357213809, + "tokens_seen": 1130758144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003320160481444333, + "loss": 1.3741, + "theoretical_loss": 3.635688911078251, + "tokens_seen": 1130823680 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003320060180541625, + "loss": 1.5103, + "theoretical_loss": 3.635669466385181, + "tokens_seen": 1130889216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003319959879638917, + "loss": 1.4538, + "theoretical_loss": 3.635650023134408, + "tokens_seen": 1130954752 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033198595787362084, + "loss": 1.5138, + "theoretical_loss": 3.635630581325742, + "tokens_seen": 1131020288 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033197592778335007, + "loss": 1.4779, + "theoretical_loss": 3.635611140958993, + "tokens_seen": 1131085824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003319658976930792, + "loss": 1.3308, + "theoretical_loss": 3.6355917020339685, + "tokens_seen": 1131151360 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033195586760280843, + "loss": 1.3638, + "theoretical_loss": 3.635572264550481, + "tokens_seen": 1131216896 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003319458375125376, + "loss": 1.4695, + "theoretical_loss": 3.635552828508337, + "tokens_seen": 1131282432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003319358074222668, + "loss": 1.4185, + "theoretical_loss": 3.635533393907349, + "tokens_seen": 1131347968 + }, + { + "epoch": 0.34, + "learning_rate": 0.000331925777331996, + "loss": 1.4474, + "theoretical_loss": 3.635513960747325, + "tokens_seen": 1131413504 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003319157472417252, + "loss": 1.4852, + "theoretical_loss": 3.6354945290280742, + "tokens_seen": 1131479040 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033190571715145434, + "loss": 1.4572, + "theoretical_loss": 3.6354750987494078, + "tokens_seen": 1131544576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318956870611836, + "loss": 1.4501, + "theoretical_loss": 3.635455669911135, + "tokens_seen": 1131610112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318856569709127, + "loss": 1.3523, + "theoretical_loss": 3.635436242513065, + "tokens_seen": 1131675648 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033187562688064194, + "loss": 1.5194, + "theoretical_loss": 3.6354168165550087, + "tokens_seen": 1131741184 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318655967903711, + "loss": 1.4633, + "theoretical_loss": 3.6353973920367757, + "tokens_seen": 1131806720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318555667001003, + "loss": 1.449, + "theoretical_loss": 3.635377968958175, + "tokens_seen": 1131872256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318455366098295, + "loss": 1.4845, + "theoretical_loss": 3.6353585473190178, + "tokens_seen": 1131937792 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033183550651955866, + "loss": 1.4126, + "theoretical_loss": 3.6353391271191136, + "tokens_seen": 1132003328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318254764292879, + "loss": 1.4765, + "theoretical_loss": 3.6353197083582725, + "tokens_seen": 1132068864 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003318154463390171, + "loss": 1.4083, + "theoretical_loss": 3.6353002910363035, + "tokens_seen": 1132134400 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033180541624874626, + "loss": 1.4544, + "theoretical_loss": 3.6352808751530183, + "tokens_seen": 1132199936 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033179538615847544, + "loss": 1.4489, + "theoretical_loss": 3.6352614607082265, + "tokens_seen": 1132265472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003317853560682046, + "loss": 1.4812, + "theoretical_loss": 3.6352420477017384, + "tokens_seen": 1132331008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003317753259779338, + "loss": 1.421, + "theoretical_loss": 3.6352226361333635, + "tokens_seen": 1132396544 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033176529588766304, + "loss": 1.4198, + "theoretical_loss": 3.635203226002912, + "tokens_seen": 1132462080 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033175526579739216, + "loss": 1.419, + "theoretical_loss": 3.6351838173101956, + "tokens_seen": 1132527616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003317452357071214, + "loss": 1.5406, + "theoretical_loss": 3.6351644100550233, + "tokens_seen": 1132593152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003317352056168506, + "loss": 1.3957, + "theoretical_loss": 3.635145004237206, + "tokens_seen": 1132658688 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033172517552657976, + "loss": 1.4079, + "theoretical_loss": 3.635125599856554, + "tokens_seen": 1132724224 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033171514543630894, + "loss": 1.4192, + "theoretical_loss": 3.6351061969128775, + "tokens_seen": 1132789760 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003317051153460381, + "loss": 1.5195, + "theoretical_loss": 3.6350867954059876, + "tokens_seen": 1132855296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003316950852557673, + "loss": 1.5616, + "theoretical_loss": 3.635067395335694, + "tokens_seen": 1132920832 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033168505516549654, + "loss": 1.3471, + "theoretical_loss": 3.6350479967018074, + "tokens_seen": 1132986368 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033167502507522567, + "loss": 1.5436, + "theoretical_loss": 3.6350285995041385, + "tokens_seen": 1133051904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003316649949849549, + "loss": 1.4547, + "theoretical_loss": 3.635009203742498, + "tokens_seen": 1133117440 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033165496489468403, + "loss": 1.4195, + "theoretical_loss": 3.6349898094166964, + "tokens_seen": 1133182976 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033164493480441326, + "loss": 1.4096, + "theoretical_loss": 3.634970416526545, + "tokens_seen": 1133248512 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033163490471414245, + "loss": 1.573, + "theoretical_loss": 3.6349510250718535, + "tokens_seen": 1133314048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003316248746238716, + "loss": 1.5171, + "theoretical_loss": 3.634931635052433, + "tokens_seen": 1133379584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003316148445336008, + "loss": 1.4936, + "theoretical_loss": 3.6349122464680947, + "tokens_seen": 1133445120 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033160481444333004, + "loss": 1.418, + "theoretical_loss": 3.6348928593186494, + "tokens_seen": 1133510656 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033159478435305917, + "loss": 1.4077, + "theoretical_loss": 3.634873473603907, + "tokens_seen": 1133576192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003315847542627884, + "loss": 1.3489, + "theoretical_loss": 3.63485408932368, + "tokens_seen": 1133641728 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033157472417251753, + "loss": 1.5111, + "theoretical_loss": 3.6348347064777773, + "tokens_seen": 1133707264 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 415442, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.445555329322815, + "objective/train/theoretical_loss": 3.6348153250660116, + "objective/train/tokens_used": 1154232800, + "theoretical_loss": 3.6348153250660116, + "tokens_seen": 1133772800 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033156469408224677, + "loss": 1.4636, + "theoretical_loss": 3.6348153250660116, + "tokens_seen": 1133772800 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033155466399197595, + "loss": 1.4946, + "theoretical_loss": 3.634795945088193, + "tokens_seen": 1133838336 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033154463390170513, + "loss": 1.3912, + "theoretical_loss": 3.634776566544133, + "tokens_seen": 1133903872 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003315346038114343, + "loss": 1.5021, + "theoretical_loss": 3.6347571894336426, + "tokens_seen": 1133969408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003315245737211635, + "loss": 1.4179, + "theoretical_loss": 3.6347378137565323, + "tokens_seen": 1134034944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003315145436308927, + "loss": 1.4529, + "theoretical_loss": 3.634718439512614, + "tokens_seen": 1134100480 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003315045135406219, + "loss": 1.4659, + "theoretical_loss": 3.6346990667016987, + "tokens_seen": 1134166016 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033149448345035104, + "loss": 1.3972, + "theoretical_loss": 3.634679695323597, + "tokens_seen": 1134231552 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033148445336008027, + "loss": 1.4344, + "theoretical_loss": 3.634660325378121, + "tokens_seen": 1134297088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003314744232698094, + "loss": 1.502, + "theoretical_loss": 3.634640956865082, + "tokens_seen": 1134362624 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033146439317953863, + "loss": 1.4026, + "theoretical_loss": 3.634621589784291, + "tokens_seen": 1134428160 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003314543630892678, + "loss": 1.4608, + "theoretical_loss": 3.6346022241355587, + "tokens_seen": 1134493696 + }, + { + "epoch": 0.34, + "learning_rate": 0.000331444332998997, + "loss": 1.4349, + "theoretical_loss": 3.634582859918697, + "tokens_seen": 1134559232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003314343029087262, + "loss": 1.5224, + "theoretical_loss": 3.6345634971335183, + "tokens_seen": 1134624768 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003314242728184554, + "loss": 1.4419, + "theoretical_loss": 3.6345441357798327, + "tokens_seen": 1134690304 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033141424272818454, + "loss": 1.4302, + "theoretical_loss": 3.634524775857452, + "tokens_seen": 1134755840 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003314042126379138, + "loss": 1.479, + "theoretical_loss": 3.6345054173661886, + "tokens_seen": 1134821376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313941825476429, + "loss": 1.5445, + "theoretical_loss": 3.6344860603058526, + "tokens_seen": 1134886912 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033138415245737214, + "loss": 1.4277, + "theoretical_loss": 3.6344667046762567, + "tokens_seen": 1134952448 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313741223671013, + "loss": 1.5231, + "theoretical_loss": 3.634447350477213, + "tokens_seen": 1135017984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313640922768305, + "loss": 1.4033, + "theoretical_loss": 3.6344279977085314, + "tokens_seen": 1135083520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313540621865597, + "loss": 1.5419, + "theoretical_loss": 3.6344086463700256, + "tokens_seen": 1135149056 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033134403209628886, + "loss": 1.4696, + "theoretical_loss": 3.6343892964615057, + "tokens_seen": 1135214592 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033133400200601804, + "loss": 1.4601, + "theoretical_loss": 3.6343699479827842, + "tokens_seen": 1135280128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313239719157473, + "loss": 1.384, + "theoretical_loss": 3.6343506009336735, + "tokens_seen": 1135345664 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003313139418254764, + "loss": 1.5521, + "theoretical_loss": 3.6343312553139846, + "tokens_seen": 1135411200 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033130391173520564, + "loss": 1.4064, + "theoretical_loss": 3.63431191112353, + "tokens_seen": 1135476736 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033129388164493477, + "loss": 1.4605, + "theoretical_loss": 3.6342925683621203, + "tokens_seen": 1135542272 + }, + { + "epoch": 0.34, + "learning_rate": 0.000331283851554664, + "loss": 1.5233, + "theoretical_loss": 3.6342732270295692, + "tokens_seen": 1135607808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003312738214643932, + "loss": 1.5028, + "theoretical_loss": 3.6342538871256878, + "tokens_seen": 1135673344 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033126379137412236, + "loss": 1.3559, + "theoretical_loss": 3.634234548650288, + "tokens_seen": 1135738880 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033125376128385155, + "loss": 1.4055, + "theoretical_loss": 3.6342152116031827, + "tokens_seen": 1135804416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003312437311935808, + "loss": 1.4006, + "theoretical_loss": 3.634195875984183, + "tokens_seen": 1135869952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003312337011033099, + "loss": 1.424, + "theoretical_loss": 3.634176541793102, + "tokens_seen": 1135935488 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033122367101303914, + "loss": 1.4247, + "theoretical_loss": 3.6341572090297514, + "tokens_seen": 1136001024 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033121364092276827, + "loss": 1.3974, + "theoretical_loss": 3.6341378776939433, + "tokens_seen": 1136066560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003312036108324975, + "loss": 1.4905, + "theoretical_loss": 3.63411854778549, + "tokens_seen": 1136132096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003311935807422267, + "loss": 1.5052, + "theoretical_loss": 3.634099219304203, + "tokens_seen": 1136197632 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033118355065195587, + "loss": 1.4979, + "theoretical_loss": 3.6340798922498965, + "tokens_seen": 1136263168 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033117352056168505, + "loss": 1.6022, + "theoretical_loss": 3.6340605666223817, + "tokens_seen": 1136328704 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033116349047141423, + "loss": 1.3931, + "theoretical_loss": 3.6340412424214708, + "tokens_seen": 1136394240 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003311534603811434, + "loss": 1.4195, + "theoretical_loss": 3.634021919646977, + "tokens_seen": 1136459776 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033114343029087265, + "loss": 1.5676, + "theoretical_loss": 3.6340025982987116, + "tokens_seen": 1136525312 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033113340020060177, + "loss": 1.476, + "theoretical_loss": 3.633983278376488, + "tokens_seen": 1136590848 + }, + { + "epoch": 0.34, + "learning_rate": 0.000331123370110331, + "loss": 1.5338, + "theoretical_loss": 3.6339639598801186, + "tokens_seen": 1136656384 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033111334002006013, + "loss": 1.4056, + "theoretical_loss": 3.633944642809416, + "tokens_seen": 1136721920 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033110330992978937, + "loss": 1.4813, + "theoretical_loss": 3.6339253271641923, + "tokens_seen": 1136787456 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033109327983951855, + "loss": 1.4006, + "theoretical_loss": 3.633906012944261, + "tokens_seen": 1136852992 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033108324974924773, + "loss": 1.4572, + "theoretical_loss": 3.633886700149434, + "tokens_seen": 1136918528 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033107321965897697, + "loss": 1.3558, + "theoretical_loss": 3.6338673887795245, + "tokens_seen": 1136984064 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 416597, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.515251874923706, + "objective/train/theoretical_loss": 3.633848078834345, + "objective/train/tokens_used": 1157509600, + "theoretical_loss": 3.633848078834345, + "tokens_seen": 1137049600 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033106318956870615, + "loss": 1.4657, + "theoretical_loss": 3.633848078834345, + "tokens_seen": 1137049600 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033105315947843533, + "loss": 1.4291, + "theoretical_loss": 3.633828770313708, + "tokens_seen": 1137115136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003310431293881645, + "loss": 1.4632, + "theoretical_loss": 3.6338094632174274, + "tokens_seen": 1137180672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003310330992978937, + "loss": 1.4917, + "theoretical_loss": 3.633790157545315, + "tokens_seen": 1137246208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003310230692076229, + "loss": 1.4157, + "theoretical_loss": 3.6337708532971833, + "tokens_seen": 1137311744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003310130391173521, + "loss": 1.4324, + "theoretical_loss": 3.6337515504728466, + "tokens_seen": 1137377280 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033100300902708124, + "loss": 1.3009, + "theoretical_loss": 3.633732249072117, + "tokens_seen": 1137442816 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033099297893681047, + "loss": 1.4007, + "theoretical_loss": 3.6337129490948072, + "tokens_seen": 1137508352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003309829488465396, + "loss": 1.4596, + "theoretical_loss": 3.6336936505407316, + "tokens_seen": 1137573888 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033097291875626883, + "loss": 1.4903, + "theoretical_loss": 3.633674353409702, + "tokens_seen": 1137639424 + }, + { + "epoch": 0.34, + "learning_rate": 0.000330962888665998, + "loss": 1.4771, + "theoretical_loss": 3.6336550577015316, + "tokens_seen": 1137704960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003309528585757272, + "loss": 1.4365, + "theoretical_loss": 3.633635763416034, + "tokens_seen": 1137770496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003309428284854564, + "loss": 1.4694, + "theoretical_loss": 3.633616470553022, + "tokens_seen": 1137836032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003309327983951856, + "loss": 1.5376, + "theoretical_loss": 3.633597179112309, + "tokens_seen": 1137901568 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033092276830491474, + "loss": 1.6363, + "theoretical_loss": 3.6335778890937083, + "tokens_seen": 1137967104 + }, + { + "epoch": 0.34, + "learning_rate": 0.000330912738214644, + "loss": 1.4996, + "theoretical_loss": 3.633558600497033, + "tokens_seen": 1138032640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003309027081243731, + "loss": 1.4964, + "theoretical_loss": 3.633539313322096, + "tokens_seen": 1138098176 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033089267803410234, + "loss": 1.4452, + "theoretical_loss": 3.6335200275687116, + "tokens_seen": 1138163712 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003308826479438315, + "loss": 1.4955, + "theoretical_loss": 3.633500743236693, + "tokens_seen": 1138229248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003308726178535607, + "loss": 1.3153, + "theoretical_loss": 3.6334814603258527, + "tokens_seen": 1138294784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0003308625877632899, + "loss": 1.5141, + "theoretical_loss": 3.6334621788360053, + "tokens_seen": 1138360320 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033085255767301906, + "loss": 1.4821, + "theoretical_loss": 3.6334428987669636, + "tokens_seen": 1138425856 + }, + { + "epoch": 0.34, + "learning_rate": 0.00033084252758274824, + "loss": 1.4435, + "theoretical_loss": 3.633423620118541, + "tokens_seen": 1138491392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003308324974924775, + "loss": 1.4304, + "theoretical_loss": 3.633404342890551, + "tokens_seen": 1138556928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003308224674022066, + "loss": 1.4879, + "theoretical_loss": 3.6333850670828083, + "tokens_seen": 1138622464 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033081243731193584, + "loss": 1.4217, + "theoretical_loss": 3.633365792695126, + "tokens_seen": 1138688000 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033080240722166497, + "loss": 1.3988, + "theoretical_loss": 3.6333465197273167, + "tokens_seen": 1138753536 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003307923771313942, + "loss": 1.4267, + "theoretical_loss": 3.633327248179195, + "tokens_seen": 1138819072 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003307823470411234, + "loss": 1.4883, + "theoretical_loss": 3.633307978050575, + "tokens_seen": 1138884608 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033077231695085256, + "loss": 1.4424, + "theoretical_loss": 3.6332887093412696, + "tokens_seen": 1138950144 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033076228686058175, + "loss": 1.5017, + "theoretical_loss": 3.6332694420510934, + "tokens_seen": 1139015680 + }, + { + "epoch": 0.35, + "learning_rate": 0.000330752256770311, + "loss": 1.407, + "theoretical_loss": 3.6332501761798595, + "tokens_seen": 1139081216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003307422266800401, + "loss": 1.391, + "theoretical_loss": 3.633230911727382, + "tokens_seen": 1139146752 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033073219658976934, + "loss": 1.5293, + "theoretical_loss": 3.633211648693475, + "tokens_seen": 1139212288 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033072216649949847, + "loss": 1.5906, + "theoretical_loss": 3.633192387077953, + "tokens_seen": 1139277824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003307121364092277, + "loss": 1.4714, + "theoretical_loss": 3.633173126880629, + "tokens_seen": 1139343360 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003307021063189569, + "loss": 1.471, + "theoretical_loss": 3.633153868101317, + "tokens_seen": 1139408896 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033069207622868607, + "loss": 1.4256, + "theoretical_loss": 3.633134610739832, + "tokens_seen": 1139474432 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033068204613841525, + "loss": 1.4817, + "theoretical_loss": 3.6331153547959874, + "tokens_seen": 1139539968 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033067201604814443, + "loss": 1.3864, + "theoretical_loss": 3.633096100269597, + "tokens_seen": 1139605504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003306619859578736, + "loss": 1.5471, + "theoretical_loss": 3.633076847160476, + "tokens_seen": 1139671040 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033065195586760285, + "loss": 1.454, + "theoretical_loss": 3.633057595468437, + "tokens_seen": 1139736576 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033064192577733197, + "loss": 1.365, + "theoretical_loss": 3.633038345193296, + "tokens_seen": 1139802112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003306318956870612, + "loss": 1.4255, + "theoretical_loss": 3.6330190963348663, + "tokens_seen": 1139867648 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033062186559679034, + "loss": 1.4722, + "theoretical_loss": 3.632999848892962, + "tokens_seen": 1139933184 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033061183550651957, + "loss": 1.4181, + "theoretical_loss": 3.6329806028673985, + "tokens_seen": 1139998720 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033060180541624875, + "loss": 1.4915, + "theoretical_loss": 3.6329613582579885, + "tokens_seen": 1140064256 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033059177532597793, + "loss": 1.4259, + "theoretical_loss": 3.632942115064548, + "tokens_seen": 1140129792 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003305817452357071, + "loss": 1.4694, + "theoretical_loss": 3.6329228732868906, + "tokens_seen": 1140195328 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033057171514543635, + "loss": 1.4321, + "theoretical_loss": 3.6329036329248305, + "tokens_seen": 1140260864 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 417664, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.422420620918274, + "objective/train/theoretical_loss": 3.632884393978183, + "objective/train/tokens_used": 1160786400, + "theoretical_loss": 3.632884393978183, + "tokens_seen": 1140326400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003305616850551655, + "loss": 1.4629, + "theoretical_loss": 3.632884393978183, + "tokens_seen": 1140326400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003305516549648947, + "loss": 1.4946, + "theoretical_loss": 3.632865156446763, + "tokens_seen": 1140391936 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033054162487462384, + "loss": 1.503, + "theoretical_loss": 3.6328459203303836, + "tokens_seen": 1140457472 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003305315947843531, + "loss": 1.4206, + "theoretical_loss": 3.63282668562886, + "tokens_seen": 1140523008 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033052156469408225, + "loss": 1.5054, + "theoretical_loss": 3.632807452342007, + "tokens_seen": 1140588544 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033051153460381144, + "loss": 1.4642, + "theoretical_loss": 3.63278822046964, + "tokens_seen": 1140654080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003305015045135406, + "loss": 1.5063, + "theoretical_loss": 3.6327689900115727, + "tokens_seen": 1140719616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304914744232698, + "loss": 1.4332, + "theoretical_loss": 3.63274976096762, + "tokens_seen": 1140785152 + }, + { + "epoch": 0.35, + "learning_rate": 0.000330481444332999, + "loss": 1.4618, + "theoretical_loss": 3.6327305333375968, + "tokens_seen": 1140850688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304714142427282, + "loss": 1.5395, + "theoretical_loss": 3.6327113071213177, + "tokens_seen": 1140916224 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033046138415245734, + "loss": 1.3918, + "theoretical_loss": 3.632692082318598, + "tokens_seen": 1140981760 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304513540621866, + "loss": 1.4371, + "theoretical_loss": 3.6326728589292525, + "tokens_seen": 1141047296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304413239719157, + "loss": 1.4878, + "theoretical_loss": 3.632653636953096, + "tokens_seen": 1141112832 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033043129388164494, + "loss": 1.4392, + "theoretical_loss": 3.6326344163899433, + "tokens_seen": 1141178368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304212637913741, + "loss": 1.4056, + "theoretical_loss": 3.63261519723961, + "tokens_seen": 1141243904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304112337011033, + "loss": 1.3442, + "theoretical_loss": 3.63259597950191, + "tokens_seen": 1141309440 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003304012036108325, + "loss": 1.5316, + "theoretical_loss": 3.63257676317666, + "tokens_seen": 1141374976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303911735205617, + "loss": 1.4985, + "theoretical_loss": 3.6325575482636734, + "tokens_seen": 1141440512 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033038114343029084, + "loss": 1.4573, + "theoretical_loss": 3.632538334762766, + "tokens_seen": 1141506048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303711133400201, + "loss": 1.4783, + "theoretical_loss": 3.6325191226737537, + "tokens_seen": 1141571584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303610832497492, + "loss": 1.5209, + "theoretical_loss": 3.6324999119964505, + "tokens_seen": 1141637120 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033035105315947844, + "loss": 1.4168, + "theoretical_loss": 3.632480702730673, + "tokens_seen": 1141702656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303410230692076, + "loss": 1.4911, + "theoretical_loss": 3.6324614948762344, + "tokens_seen": 1141768192 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303309929789368, + "loss": 1.3818, + "theoretical_loss": 3.6324422884329524, + "tokens_seen": 1141833728 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033032096288866604, + "loss": 1.4835, + "theoretical_loss": 3.6324230834006404, + "tokens_seen": 1141899264 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033031093279839517, + "loss": 1.4406, + "theoretical_loss": 3.632403879779115, + "tokens_seen": 1141964800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003303009027081244, + "loss": 1.492, + "theoretical_loss": 3.632384677568191, + "tokens_seen": 1142030336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003302908726178536, + "loss": 1.4964, + "theoretical_loss": 3.632365476767684, + "tokens_seen": 1142095872 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033028084252758276, + "loss": 1.4763, + "theoretical_loss": 3.63234627737741, + "tokens_seen": 1142161408 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033027081243731195, + "loss": 1.2988, + "theoretical_loss": 3.6323270793971836, + "tokens_seen": 1142226944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003302607823470412, + "loss": 1.4048, + "theoretical_loss": 3.632307882826821, + "tokens_seen": 1142292480 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003302507522567703, + "loss": 1.4563, + "theoretical_loss": 3.6322886876661373, + "tokens_seen": 1142358016 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033024072216649954, + "loss": 1.4745, + "theoretical_loss": 3.632269493914949, + "tokens_seen": 1142423552 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033023069207622867, + "loss": 1.5479, + "theoretical_loss": 3.6322503015730705, + "tokens_seen": 1142489088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003302206619859579, + "loss": 1.3767, + "theoretical_loss": 3.6322311106403182, + "tokens_seen": 1142554624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003302106318956871, + "loss": 1.4, + "theoretical_loss": 3.632211921116508, + "tokens_seen": 1142620160 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033020060180541627, + "loss": 1.4839, + "theoretical_loss": 3.632192733001455, + "tokens_seen": 1142685696 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033019057171514545, + "loss": 1.4055, + "theoretical_loss": 3.632173546294976, + "tokens_seen": 1142751232 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033018054162487463, + "loss": 1.3362, + "theoretical_loss": 3.6321543609968856, + "tokens_seen": 1142816768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003301705115346038, + "loss": 1.4383, + "theoretical_loss": 3.6321351771070005, + "tokens_seen": 1142882304 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033016048144433305, + "loss": 1.4343, + "theoretical_loss": 3.6321159946251367, + "tokens_seen": 1142947840 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003301504513540622, + "loss": 1.4539, + "theoretical_loss": 3.6320968135511094, + "tokens_seen": 1143013376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003301404212637914, + "loss": 1.3496, + "theoretical_loss": 3.632077633884735, + "tokens_seen": 1143078912 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033013039117352054, + "loss": 1.4812, + "theoretical_loss": 3.6320584556258297, + "tokens_seen": 1143144448 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033012036108324977, + "loss": 1.3241, + "theoretical_loss": 3.6320392787742093, + "tokens_seen": 1143209984 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033011033099297895, + "loss": 1.4167, + "theoretical_loss": 3.6320201033296895, + "tokens_seen": 1143275520 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033010030090270813, + "loss": 1.4478, + "theoretical_loss": 3.6320009292920874, + "tokens_seen": 1143341056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003300902708124373, + "loss": 1.4663, + "theoretical_loss": 3.6319817566612183, + "tokens_seen": 1143406592 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033008024072216655, + "loss": 1.362, + "theoretical_loss": 3.6319625854368986, + "tokens_seen": 1143472128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003300702106318957, + "loss": 1.4839, + "theoretical_loss": 3.6319434156189443, + "tokens_seen": 1143537664 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 418760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3668909072875977, + "objective/train/theoretical_loss": 3.6319242472071718, + "objective/train/tokens_used": 1164063200, + "theoretical_loss": 3.6319242472071718, + "tokens_seen": 1143603200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003300601805416249, + "loss": 1.4044, + "theoretical_loss": 3.6319242472071718, + "tokens_seen": 1143603200 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033005015045135404, + "loss": 1.5289, + "theoretical_loss": 3.631905080201398, + "tokens_seen": 1143668736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003300401203610833, + "loss": 1.3727, + "theoretical_loss": 3.6318859146014377, + "tokens_seen": 1143734272 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033003009027081245, + "loss": 1.381, + "theoretical_loss": 3.6318667504071094, + "tokens_seen": 1143799808 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033002006018054164, + "loss": 1.4041, + "theoretical_loss": 3.6318475876182275, + "tokens_seen": 1143865344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003300100300902708, + "loss": 1.4674, + "theoretical_loss": 3.631828426234609, + "tokens_seen": 1143930880 + }, + { + "epoch": 0.35, + "learning_rate": 0.00033, + "loss": 1.3682, + "theoretical_loss": 3.631809266256071, + "tokens_seen": 1143996416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299899699097292, + "loss": 1.381, + "theoretical_loss": 3.631790107682429, + "tokens_seen": 1144061952 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299799398194584, + "loss": 1.2986, + "theoretical_loss": 3.6317709505135003, + "tokens_seen": 1144127488 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032996990972918754, + "loss": 1.3103, + "theoretical_loss": 3.631751794749101, + "tokens_seen": 1144193024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299598796389168, + "loss": 1.4855, + "theoretical_loss": 3.6317326403890484, + "tokens_seen": 1144258560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299498495486459, + "loss": 1.3159, + "theoretical_loss": 3.6317134874331582, + "tokens_seen": 1144324096 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032993981945837514, + "loss": 1.2826, + "theoretical_loss": 3.6316943358812477, + "tokens_seen": 1144389632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299297893681043, + "loss": 1.4825, + "theoretical_loss": 3.631675185733133, + "tokens_seen": 1144455168 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299197592778335, + "loss": 1.4378, + "theoretical_loss": 3.6316560369886313, + "tokens_seen": 1144520704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003299097291875627, + "loss": 1.4757, + "theoretical_loss": 3.6316368896475586, + "tokens_seen": 1144586240 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003298996990972919, + "loss": 1.461, + "theoretical_loss": 3.6316177437097332, + "tokens_seen": 1144651776 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032988966900702104, + "loss": 1.3015, + "theoretical_loss": 3.6315985991749704, + "tokens_seen": 1144717312 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003298796389167503, + "loss": 1.3439, + "theoretical_loss": 3.631579456043088, + "tokens_seen": 1144782848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003298696088264794, + "loss": 1.3066, + "theoretical_loss": 3.631560314313903, + "tokens_seen": 1144848384 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032985957873620864, + "loss": 1.4156, + "theoretical_loss": 3.631541173987231, + "tokens_seen": 1144913920 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003298495486459378, + "loss": 1.4297, + "theoretical_loss": 3.6315220350628907, + "tokens_seen": 1144979456 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329839518555667, + "loss": 1.347, + "theoretical_loss": 3.631502897540698, + "tokens_seen": 1145044992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003298294884653962, + "loss": 1.4379, + "theoretical_loss": 3.6314837614204705, + "tokens_seen": 1145110528 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032981945837512537, + "loss": 1.4056, + "theoretical_loss": 3.6314646267020247, + "tokens_seen": 1145176064 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032980942828485455, + "loss": 1.5069, + "theoretical_loss": 3.6314454933851783, + "tokens_seen": 1145241600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297993981945838, + "loss": 1.4152, + "theoretical_loss": 3.6314263614697477, + "tokens_seen": 1145307136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297893681043129, + "loss": 1.4577, + "theoretical_loss": 3.6314072309555505, + "tokens_seen": 1145372672 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032977933801404215, + "loss": 1.5353, + "theoretical_loss": 3.631388101842404, + "tokens_seen": 1145438208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297693079237713, + "loss": 1.3708, + "theoretical_loss": 3.6313689741301256, + "tokens_seen": 1145503744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297592778335005, + "loss": 1.4718, + "theoretical_loss": 3.631349847818532, + "tokens_seen": 1145569280 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297492477432297, + "loss": 1.4186, + "theoretical_loss": 3.6313307229074407, + "tokens_seen": 1145634816 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032973921765295887, + "loss": 1.3098, + "theoretical_loss": 3.6313115993966694, + "tokens_seen": 1145700352 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032972918756268805, + "loss": 1.4845, + "theoretical_loss": 3.6312924772860358, + "tokens_seen": 1145765888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297191574724173, + "loss": 1.5262, + "theoretical_loss": 3.631273356575356, + "tokens_seen": 1145831424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003297091273821464, + "loss": 1.439, + "theoretical_loss": 3.6312542372644483, + "tokens_seen": 1145896960 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032969909729187565, + "loss": 1.407, + "theoretical_loss": 3.6312351193531303, + "tokens_seen": 1145962496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003296890672016048, + "loss": 1.4704, + "theoretical_loss": 3.6312160028412195, + "tokens_seen": 1146028032 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329679037111334, + "loss": 1.4509, + "theoretical_loss": 3.631196887728533, + "tokens_seen": 1146093568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003296690070210632, + "loss": 1.4227, + "theoretical_loss": 3.6311777740148887, + "tokens_seen": 1146159104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003296589769307924, + "loss": 1.4835, + "theoretical_loss": 3.631158661700104, + "tokens_seen": 1146224640 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032964894684052155, + "loss": 1.5523, + "theoretical_loss": 3.631139550783997, + "tokens_seen": 1146290176 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032963891675025074, + "loss": 1.3669, + "theoretical_loss": 3.631120441266385, + "tokens_seen": 1146355712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003296288866599799, + "loss": 1.4184, + "theoretical_loss": 3.6311013331470856, + "tokens_seen": 1146421248 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032961885656970915, + "loss": 1.4031, + "theoretical_loss": 3.631082226425917, + "tokens_seen": 1146486784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003296088264794383, + "loss": 1.3807, + "theoretical_loss": 3.6310631211026965, + "tokens_seen": 1146552320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295987963891675, + "loss": 1.5127, + "theoretical_loss": 3.6310440171772425, + "tokens_seen": 1146617856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295887662988967, + "loss": 1.3977, + "theoretical_loss": 3.6310249146493723, + "tokens_seen": 1146683392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295787362086259, + "loss": 1.4371, + "theoretical_loss": 3.631005813518904, + "tokens_seen": 1146748928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295687061183551, + "loss": 1.4047, + "theoretical_loss": 3.6309867137856555, + "tokens_seen": 1146814464 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 419954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4015012979507446, + "objective/train/theoretical_loss": 3.630967615449445, + "objective/train/tokens_used": 1167340000, + "theoretical_loss": 3.630967615449445, + "tokens_seen": 1146880000 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032955867602808424, + "loss": 1.3224, + "theoretical_loss": 3.630967615449445, + "tokens_seen": 1146880000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295486459378135, + "loss": 1.3529, + "theoretical_loss": 3.63094851851009, + "tokens_seen": 1146945536 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032953861584754265, + "loss": 1.5142, + "theoretical_loss": 3.6309294229674087, + "tokens_seen": 1147011072 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032952858575727184, + "loss": 1.5324, + "theoretical_loss": 3.63091032882122, + "tokens_seen": 1147076608 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329518555667001, + "loss": 1.3994, + "theoretical_loss": 3.630891236071341, + "tokens_seen": 1147142144 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003295085255767302, + "loss": 1.4113, + "theoretical_loss": 3.63087214471759, + "tokens_seen": 1147207680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294984954864594, + "loss": 1.3345, + "theoretical_loss": 3.6308530547597853, + "tokens_seen": 1147273216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294884653961886, + "loss": 1.3883, + "theoretical_loss": 3.6308339661977445, + "tokens_seen": 1147338752 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032947843530591774, + "loss": 1.4323, + "theoretical_loss": 3.630814879031287, + "tokens_seen": 1147404288 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329468405215647, + "loss": 1.4325, + "theoretical_loss": 3.6307957932602304, + "tokens_seen": 1147469824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294583751253761, + "loss": 1.5168, + "theoretical_loss": 3.630776708884393, + "tokens_seen": 1147535360 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032944834503510534, + "loss": 1.4309, + "theoretical_loss": 3.630757625903593, + "tokens_seen": 1147600896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294383149448345, + "loss": 1.3621, + "theoretical_loss": 3.630738544317649, + "tokens_seen": 1147666432 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294282848545637, + "loss": 1.3678, + "theoretical_loss": 3.6307194641263796, + "tokens_seen": 1147731968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294182547642929, + "loss": 1.5602, + "theoretical_loss": 3.630700385329603, + "tokens_seen": 1147797504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003294082246740221, + "loss": 1.4308, + "theoretical_loss": 3.6306813079271376, + "tokens_seen": 1147863040 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032939819458375124, + "loss": 1.464, + "theoretical_loss": 3.630662231918802, + "tokens_seen": 1147928576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003293881644934805, + "loss": 1.5032, + "theoretical_loss": 3.630643157304415, + "tokens_seen": 1147994112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003293781344032096, + "loss": 1.5216, + "theoretical_loss": 3.6306240840837942, + "tokens_seen": 1148059648 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032936810431293884, + "loss": 1.4508, + "theoretical_loss": 3.6306050122567592, + "tokens_seen": 1148125184 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329358074222668, + "loss": 1.4175, + "theoretical_loss": 3.6305859418231283, + "tokens_seen": 1148190720 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003293480441323972, + "loss": 1.3604, + "theoretical_loss": 3.6305668727827203, + "tokens_seen": 1148256256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003293380140421264, + "loss": 1.4358, + "theoretical_loss": 3.630547805135354, + "tokens_seen": 1148321792 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032932798395185557, + "loss": 1.2939, + "theoretical_loss": 3.630528738880847, + "tokens_seen": 1148387328 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032931795386158475, + "loss": 1.4092, + "theoretical_loss": 3.63050967401902, + "tokens_seen": 1148452864 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329307923771314, + "loss": 1.3879, + "theoretical_loss": 3.630490610549691, + "tokens_seen": 1148518400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292978936810431, + "loss": 1.4061, + "theoretical_loss": 3.630471548472678, + "tokens_seen": 1148583936 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032928786359077235, + "loss": 1.5069, + "theoretical_loss": 3.6304524877878004, + "tokens_seen": 1148649472 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292778335005015, + "loss": 1.5159, + "theoretical_loss": 3.630433428494878, + "tokens_seen": 1148715008 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292678034102307, + "loss": 1.3935, + "theoretical_loss": 3.630414370593728, + "tokens_seen": 1148780544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292577733199599, + "loss": 1.4678, + "theoretical_loss": 3.6303953140841716, + "tokens_seen": 1148846080 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032924774322968907, + "loss": 1.45, + "theoretical_loss": 3.6303762589660256, + "tokens_seen": 1148911616 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032923771313941825, + "loss": 1.4513, + "theoretical_loss": 3.6303572052391107, + "tokens_seen": 1148977152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292276830491475, + "loss": 1.346, + "theoretical_loss": 3.630338152903245, + "tokens_seen": 1149042688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003292176529588766, + "loss": 1.4799, + "theoretical_loss": 3.630319101958248, + "tokens_seen": 1149108224 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032920762286860585, + "loss": 1.3483, + "theoretical_loss": 3.630300052403939, + "tokens_seen": 1149173760 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329197592778335, + "loss": 1.3986, + "theoretical_loss": 3.6302810042401368, + "tokens_seen": 1149239296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291875626880642, + "loss": 1.3579, + "theoretical_loss": 3.6302619574666606, + "tokens_seen": 1149304832 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291775325977934, + "loss": 1.3869, + "theoretical_loss": 3.6302429120833297, + "tokens_seen": 1149370368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291675025075226, + "loss": 1.3362, + "theoretical_loss": 3.630223868089964, + "tokens_seen": 1149435904 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032915747241725175, + "loss": 1.4512, + "theoretical_loss": 3.6302048254863823, + "tokens_seen": 1149501440 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032914744232698094, + "loss": 1.4161, + "theoretical_loss": 3.6301857842724035, + "tokens_seen": 1149566976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291374122367101, + "loss": 1.3214, + "theoretical_loss": 3.6301667444478474, + "tokens_seen": 1149632512 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032912738214643935, + "loss": 1.6417, + "theoretical_loss": 3.630147706012534, + "tokens_seen": 1149698048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291173520561685, + "loss": 1.4788, + "theoretical_loss": 3.6301286689662824, + "tokens_seen": 1149763584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003291073219658977, + "loss": 1.377, + "theoretical_loss": 3.630109633308911, + "tokens_seen": 1149829120 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003290972918756269, + "loss": 1.4202, + "theoretical_loss": 3.6300905990402415, + "tokens_seen": 1149894656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003290872617853561, + "loss": 1.4001, + "theoretical_loss": 3.6300715661600913, + "tokens_seen": 1149960192 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032907723169508526, + "loss": 1.3866, + "theoretical_loss": 3.630052534668281, + "tokens_seen": 1150025728 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032906720160481444, + "loss": 1.4353, + "theoretical_loss": 3.63003350456463, + "tokens_seen": 1150091264 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 421268, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4163382053375244, + "objective/train/theoretical_loss": 3.630014475848958, + "objective/train/tokens_used": 1170616800, + "theoretical_loss": 3.630014475848958, + "tokens_seen": 1150156800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003290571715145436, + "loss": 1.3911, + "theoretical_loss": 3.630014475848958, + "tokens_seen": 1150156800 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032904714142427286, + "loss": 1.4512, + "theoretical_loss": 3.6299954485210852, + "tokens_seen": 1150222336 + }, + { + "epoch": 0.35, + "learning_rate": 0.000329037111334002, + "loss": 1.3476, + "theoretical_loss": 3.629976422580831, + "tokens_seen": 1150287872 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003290270812437312, + "loss": 1.4192, + "theoretical_loss": 3.629957398028014, + "tokens_seen": 1150353408 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032901705115346034, + "loss": 1.3807, + "theoretical_loss": 3.629938374862456, + "tokens_seen": 1150418944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003290070210631896, + "loss": 1.3745, + "theoretical_loss": 3.6299193530839755, + "tokens_seen": 1150484480 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032899699097291876, + "loss": 1.3136, + "theoretical_loss": 3.629900332692392, + "tokens_seen": 1150550016 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032898696088264794, + "loss": 1.4627, + "theoretical_loss": 3.629881313687527, + "tokens_seen": 1150615552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003289769307923771, + "loss": 1.4088, + "theoretical_loss": 3.6298622960691995, + "tokens_seen": 1150681088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003289669007021063, + "loss": 1.4068, + "theoretical_loss": 3.629843279837229, + "tokens_seen": 1150746624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003289568706118355, + "loss": 1.4295, + "theoretical_loss": 3.629824264991436, + "tokens_seen": 1150812160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003289468405215647, + "loss": 1.3966, + "theoretical_loss": 3.6298052515316406, + "tokens_seen": 1150877696 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032893681043129385, + "loss": 1.503, + "theoretical_loss": 3.6297862394576628, + "tokens_seen": 1150943232 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003289267803410231, + "loss": 1.3821, + "theoretical_loss": 3.629767228769323, + "tokens_seen": 1151008768 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032891675025075226, + "loss": 1.3381, + "theoretical_loss": 3.62974821946644, + "tokens_seen": 1151074304 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032890672016048145, + "loss": 1.3813, + "theoretical_loss": 3.6297292115488355, + "tokens_seen": 1151139840 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288966900702106, + "loss": 1.2134, + "theoretical_loss": 3.6297102050163286, + "tokens_seen": 1151205376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288866599799398, + "loss": 1.5468, + "theoretical_loss": 3.6296911998687404, + "tokens_seen": 1151270912 + }, + { + "epoch": 0.35, + "learning_rate": 0.000328876629889669, + "loss": 1.4022, + "theoretical_loss": 3.629672196105891, + "tokens_seen": 1151336448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288665997993982, + "loss": 1.4435, + "theoretical_loss": 3.6296531937276004, + "tokens_seen": 1151401984 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032885656970912735, + "loss": 1.4785, + "theoretical_loss": 3.629634192733689, + "tokens_seen": 1151467520 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288465396188566, + "loss": 1.3179, + "theoretical_loss": 3.6296151931239766, + "tokens_seen": 1151533056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288365095285857, + "loss": 1.4977, + "theoretical_loss": 3.629596194898285, + "tokens_seen": 1151598592 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032882647943831495, + "loss": 1.4635, + "theoretical_loss": 3.6295771980564333, + "tokens_seen": 1151664128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288164493480442, + "loss": 1.3405, + "theoretical_loss": 3.6295582025982425, + "tokens_seen": 1151729664 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003288064192577733, + "loss": 1.4369, + "theoretical_loss": 3.629539208523533, + "tokens_seen": 1151795200 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032879638916750255, + "loss": 1.4076, + "theoretical_loss": 3.6295202158321254, + "tokens_seen": 1151860736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287863590772317, + "loss": 1.3861, + "theoretical_loss": 3.6295012245238403, + "tokens_seen": 1151926272 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287763289869609, + "loss": 1.344, + "theoretical_loss": 3.6294822345984983, + "tokens_seen": 1151991808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287662988966901, + "loss": 1.4048, + "theoretical_loss": 3.62946324605592, + "tokens_seen": 1152057344 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032875626880641927, + "loss": 1.289, + "theoretical_loss": 3.6294442588959255, + "tokens_seen": 1152122880 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032874623871614845, + "loss": 1.4023, + "theoretical_loss": 3.6294252731183363, + "tokens_seen": 1152188416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287362086258777, + "loss": 1.4859, + "theoretical_loss": 3.629406288722973, + "tokens_seen": 1152253952 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287261785356068, + "loss": 1.3559, + "theoretical_loss": 3.629387305709656, + "tokens_seen": 1152319488 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032871614844533605, + "loss": 1.4315, + "theoretical_loss": 3.6293683240782064, + "tokens_seen": 1152385024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003287061183550652, + "loss": 1.351, + "theoretical_loss": 3.629349343828445, + "tokens_seen": 1152450560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286960882647944, + "loss": 1.4135, + "theoretical_loss": 3.6293303649601922, + "tokens_seen": 1152516096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286860581745236, + "loss": 1.4266, + "theoretical_loss": 3.62931138747327, + "tokens_seen": 1152581632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286760280842528, + "loss": 1.3918, + "theoretical_loss": 3.6292924113674983, + "tokens_seen": 1152647168 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032866599799398195, + "loss": 1.3302, + "theoretical_loss": 3.629273436642698, + "tokens_seen": 1152712704 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032865596790371114, + "loss": 1.4561, + "theoretical_loss": 3.6292544632986905, + "tokens_seen": 1152778240 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286459378134403, + "loss": 1.4175, + "theoretical_loss": 3.6292354913352973, + "tokens_seen": 1152843776 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032863590772316955, + "loss": 1.5683, + "theoretical_loss": 3.6292165207523386, + "tokens_seen": 1152909312 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286258776328987, + "loss": 1.4997, + "theoretical_loss": 3.629197551549636, + "tokens_seen": 1152974848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286158475426279, + "loss": 1.3429, + "theoretical_loss": 3.6291785837270103, + "tokens_seen": 1153040384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003286058174523571, + "loss": 1.4613, + "theoretical_loss": 3.6291596172842833, + "tokens_seen": 1153105920 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003285957873620863, + "loss": 1.5513, + "theoretical_loss": 3.6291406522212757, + "tokens_seen": 1153171456 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032858575727181546, + "loss": 1.3953, + "theoretical_loss": 3.629121688537808, + "tokens_seen": 1153236992 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032857572718154464, + "loss": 1.2668, + "theoretical_loss": 3.629102726233703, + "tokens_seen": 1153302528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003285656970912738, + "loss": 1.4246, + "theoretical_loss": 3.6290837653087813, + "tokens_seen": 1153368064 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 422447, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.571091890335083, + "objective/train/theoretical_loss": 3.6290648057628636, + "objective/train/tokens_used": 1173893600, + "theoretical_loss": 3.6290648057628636, + "tokens_seen": 1153433600 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032855566700100306, + "loss": 1.4624, + "theoretical_loss": 3.6290648057628636, + "tokens_seen": 1153433600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003285456369107322, + "loss": 1.328, + "theoretical_loss": 3.6290458475957723, + "tokens_seen": 1153499136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003285356068204614, + "loss": 1.4386, + "theoretical_loss": 3.629026890807328, + "tokens_seen": 1153564672 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032852557673019054, + "loss": 1.3854, + "theoretical_loss": 3.6290079353973526, + "tokens_seen": 1153630208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003285155466399198, + "loss": 1.4129, + "theoretical_loss": 3.6289889813656675, + "tokens_seen": 1153695744 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032850551654964896, + "loss": 1.353, + "theoretical_loss": 3.628970028712094, + "tokens_seen": 1153761280 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032849548645937814, + "loss": 1.3213, + "theoretical_loss": 3.6289510774364535, + "tokens_seen": 1153826816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284854563691073, + "loss": 1.4292, + "theoretical_loss": 3.628932127538568, + "tokens_seen": 1153892352 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284754262788365, + "loss": 1.3653, + "theoretical_loss": 3.6289131790182587, + "tokens_seen": 1153957888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284653961885657, + "loss": 1.3699, + "theoretical_loss": 3.628894231875348, + "tokens_seen": 1154023424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284553660982949, + "loss": 1.4069, + "theoretical_loss": 3.6288752861096567, + "tokens_seen": 1154088960 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032844533600802405, + "loss": 1.4153, + "theoretical_loss": 3.628856341721007, + "tokens_seen": 1154154496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284353059177533, + "loss": 1.4462, + "theoretical_loss": 3.62883739870922, + "tokens_seen": 1154220032 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032842527582748246, + "loss": 1.3154, + "theoretical_loss": 3.628818457074118, + "tokens_seen": 1154285568 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032841524573721165, + "loss": 1.3986, + "theoretical_loss": 3.6287995168155227, + "tokens_seen": 1154351104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003284052156469408, + "loss": 1.4427, + "theoretical_loss": 3.628780577933256, + "tokens_seen": 1154416640 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032839518555667, + "loss": 1.4738, + "theoretical_loss": 3.6287616404271397, + "tokens_seen": 1154482176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283851554663992, + "loss": 1.4309, + "theoretical_loss": 3.6287427042969953, + "tokens_seen": 1154547712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283751253761284, + "loss": 1.5021, + "theoretical_loss": 3.6287237695426455, + "tokens_seen": 1154613248 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032836509528585755, + "loss": 1.4171, + "theoretical_loss": 3.6287048361639114, + "tokens_seen": 1154678784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283550651955868, + "loss": 1.4447, + "theoretical_loss": 3.6286859041606156, + "tokens_seen": 1154744320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283450351053159, + "loss": 1.3687, + "theoretical_loss": 3.62866697353258, + "tokens_seen": 1154809856 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032833500501504515, + "loss": 1.3424, + "theoretical_loss": 3.6286480442796267, + "tokens_seen": 1154875392 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032832497492477433, + "loss": 1.4019, + "theoretical_loss": 3.628629116401578, + "tokens_seen": 1154940928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283149448345035, + "loss": 1.4184, + "theoretical_loss": 3.6286101898982555, + "tokens_seen": 1155006464 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003283049147442327, + "loss": 1.345, + "theoretical_loss": 3.628591264769481, + "tokens_seen": 1155072000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003282948846539619, + "loss": 1.3932, + "theoretical_loss": 3.628572341015078, + "tokens_seen": 1155137536 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032828485456369105, + "loss": 1.3995, + "theoretical_loss": 3.6285534186348682, + "tokens_seen": 1155203072 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003282748244734203, + "loss": 1.3122, + "theoretical_loss": 3.628534497628673, + "tokens_seen": 1155268608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003282647943831494, + "loss": 1.4221, + "theoretical_loss": 3.628515577996316, + "tokens_seen": 1155334144 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032825476429287865, + "loss": 1.3203, + "theoretical_loss": 3.6284966597376185, + "tokens_seen": 1155399680 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032824473420260783, + "loss": 1.4405, + "theoretical_loss": 3.628477742852403, + "tokens_seen": 1155465216 + }, + { + "epoch": 0.35, + "learning_rate": 0.000328234704112337, + "loss": 1.2755, + "theoretical_loss": 3.6284588273404927, + "tokens_seen": 1155530752 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003282246740220662, + "loss": 1.3869, + "theoretical_loss": 3.628439913201709, + "tokens_seen": 1155596288 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003282146439317954, + "loss": 1.4132, + "theoretical_loss": 3.6284210004358757, + "tokens_seen": 1155661824 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032820461384152456, + "loss": 1.3998, + "theoretical_loss": 3.6284020890428135, + "tokens_seen": 1155727360 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281945837512538, + "loss": 1.3697, + "theoretical_loss": 3.6283831790223458, + "tokens_seen": 1155792896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281845536609829, + "loss": 1.3577, + "theoretical_loss": 3.6283642703742958, + "tokens_seen": 1155858432 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032817452357071215, + "loss": 1.4293, + "theoretical_loss": 3.628345363098485, + "tokens_seen": 1155923968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281644934804413, + "loss": 1.3685, + "theoretical_loss": 3.6283264571947367, + "tokens_seen": 1155989504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281544633901705, + "loss": 1.3567, + "theoretical_loss": 3.628307552662873, + "tokens_seen": 1156055040 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281444332998997, + "loss": 1.4586, + "theoretical_loss": 3.628288649502718, + "tokens_seen": 1156120576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281344032096289, + "loss": 1.3618, + "theoretical_loss": 3.6282697477140924, + "tokens_seen": 1156186112 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032812437311935806, + "loss": 1.2973, + "theoretical_loss": 3.6282508472968202, + "tokens_seen": 1156251648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281143430290873, + "loss": 1.4, + "theoretical_loss": 3.6282319482507237, + "tokens_seen": 1156317184 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003281043129388164, + "loss": 1.398, + "theoretical_loss": 3.628213050575626, + "tokens_seen": 1156382720 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032809428284854566, + "loss": 1.2891, + "theoretical_loss": 3.62819415427135, + "tokens_seen": 1156448256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003280842527582748, + "loss": 1.3606, + "theoretical_loss": 3.628175259337719, + "tokens_seen": 1156513792 + }, + { + "epoch": 0.35, + "learning_rate": 0.000328074222668004, + "loss": 1.33, + "theoretical_loss": 3.628156365774555, + "tokens_seen": 1156579328 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032806419257773326, + "loss": 1.5053, + "theoretical_loss": 3.628137473581681, + "tokens_seen": 1156644864 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 423626, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2291412353515625, + "objective/train/theoretical_loss": 3.628118582758921, + "objective/train/tokens_used": 1177170400, + "theoretical_loss": 3.628118582758921, + "tokens_seen": 1156710400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003280541624874624, + "loss": 1.2974, + "theoretical_loss": 3.628118582758921, + "tokens_seen": 1156710400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003280441323971916, + "loss": 1.3133, + "theoretical_loss": 3.6280996933060967, + "tokens_seen": 1156775936 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032803410230692074, + "loss": 1.3514, + "theoretical_loss": 3.6280808052230316, + "tokens_seen": 1156841472 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032802407221665, + "loss": 1.3142, + "theoretical_loss": 3.6280619185095495, + "tokens_seen": 1156907008 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032801404212637916, + "loss": 1.4337, + "theoretical_loss": 3.628043033165473, + "tokens_seen": 1156972544 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032800401203610834, + "loss": 1.422, + "theoretical_loss": 3.6280241491906247, + "tokens_seen": 1157038080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279939819458375, + "loss": 1.2687, + "theoretical_loss": 3.628005266584829, + "tokens_seen": 1157103616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279839518555667, + "loss": 1.4309, + "theoretical_loss": 3.6279863853479077, + "tokens_seen": 1157169152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279739217652959, + "loss": 1.4561, + "theoretical_loss": 3.6279675054796856, + "tokens_seen": 1157234688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279638916750251, + "loss": 1.485, + "theoretical_loss": 3.6279486269799843, + "tokens_seen": 1157300224 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032795386158475425, + "loss": 1.3435, + "theoretical_loss": 3.6279297498486285, + "tokens_seen": 1157365760 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279438314944835, + "loss": 1.3587, + "theoretical_loss": 3.6279108740854413, + "tokens_seen": 1157431296 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032793380140421266, + "loss": 1.2666, + "theoretical_loss": 3.6278919996902452, + "tokens_seen": 1157496832 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032792377131394185, + "loss": 1.3931, + "theoretical_loss": 3.6278731266628643, + "tokens_seen": 1157562368 + }, + { + "epoch": 0.35, + "learning_rate": 0.000327913741223671, + "loss": 1.4655, + "theoretical_loss": 3.6278542550031223, + "tokens_seen": 1157627904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003279037111334002, + "loss": 1.4056, + "theoretical_loss": 3.627835384710842, + "tokens_seen": 1157693440 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003278936810431294, + "loss": 1.4133, + "theoretical_loss": 3.6278165157858475, + "tokens_seen": 1157758976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003278836509528586, + "loss": 1.3486, + "theoretical_loss": 3.627797648227962, + "tokens_seen": 1157824512 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032787362086258775, + "loss": 1.5089, + "theoretical_loss": 3.6277787820370087, + "tokens_seen": 1157890048 + }, + { + "epoch": 0.35, + "learning_rate": 0.000327863590772317, + "loss": 1.3805, + "theoretical_loss": 3.6277599172128125, + "tokens_seen": 1157955584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003278535606820461, + "loss": 1.3818, + "theoretical_loss": 3.6277410537551953, + "tokens_seen": 1158021120 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032784353059177535, + "loss": 1.3393, + "theoretical_loss": 3.627722191663983, + "tokens_seen": 1158086656 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032783350050150453, + "loss": 1.3119, + "theoretical_loss": 3.6277033309389966, + "tokens_seen": 1158152192 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003278234704112337, + "loss": 1.3819, + "theoretical_loss": 3.627684471580062, + "tokens_seen": 1158217728 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003278134403209629, + "loss": 1.4281, + "theoretical_loss": 3.627665613587002, + "tokens_seen": 1158283264 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032780341023069213, + "loss": 1.4427, + "theoretical_loss": 3.6276467569596402, + "tokens_seen": 1158348800 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032779338014042125, + "loss": 1.2714, + "theoretical_loss": 3.6276279016978012, + "tokens_seen": 1158414336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003277833500501505, + "loss": 1.3626, + "theoretical_loss": 3.6276090478013088, + "tokens_seen": 1158479872 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003277733199598796, + "loss": 1.5187, + "theoretical_loss": 3.6275901952699865, + "tokens_seen": 1158545408 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032776328986960885, + "loss": 1.4779, + "theoretical_loss": 3.627571344103658, + "tokens_seen": 1158610944 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032775325977933803, + "loss": 1.4114, + "theoretical_loss": 3.6275524943021478, + "tokens_seen": 1158676480 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003277432296890672, + "loss": 1.3993, + "theoretical_loss": 3.62753364586528, + "tokens_seen": 1158742016 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003277331995987964, + "loss": 1.4395, + "theoretical_loss": 3.6275147987928777, + "tokens_seen": 1158807552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003277231695085256, + "loss": 1.3208, + "theoretical_loss": 3.6274959530847664, + "tokens_seen": 1158873088 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032771313941825476, + "loss": 1.395, + "theoretical_loss": 3.6274771087407687, + "tokens_seen": 1158938624 + }, + { + "epoch": 0.35, + "learning_rate": 0.000327703109327984, + "loss": 1.4591, + "theoretical_loss": 3.6274582657607097, + "tokens_seen": 1159004160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276930792377131, + "loss": 1.3865, + "theoretical_loss": 3.6274394241444132, + "tokens_seen": 1159069696 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032768304914744235, + "loss": 1.4966, + "theoretical_loss": 3.6274205838917037, + "tokens_seen": 1159135232 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276730190571715, + "loss": 1.4638, + "theoretical_loss": 3.6274017450024054, + "tokens_seen": 1159200768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276629889669007, + "loss": 1.3481, + "theoretical_loss": 3.627382907476342, + "tokens_seen": 1159266304 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276529588766299, + "loss": 1.2381, + "theoretical_loss": 3.6273640713133384, + "tokens_seen": 1159331840 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276429287863591, + "loss": 1.3167, + "theoretical_loss": 3.6273452365132184, + "tokens_seen": 1159397376 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032763289869608826, + "loss": 1.4117, + "theoretical_loss": 3.627326403075807, + "tokens_seen": 1159462912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276228686058175, + "loss": 1.4361, + "theoretical_loss": 3.6273075710009284, + "tokens_seen": 1159528448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003276128385155466, + "loss": 1.3694, + "theoretical_loss": 3.6272887402884066, + "tokens_seen": 1159593984 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032760280842527586, + "loss": 1.4574, + "theoretical_loss": 3.6272699109380664, + "tokens_seen": 1159659520 + }, + { + "epoch": 0.35, + "learning_rate": 0.000327592778335005, + "loss": 1.3446, + "theoretical_loss": 3.627251082949732, + "tokens_seen": 1159725056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275827482447342, + "loss": 1.4221, + "theoretical_loss": 3.6272322563232287, + "tokens_seen": 1159790592 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275727181544634, + "loss": 1.3017, + "theoretical_loss": 3.6272134310583803, + "tokens_seen": 1159856128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275626880641926, + "loss": 1.3857, + "theoretical_loss": 3.627194607155011, + "tokens_seen": 1159921664 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 424758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4573270082473755, + "objective/train/theoretical_loss": 3.6271757846129464, + "objective/train/tokens_used": 1180447200, + "theoretical_loss": 3.6271757846129464, + "tokens_seen": 1159987200 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032755265797392176, + "loss": 1.4994, + "theoretical_loss": 3.6271757846129464, + "tokens_seen": 1159987200 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032754262788365094, + "loss": 1.3898, + "theoretical_loss": 3.6271569634320113, + "tokens_seen": 1160052736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275325977933801, + "loss": 1.4416, + "theoretical_loss": 3.6271381436120294, + "tokens_seen": 1160118272 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032752256770310936, + "loss": 1.4131, + "theoretical_loss": 3.627119325152826, + "tokens_seen": 1160183808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275125376128385, + "loss": 1.4187, + "theoretical_loss": 3.627100508054226, + "tokens_seen": 1160249344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003275025075225677, + "loss": 1.3965, + "theoretical_loss": 3.627081692316053, + "tokens_seen": 1160314880 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032749247743229685, + "loss": 1.3984, + "theoretical_loss": 3.6270628779381333, + "tokens_seen": 1160380416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003274824473420261, + "loss": 1.4381, + "theoretical_loss": 3.627044064920291, + "tokens_seen": 1160445952 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032747241725175527, + "loss": 1.4292, + "theoretical_loss": 3.6270252532623513, + "tokens_seen": 1160511488 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032746238716148445, + "loss": 1.3688, + "theoretical_loss": 3.627006442964139, + "tokens_seen": 1160577024 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032745235707121363, + "loss": 1.4584, + "theoretical_loss": 3.626987634025479, + "tokens_seen": 1160642560 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032744232698094286, + "loss": 1.3765, + "theoretical_loss": 3.626968826446196, + "tokens_seen": 1160708096 + }, + { + "epoch": 0.35, + "learning_rate": 0.000327432296890672, + "loss": 1.3402, + "theoretical_loss": 3.626950020226116, + "tokens_seen": 1160773632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003274222668004012, + "loss": 1.3415, + "theoretical_loss": 3.6269312153650626, + "tokens_seen": 1160839168 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032741223671013035, + "loss": 1.3667, + "theoretical_loss": 3.6269124118628615, + "tokens_seen": 1160904704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003274022066198596, + "loss": 1.4296, + "theoretical_loss": 3.626893609719338, + "tokens_seen": 1160970240 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032739217652958877, + "loss": 1.5192, + "theoretical_loss": 3.6268748089343177, + "tokens_seen": 1161035776 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032738214643931795, + "loss": 1.2964, + "theoretical_loss": 3.6268560095076245, + "tokens_seen": 1161101312 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032737211634904713, + "loss": 1.4664, + "theoretical_loss": 3.626837211439085, + "tokens_seen": 1161166848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003273620862587763, + "loss": 1.324, + "theoretical_loss": 3.626818414728523, + "tokens_seen": 1161232384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003273520561685055, + "loss": 1.3525, + "theoretical_loss": 3.626799619375765, + "tokens_seen": 1161297920 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032734202607823473, + "loss": 1.3295, + "theoretical_loss": 3.626780825380636, + "tokens_seen": 1161363456 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032733199598796386, + "loss": 1.3983, + "theoretical_loss": 3.6267620327429606, + "tokens_seen": 1161428992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003273219658976931, + "loss": 1.4358, + "theoretical_loss": 3.626743241462565, + "tokens_seen": 1161494528 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032731193580742233, + "loss": 1.5222, + "theoretical_loss": 3.6267244515392743, + "tokens_seen": 1161560064 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032730190571715145, + "loss": 1.3894, + "theoretical_loss": 3.626705662972914, + "tokens_seen": 1161625600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272918756268807, + "loss": 1.3471, + "theoretical_loss": 3.6266868757633093, + "tokens_seen": 1161691136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272818455366098, + "loss": 1.4162, + "theoretical_loss": 3.626668089910286, + "tokens_seen": 1161756672 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032727181544633905, + "loss": 1.4729, + "theoretical_loss": 3.62664930541367, + "tokens_seen": 1161822208 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032726178535606823, + "loss": 1.4476, + "theoretical_loss": 3.6266305222732855, + "tokens_seen": 1161887744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272517552657974, + "loss": 1.4199, + "theoretical_loss": 3.6266117404889595, + "tokens_seen": 1161953280 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272417251755266, + "loss": 1.4645, + "theoretical_loss": 3.626592960060517, + "tokens_seen": 1162018816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272316950852558, + "loss": 1.3503, + "theoretical_loss": 3.6265741809877836, + "tokens_seen": 1162084352 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032722166499498496, + "loss": 1.308, + "theoretical_loss": 3.6265554032705856, + "tokens_seen": 1162149888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272116349047142, + "loss": 1.3218, + "theoretical_loss": 3.626536626908748, + "tokens_seen": 1162215424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003272016048144433, + "loss": 1.4187, + "theoretical_loss": 3.626517851902097, + "tokens_seen": 1162280960 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032719157472417256, + "loss": 1.494, + "theoretical_loss": 3.6264990782504576, + "tokens_seen": 1162346496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271815446339017, + "loss": 1.2081, + "theoretical_loss": 3.626480305953657, + "tokens_seen": 1162412032 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271715145436309, + "loss": 1.5572, + "theoretical_loss": 3.6264615350115195, + "tokens_seen": 1162477568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271614844533601, + "loss": 1.3193, + "theoretical_loss": 3.6264427654238727, + "tokens_seen": 1162543104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271514543630893, + "loss": 1.3366, + "theoretical_loss": 3.6264239971905408, + "tokens_seen": 1162608640 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032714142427281846, + "loss": 1.3355, + "theoretical_loss": 3.6264052303113505, + "tokens_seen": 1162674176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271313941825477, + "loss": 1.297, + "theoretical_loss": 3.626386464786128, + "tokens_seen": 1162739712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271213640922768, + "loss": 1.3009, + "theoretical_loss": 3.626367700614699, + "tokens_seen": 1162805248 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032711133400200606, + "loss": 1.4773, + "theoretical_loss": 3.6263489377968896, + "tokens_seen": 1162870784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003271013039117352, + "loss": 1.4886, + "theoretical_loss": 3.6263301763325257, + "tokens_seen": 1162936320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270912738214644, + "loss": 1.5016, + "theoretical_loss": 3.6263114162214345, + "tokens_seen": 1163001856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270812437311936, + "loss": 1.4006, + "theoretical_loss": 3.6262926574634404, + "tokens_seen": 1163067392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270712136409228, + "loss": 1.4251, + "theoretical_loss": 3.6262739000583704, + "tokens_seen": 1163132928 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032706118355065196, + "loss": 1.3565, + "theoretical_loss": 3.626255144006051, + "tokens_seen": 1163198464 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 425912, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.398533582687378, + "objective/train/theoretical_loss": 3.6262363893063085, + "objective/train/tokens_used": 1183724000, + "theoretical_loss": 3.6262363893063085, + "tokens_seen": 1163264000 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032705115346038114, + "loss": 1.3489, + "theoretical_loss": 3.6262363893063085, + "tokens_seen": 1163264000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270411233701103, + "loss": 1.3636, + "theoretical_loss": 3.626217635958968, + "tokens_seen": 1163329536 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032703109327983956, + "loss": 1.4409, + "theoretical_loss": 3.6261988839638573, + "tokens_seen": 1163395072 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270210631895687, + "loss": 1.3665, + "theoretical_loss": 3.626180133320802, + "tokens_seen": 1163460608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003270110330992979, + "loss": 1.382, + "theoretical_loss": 3.6261613840296283, + "tokens_seen": 1163526144 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032700100300902705, + "loss": 1.2392, + "theoretical_loss": 3.6261426360901625, + "tokens_seen": 1163591680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003269909729187563, + "loss": 1.4114, + "theoretical_loss": 3.626123889502232, + "tokens_seen": 1163657216 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032698094282848547, + "loss": 1.3612, + "theoretical_loss": 3.6261051442656624, + "tokens_seen": 1163722752 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032697091273821465, + "loss": 1.4632, + "theoretical_loss": 3.62608640038028, + "tokens_seen": 1163788288 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032696088264794383, + "loss": 1.3416, + "theoretical_loss": 3.6260676578459123, + "tokens_seen": 1163853824 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032695085255767306, + "loss": 1.461, + "theoretical_loss": 3.626048916662385, + "tokens_seen": 1163919360 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003269408224674022, + "loss": 1.3967, + "theoretical_loss": 3.6260301768295253, + "tokens_seen": 1163984896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003269307923771314, + "loss": 1.359, + "theoretical_loss": 3.6260114383471587, + "tokens_seen": 1164050432 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032692076228686055, + "loss": 1.4352, + "theoretical_loss": 3.625992701215113, + "tokens_seen": 1164115968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003269107321965898, + "loss": 1.2683, + "theoretical_loss": 3.625973965433215, + "tokens_seen": 1164181504 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032690070210631897, + "loss": 1.3451, + "theoretical_loss": 3.6259552310012904, + "tokens_seen": 1164247040 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032689067201604815, + "loss": 1.3848, + "theoretical_loss": 3.6259364979191666, + "tokens_seen": 1164312576 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032688064192577733, + "loss": 1.3658, + "theoretical_loss": 3.62591776618667, + "tokens_seen": 1164378112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003268706118355065, + "loss": 1.2376, + "theoretical_loss": 3.625899035803628, + "tokens_seen": 1164443648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003268605817452357, + "loss": 1.3835, + "theoretical_loss": 3.625880306769867, + "tokens_seen": 1164509184 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032685055165496493, + "loss": 1.377, + "theoretical_loss": 3.625861579085214, + "tokens_seen": 1164574720 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032684052156469406, + "loss": 1.4687, + "theoretical_loss": 3.625842852749496, + "tokens_seen": 1164640256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003268304914744233, + "loss": 1.2917, + "theoretical_loss": 3.62582412776254, + "tokens_seen": 1164705792 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003268204613841524, + "loss": 1.4236, + "theoretical_loss": 3.625805404124172, + "tokens_seen": 1164771328 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032681043129388165, + "loss": 1.4642, + "theoretical_loss": 3.6257866818342204, + "tokens_seen": 1164836864 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032680040120361084, + "loss": 1.3767, + "theoretical_loss": 3.6257679608925115, + "tokens_seen": 1164902400 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032679037111334, + "loss": 1.4303, + "theoretical_loss": 3.6257492412988723, + "tokens_seen": 1164967936 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003267803410230692, + "loss": 1.4376, + "theoretical_loss": 3.6257305230531305, + "tokens_seen": 1165033472 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032677031093279843, + "loss": 1.3778, + "theoretical_loss": 3.6257118061551132, + "tokens_seen": 1165099008 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032676028084252756, + "loss": 1.3655, + "theoretical_loss": 3.6256930906046465, + "tokens_seen": 1165164544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003267502507522568, + "loss": 1.4835, + "theoretical_loss": 3.625674376401559, + "tokens_seen": 1165230080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003267402206619859, + "loss": 1.4438, + "theoretical_loss": 3.6256556635456763, + "tokens_seen": 1165295616 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032673019057171516, + "loss": 1.3566, + "theoretical_loss": 3.6256369520368272, + "tokens_seen": 1165361152 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032672016048144434, + "loss": 1.3562, + "theoretical_loss": 3.6256182418748377, + "tokens_seen": 1165426688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003267101303911735, + "loss": 1.4028, + "theoretical_loss": 3.6255995330595363, + "tokens_seen": 1165492224 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003267001003009027, + "loss": 1.4752, + "theoretical_loss": 3.6255808255907502, + "tokens_seen": 1165557760 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003266900702106319, + "loss": 1.3983, + "theoretical_loss": 3.625562119468306, + "tokens_seen": 1165623296 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032668004012036106, + "loss": 1.5127, + "theoretical_loss": 3.6255434146920313, + "tokens_seen": 1165688832 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003266700100300903, + "loss": 1.4223, + "theoretical_loss": 3.6255247112617544, + "tokens_seen": 1165754368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003266599799398194, + "loss": 1.3006, + "theoretical_loss": 3.625506009177302, + "tokens_seen": 1165819904 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032664994984954866, + "loss": 1.3518, + "theoretical_loss": 3.625487308438502, + "tokens_seen": 1165885440 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032663991975927784, + "loss": 1.4359, + "theoretical_loss": 3.625468609045181, + "tokens_seen": 1165950976 + }, + { + "epoch": 0.35, + "learning_rate": 0.000326629889669007, + "loss": 1.4643, + "theoretical_loss": 3.625449910997168, + "tokens_seen": 1166016512 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003266198595787362, + "loss": 1.3496, + "theoretical_loss": 3.6254312142942897, + "tokens_seen": 1166082048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003266098294884654, + "loss": 1.3278, + "theoretical_loss": 3.6254125189363737, + "tokens_seen": 1166147584 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032659979939819457, + "loss": 1.3691, + "theoretical_loss": 3.6253938249232482, + "tokens_seen": 1166213120 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003265897693079238, + "loss": 1.3274, + "theoretical_loss": 3.6253751322547405, + "tokens_seen": 1166278656 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032657973921765293, + "loss": 1.4243, + "theoretical_loss": 3.6253564409306787, + "tokens_seen": 1166344192 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032656970912738216, + "loss": 1.5144, + "theoretical_loss": 3.62533775095089, + "tokens_seen": 1166409728 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032655967903711135, + "loss": 1.4106, + "theoretical_loss": 3.625319062315203, + "tokens_seen": 1166475264 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 427016, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4649659395217896, + "objective/train/theoretical_loss": 3.625300375023445, + "objective/train/tokens_used": 1187000800, + "theoretical_loss": 3.625300375023445, + "tokens_seen": 1166540800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003265496489468405, + "loss": 1.4446, + "theoretical_loss": 3.625300375023445, + "tokens_seen": 1166540800 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032653961885656976, + "loss": 1.3601, + "theoretical_loss": 3.625281689075444, + "tokens_seen": 1166606336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003265295887662989, + "loss": 1.4348, + "theoretical_loss": 3.6252630044710275, + "tokens_seen": 1166671872 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003265195586760281, + "loss": 1.5032, + "theoretical_loss": 3.625244321210024, + "tokens_seen": 1166737408 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032650952858575725, + "loss": 1.3752, + "theoretical_loss": 3.625225639292261, + "tokens_seen": 1166802944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003264994984954865, + "loss": 1.398, + "theoretical_loss": 3.6252069587175666, + "tokens_seen": 1166868480 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032648946840521567, + "loss": 1.3841, + "theoretical_loss": 3.6251882794857693, + "tokens_seen": 1166934016 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032647943831494485, + "loss": 1.449, + "theoretical_loss": 3.6251696015966965, + "tokens_seen": 1166999552 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032646940822467403, + "loss": 1.4734, + "theoretical_loss": 3.6251509250501766, + "tokens_seen": 1167065088 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032645937813440326, + "loss": 1.4472, + "theoretical_loss": 3.625132249846038, + "tokens_seen": 1167130624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003264493480441324, + "loss": 1.4674, + "theoretical_loss": 3.625113575984108, + "tokens_seen": 1167196160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003264393179538616, + "loss": 1.3378, + "theoretical_loss": 3.6250949034642153, + "tokens_seen": 1167261696 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032642928786359075, + "loss": 1.3021, + "theoretical_loss": 3.6250762322861885, + "tokens_seen": 1167327232 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032641925777332, + "loss": 1.4038, + "theoretical_loss": 3.6250575624498556, + "tokens_seen": 1167392768 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032640922768304917, + "loss": 1.4039, + "theoretical_loss": 3.625038893955044, + "tokens_seen": 1167458304 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032639919759277835, + "loss": 1.4313, + "theoretical_loss": 3.625020226801584, + "tokens_seen": 1167523840 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032638916750250753, + "loss": 1.4287, + "theoretical_loss": 3.6250015609893014, + "tokens_seen": 1167589376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003263791374122367, + "loss": 1.4124, + "theoretical_loss": 3.624982896518026, + "tokens_seen": 1167654912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003263691073219659, + "loss": 1.3449, + "theoretical_loss": 3.624964233387587, + "tokens_seen": 1167720448 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032635907723169513, + "loss": 1.3533, + "theoretical_loss": 3.624945571597811, + "tokens_seen": 1167785984 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032634904714142426, + "loss": 1.3996, + "theoretical_loss": 3.6249269111485276, + "tokens_seen": 1167851520 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003263390170511535, + "loss": 1.3267, + "theoretical_loss": 3.624908252039565, + "tokens_seen": 1167917056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003263289869608826, + "loss": 1.4051, + "theoretical_loss": 3.6248895942707517, + "tokens_seen": 1167982592 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032631895687061185, + "loss": 1.3994, + "theoretical_loss": 3.6248709378419166, + "tokens_seen": 1168048128 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032630892678034104, + "loss": 1.4484, + "theoretical_loss": 3.6248522827528875, + "tokens_seen": 1168113664 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003262988966900702, + "loss": 1.4479, + "theoretical_loss": 3.624833629003494, + "tokens_seen": 1168179200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003262888665997994, + "loss": 1.4069, + "theoretical_loss": 3.624814976593564, + "tokens_seen": 1168244736 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032627883650952863, + "loss": 1.3865, + "theoretical_loss": 3.6247963255229267, + "tokens_seen": 1168310272 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032626880641925776, + "loss": 1.4541, + "theoretical_loss": 3.6247776757914103, + "tokens_seen": 1168375808 + }, + { + "epoch": 0.35, + "learning_rate": 0.000326258776328987, + "loss": 1.3537, + "theoretical_loss": 3.624759027398844, + "tokens_seen": 1168441344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003262487462387161, + "loss": 1.2884, + "theoretical_loss": 3.624740380345057, + "tokens_seen": 1168506880 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032623871614844536, + "loss": 1.4401, + "theoretical_loss": 3.6247217346298766, + "tokens_seen": 1168572416 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032622868605817454, + "loss": 1.4018, + "theoretical_loss": 3.624703090253133, + "tokens_seen": 1168637952 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003262186559679037, + "loss": 1.3571, + "theoretical_loss": 3.6246844472146544, + "tokens_seen": 1168703488 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003262086258776329, + "loss": 1.3408, + "theoretical_loss": 3.62466580551427, + "tokens_seen": 1168769024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261985957873621, + "loss": 1.2903, + "theoretical_loss": 3.624647165151809, + "tokens_seen": 1168834560 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032618856569709126, + "loss": 1.4892, + "theoretical_loss": 3.6246285261270996, + "tokens_seen": 1168900096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261785356068205, + "loss": 1.371, + "theoretical_loss": 3.624609888439972, + "tokens_seen": 1168965632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261685055165496, + "loss": 1.3412, + "theoretical_loss": 3.6245912520902537, + "tokens_seen": 1169031168 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032615847542627886, + "loss": 1.5237, + "theoretical_loss": 3.624572617077775, + "tokens_seen": 1169096704 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032614844533600804, + "loss": 1.3283, + "theoretical_loss": 3.624553983402364, + "tokens_seen": 1169162240 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261384152457372, + "loss": 1.3554, + "theoretical_loss": 3.6245353510638507, + "tokens_seen": 1169227776 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261283851554664, + "loss": 1.2839, + "theoretical_loss": 3.624516720062064, + "tokens_seen": 1169293312 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003261183550651956, + "loss": 1.3685, + "theoretical_loss": 3.624498090396833, + "tokens_seen": 1169358848 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032610832497492477, + "loss": 1.3823, + "theoretical_loss": 3.6244794620679865, + "tokens_seen": 1169424384 + }, + { + "epoch": 0.35, + "learning_rate": 0.000326098294884654, + "loss": 1.4728, + "theoretical_loss": 3.6244608350753547, + "tokens_seen": 1169489920 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032608826479438313, + "loss": 1.3204, + "theoretical_loss": 3.6244422094187665, + "tokens_seen": 1169555456 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032607823470411236, + "loss": 1.3668, + "theoretical_loss": 3.6244235850980506, + "tokens_seen": 1169620992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003260682046138415, + "loss": 1.4646, + "theoretical_loss": 3.624404962113037, + "tokens_seen": 1169686528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003260581745235707, + "loss": 1.4969, + "theoretical_loss": 3.6243863404635546, + "tokens_seen": 1169752064 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 428177, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.430919885635376, + "objective/train/theoretical_loss": 3.6243677201494338, + "objective/train/tokens_used": 1190277600, + "theoretical_loss": 3.6243677201494338, + "tokens_seen": 1169817600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003260481444332999, + "loss": 1.465, + "theoretical_loss": 3.6243677201494338, + "tokens_seen": 1169817600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003260381143430291, + "loss": 1.482, + "theoretical_loss": 3.6243491011705027, + "tokens_seen": 1169883136 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032602808425275827, + "loss": 1.3814, + "theoretical_loss": 3.6243304835265917, + "tokens_seen": 1169948672 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032601805416248745, + "loss": 1.3172, + "theoretical_loss": 3.62431186721753, + "tokens_seen": 1170014208 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032600802407221663, + "loss": 1.4712, + "theoretical_loss": 3.624293252243147, + "tokens_seen": 1170079744 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032599799398194587, + "loss": 1.4972, + "theoretical_loss": 3.6242746386032723, + "tokens_seen": 1170145280 + }, + { + "epoch": 0.35, + "learning_rate": 0.000325987963891675, + "loss": 1.3492, + "theoretical_loss": 3.624256026297736, + "tokens_seen": 1170210816 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032597793380140423, + "loss": 1.3863, + "theoretical_loss": 3.624237415326367, + "tokens_seen": 1170276352 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003259679037111334, + "loss": 1.2787, + "theoretical_loss": 3.6242188056889955, + "tokens_seen": 1170341888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003259578736208626, + "loss": 1.3474, + "theoretical_loss": 3.6242001973854507, + "tokens_seen": 1170407424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003259478435305918, + "loss": 1.4121, + "theoretical_loss": 3.6241815904155628, + "tokens_seen": 1170472960 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032593781344032095, + "loss": 1.4237, + "theoretical_loss": 3.6241629847791614, + "tokens_seen": 1170538496 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032592778335005014, + "loss": 1.4816, + "theoretical_loss": 3.6241443804760762, + "tokens_seen": 1170604032 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032591775325977937, + "loss": 1.403, + "theoretical_loss": 3.624125777506137, + "tokens_seen": 1170669568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003259077231695085, + "loss": 1.4038, + "theoretical_loss": 3.624107175869174, + "tokens_seen": 1170735104 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032589769307923773, + "loss": 1.3582, + "theoretical_loss": 3.6240885755650165, + "tokens_seen": 1170800640 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032588766298896686, + "loss": 1.3919, + "theoretical_loss": 3.6240699765934945, + "tokens_seen": 1170866176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003258776328986961, + "loss": 1.4703, + "theoretical_loss": 3.624051378954438, + "tokens_seen": 1170931712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003258676028084253, + "loss": 1.4701, + "theoretical_loss": 3.624032782647678, + "tokens_seen": 1170997248 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032585757271815446, + "loss": 1.4215, + "theoretical_loss": 3.6240141876730423, + "tokens_seen": 1171062784 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032584754262788364, + "loss": 1.4615, + "theoretical_loss": 3.6239955940303634, + "tokens_seen": 1171128320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003258375125376128, + "loss": 1.3152, + "theoretical_loss": 3.6239770017194695, + "tokens_seen": 1171193856 + }, + { + "epoch": 0.35, + "learning_rate": 0.000325827482447342, + "loss": 1.5145, + "theoretical_loss": 3.6239584107401916, + "tokens_seen": 1171259392 + }, + { + "epoch": 0.35, + "learning_rate": 0.00032581745235707124, + "loss": 1.3908, + "theoretical_loss": 3.6239398210923595, + "tokens_seen": 1171324928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003258074222668004, + "loss": 1.4418, + "theoretical_loss": 3.6239212327758037, + "tokens_seen": 1171390464 + }, + { + "epoch": 0.35, + "learning_rate": 0.0003257973921765296, + "loss": 1.333, + "theoretical_loss": 3.623902645790354, + "tokens_seen": 1171456000 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032578736208625883, + "loss": 1.3919, + "theoretical_loss": 3.6238840601358406, + "tokens_seen": 1171521536 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032577733199598796, + "loss": 1.4269, + "theoretical_loss": 3.6238654758120945, + "tokens_seen": 1171587072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003257673019057172, + "loss": 1.4066, + "theoretical_loss": 3.6238468928189453, + "tokens_seen": 1171652608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003257572718154463, + "loss": 1.4204, + "theoretical_loss": 3.623828311156223, + "tokens_seen": 1171718144 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032574724172517556, + "loss": 1.4176, + "theoretical_loss": 3.6238097308237585, + "tokens_seen": 1171783680 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032573721163490474, + "loss": 1.3738, + "theoretical_loss": 3.623791151821382, + "tokens_seen": 1171849216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003257271815446339, + "loss": 1.4786, + "theoretical_loss": 3.623772574148924, + "tokens_seen": 1171914752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003257171514543631, + "loss": 1.3186, + "theoretical_loss": 3.623753997806215, + "tokens_seen": 1171980288 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003257071213640923, + "loss": 1.3642, + "theoretical_loss": 3.6237354227930854, + "tokens_seen": 1172045824 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032569709127382146, + "loss": 1.4666, + "theoretical_loss": 3.6237168491093654, + "tokens_seen": 1172111360 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256870611835507, + "loss": 1.2518, + "theoretical_loss": 3.6236982767548858, + "tokens_seen": 1172176896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256770310932798, + "loss": 1.3529, + "theoretical_loss": 3.6236797057294776, + "tokens_seen": 1172242432 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032566700100300906, + "loss": 1.3126, + "theoretical_loss": 3.623661136032971, + "tokens_seen": 1172307968 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032565697091273824, + "loss": 1.1895, + "theoretical_loss": 3.623642567665196, + "tokens_seen": 1172373504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256469408224674, + "loss": 1.3927, + "theoretical_loss": 3.623624000625984, + "tokens_seen": 1172439040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256369107321966, + "loss": 1.3513, + "theoretical_loss": 3.6236054349151656, + "tokens_seen": 1172504576 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256268806419258, + "loss": 1.3926, + "theoretical_loss": 3.623586870532572, + "tokens_seen": 1172570112 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032561685055165497, + "loss": 1.3965, + "theoretical_loss": 3.623568307478032, + "tokens_seen": 1172635648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003256068204613842, + "loss": 1.3084, + "theoretical_loss": 3.623549745751379, + "tokens_seen": 1172701184 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032559679037111333, + "loss": 1.4868, + "theoretical_loss": 3.623531185352442, + "tokens_seen": 1172766720 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032558676028084256, + "loss": 1.4239, + "theoretical_loss": 3.6235126262810526, + "tokens_seen": 1172832256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003255767301905717, + "loss": 1.3772, + "theoretical_loss": 3.623494068537042, + "tokens_seen": 1172897792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003255667001003009, + "loss": 1.5093, + "theoretical_loss": 3.6234755121202396, + "tokens_seen": 1172963328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003255566700100301, + "loss": 1.4192, + "theoretical_loss": 3.623456957030478, + "tokens_seen": 1173028864 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 429267, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.404681921005249, + "objective/train/theoretical_loss": 3.623438403267587, + "objective/train/tokens_used": 1193554400, + "theoretical_loss": 3.623438403267587, + "tokens_seen": 1173094400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003255466399197593, + "loss": 1.3204, + "theoretical_loss": 3.623438403267587, + "tokens_seen": 1173094400 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032553660982948847, + "loss": 1.3692, + "theoretical_loss": 3.6234198508313984, + "tokens_seen": 1173159936 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032552657973921765, + "loss": 1.4268, + "theoretical_loss": 3.623401299721743, + "tokens_seen": 1173225472 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032551654964894683, + "loss": 1.4119, + "theoretical_loss": 3.6233827499384517, + "tokens_seen": 1173291008 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032550651955867607, + "loss": 1.4398, + "theoretical_loss": 3.6233642014813556, + "tokens_seen": 1173356544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003254964894684052, + "loss": 1.4005, + "theoretical_loss": 3.6233456543502855, + "tokens_seen": 1173422080 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032548645937813443, + "loss": 1.3884, + "theoretical_loss": 3.623327108545074, + "tokens_seen": 1173487616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003254764292878636, + "loss": 1.3199, + "theoretical_loss": 3.6233085640655505, + "tokens_seen": 1173553152 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003254663991975928, + "loss": 1.3384, + "theoretical_loss": 3.623290020911547, + "tokens_seen": 1173618688 + }, + { + "epoch": 0.36, + "learning_rate": 0.000325456369107322, + "loss": 1.3924, + "theoretical_loss": 3.6232714790828946, + "tokens_seen": 1173684224 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032544633901705115, + "loss": 1.4426, + "theoretical_loss": 3.6232529385794248, + "tokens_seen": 1173749760 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032543630892678034, + "loss": 1.4827, + "theoretical_loss": 3.6232343994009684, + "tokens_seen": 1173815296 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032542627883650957, + "loss": 1.437, + "theoretical_loss": 3.623215861547358, + "tokens_seen": 1173880832 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003254162487462387, + "loss": 1.4054, + "theoretical_loss": 3.623197325018423, + "tokens_seen": 1173946368 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032540621865596793, + "loss": 1.3115, + "theoretical_loss": 3.6231787898139967, + "tokens_seen": 1174011904 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032539618856569706, + "loss": 1.4677, + "theoretical_loss": 3.623160255933909, + "tokens_seen": 1174077440 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003253861584754263, + "loss": 1.4082, + "theoretical_loss": 3.6231417233779926, + "tokens_seen": 1174142976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003253761283851555, + "loss": 1.333, + "theoretical_loss": 3.6231231921460783, + "tokens_seen": 1174208512 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032536609829488466, + "loss": 1.2979, + "theoretical_loss": 3.6231046622379974, + "tokens_seen": 1174274048 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032535606820461384, + "loss": 1.3654, + "theoretical_loss": 3.6230861336535822, + "tokens_seen": 1174339584 + }, + { + "epoch": 0.36, + "learning_rate": 0.000325346038114343, + "loss": 1.344, + "theoretical_loss": 3.623067606392664, + "tokens_seen": 1174405120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003253360080240722, + "loss": 1.3104, + "theoretical_loss": 3.6230490804550737, + "tokens_seen": 1174470656 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032532597793380144, + "loss": 1.3464, + "theoretical_loss": 3.6230305558406446, + "tokens_seen": 1174536192 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032531594784353056, + "loss": 1.4714, + "theoretical_loss": 3.623012032549206, + "tokens_seen": 1174601728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003253059177532598, + "loss": 1.4544, + "theoretical_loss": 3.622993510580592, + "tokens_seen": 1174667264 + }, + { + "epoch": 0.36, + "learning_rate": 0.000325295887662989, + "loss": 1.4894, + "theoretical_loss": 3.622974989934633, + "tokens_seen": 1174732800 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032528585757271816, + "loss": 1.4627, + "theoretical_loss": 3.622956470611161, + "tokens_seen": 1174798336 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032527582748244734, + "loss": 1.3603, + "theoretical_loss": 3.622937952610008, + "tokens_seen": 1174863872 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003252657973921765, + "loss": 1.3223, + "theoretical_loss": 3.6229194359310055, + "tokens_seen": 1174929408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003252557673019057, + "loss": 1.241, + "theoretical_loss": 3.6229009205739855, + "tokens_seen": 1174994944 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032524573721163494, + "loss": 1.3912, + "theoretical_loss": 3.62288240653878, + "tokens_seen": 1175060480 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032523570712136407, + "loss": 1.4425, + "theoretical_loss": 3.6228638938252216, + "tokens_seen": 1175126016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003252256770310933, + "loss": 1.4119, + "theoretical_loss": 3.6228453824331406, + "tokens_seen": 1175191552 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032521564694082243, + "loss": 1.3252, + "theoretical_loss": 3.6228268723623698, + "tokens_seen": 1175257088 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032520561685055166, + "loss": 1.4063, + "theoretical_loss": 3.6228083636127417, + "tokens_seen": 1175322624 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032519558676028084, + "loss": 1.4375, + "theoretical_loss": 3.622789856184088, + "tokens_seen": 1175388160 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032518555667001, + "loss": 1.4354, + "theoretical_loss": 3.6227713500762406, + "tokens_seen": 1175453696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003251755265797392, + "loss": 1.4096, + "theoretical_loss": 3.622752845289032, + "tokens_seen": 1175519232 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032516549648946844, + "loss": 1.3098, + "theoretical_loss": 3.6227343418222935, + "tokens_seen": 1175584768 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032515546639919757, + "loss": 1.4671, + "theoretical_loss": 3.622715839675858, + "tokens_seen": 1175650304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003251454363089268, + "loss": 1.3032, + "theoretical_loss": 3.622697338849558, + "tokens_seen": 1175715840 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032513540621865593, + "loss": 1.5451, + "theoretical_loss": 3.6226788393432248, + "tokens_seen": 1175781376 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032512537612838517, + "loss": 1.2377, + "theoretical_loss": 3.6226603411566916, + "tokens_seen": 1175846912 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032511534603811435, + "loss": 1.4305, + "theoretical_loss": 3.62264184428979, + "tokens_seen": 1175912448 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032510531594784353, + "loss": 1.2548, + "theoretical_loss": 3.6226233487423523, + "tokens_seen": 1175977984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003250952858575727, + "loss": 1.3783, + "theoretical_loss": 3.6226048545142113, + "tokens_seen": 1176043520 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003250852557673019, + "loss": 1.4614, + "theoretical_loss": 3.622586361605199, + "tokens_seen": 1176109056 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032507522567703107, + "loss": 1.4486, + "theoretical_loss": 3.6225678700151485, + "tokens_seen": 1176174592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003250651955867603, + "loss": 1.3086, + "theoretical_loss": 3.6225493797438917, + "tokens_seen": 1176240128 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003250551654964895, + "loss": 1.4105, + "theoretical_loss": 3.6225308907912606, + "tokens_seen": 1176305664 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 430645, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.510002851486206, + "objective/train/theoretical_loss": 3.6225124031570886, + "objective/train/tokens_used": 1196831200, + "theoretical_loss": 3.6225124031570886, + "tokens_seen": 1176371200 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032504513540621867, + "loss": 1.4026, + "theoretical_loss": 3.6225124031570886, + "tokens_seen": 1176371200 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032503510531594785, + "loss": 1.2886, + "theoretical_loss": 3.622493916841208, + "tokens_seen": 1176436736 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032502507522567703, + "loss": 1.454, + "theoretical_loss": 3.6224754318434504, + "tokens_seen": 1176502272 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032501504513540627, + "loss": 1.4953, + "theoretical_loss": 3.62245694816365, + "tokens_seen": 1176567808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003250050150451354, + "loss": 1.4779, + "theoretical_loss": 3.622438465801638, + "tokens_seen": 1176633344 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032499498495486463, + "loss": 1.4226, + "theoretical_loss": 3.622419984757248, + "tokens_seen": 1176698880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003249849548645938, + "loss": 1.4861, + "theoretical_loss": 3.6224015050303127, + "tokens_seen": 1176764416 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324974924774323, + "loss": 1.4264, + "theoretical_loss": 3.6223830266206645, + "tokens_seen": 1176829952 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003249648946840522, + "loss": 1.4139, + "theoretical_loss": 3.622364549528136, + "tokens_seen": 1176895488 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032495486459378135, + "loss": 1.3315, + "theoretical_loss": 3.62234607375256, + "tokens_seen": 1176961024 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032494483450351054, + "loss": 1.3811, + "theoretical_loss": 3.6223275992937696, + "tokens_seen": 1177026560 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032493480441323977, + "loss": 1.4786, + "theoretical_loss": 3.6223091261515976, + "tokens_seen": 1177092096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003249247743229689, + "loss": 1.5033, + "theoretical_loss": 3.6222906543258766, + "tokens_seen": 1177157632 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032491474423269813, + "loss": 1.2722, + "theoretical_loss": 3.6222721838164396, + "tokens_seen": 1177223168 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032490471414242726, + "loss": 1.4556, + "theoretical_loss": 3.6222537146231195, + "tokens_seen": 1177288704 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003248946840521565, + "loss": 1.4206, + "theoretical_loss": 3.62223524674575, + "tokens_seen": 1177354240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003248846539618857, + "loss": 1.291, + "theoretical_loss": 3.622216780184163, + "tokens_seen": 1177419776 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032487462387161486, + "loss": 1.3664, + "theoretical_loss": 3.622198314938192, + "tokens_seen": 1177485312 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032486459378134404, + "loss": 1.4673, + "theoretical_loss": 3.6221798510076706, + "tokens_seen": 1177550848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003248545636910732, + "loss": 1.427, + "theoretical_loss": 3.6221613883924304, + "tokens_seen": 1177616384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003248445336008024, + "loss": 1.4545, + "theoretical_loss": 3.622142927092306, + "tokens_seen": 1177681920 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032483450351053164, + "loss": 1.4335, + "theoretical_loss": 3.6221244671071298, + "tokens_seen": 1177747456 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032482447342026076, + "loss": 1.3649, + "theoretical_loss": 3.6221060084367354, + "tokens_seen": 1177812992 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032481444332999, + "loss": 1.5048, + "theoretical_loss": 3.6220875510809556, + "tokens_seen": 1177878528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003248044132397192, + "loss": 1.3738, + "theoretical_loss": 3.6220690950396235, + "tokens_seen": 1177944064 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032479438314944836, + "loss": 1.4034, + "theoretical_loss": 3.622050640312573, + "tokens_seen": 1178009600 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032478435305917754, + "loss": 1.4478, + "theoretical_loss": 3.622032186899637, + "tokens_seen": 1178075136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003247743229689067, + "loss": 1.3066, + "theoretical_loss": 3.622013734800649, + "tokens_seen": 1178140672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003247642928786359, + "loss": 1.3535, + "theoretical_loss": 3.6219952840154424, + "tokens_seen": 1178206208 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032475426278836514, + "loss": 1.4774, + "theoretical_loss": 3.62197683454385, + "tokens_seen": 1178271744 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032474423269809427, + "loss": 1.3111, + "theoretical_loss": 3.6219583863857068, + "tokens_seen": 1178337280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003247342026078235, + "loss": 1.5437, + "theoretical_loss": 3.6219399395408436, + "tokens_seen": 1178402816 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032472417251755263, + "loss": 1.4057, + "theoretical_loss": 3.6219214940090962, + "tokens_seen": 1178468352 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032471414242728186, + "loss": 1.434, + "theoretical_loss": 3.6219030497902978, + "tokens_seen": 1178533888 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032470411233701104, + "loss": 1.4376, + "theoretical_loss": 3.6218846068842807, + "tokens_seen": 1178599424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003246940822467402, + "loss": 1.3475, + "theoretical_loss": 3.6218661652908795, + "tokens_seen": 1178664960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003246840521564694, + "loss": 1.5048, + "theoretical_loss": 3.6218477250099275, + "tokens_seen": 1178730496 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032467402206619864, + "loss": 1.3507, + "theoretical_loss": 3.621829286041258, + "tokens_seen": 1178796032 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032466399197592777, + "loss": 1.3164, + "theoretical_loss": 3.6218108483847056, + "tokens_seen": 1178861568 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324653961885657, + "loss": 1.4683, + "theoretical_loss": 3.6217924120401035, + "tokens_seen": 1178927104 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032464393179538613, + "loss": 1.5227, + "theoretical_loss": 3.621773977007285, + "tokens_seen": 1178992640 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032463390170511537, + "loss": 1.4511, + "theoretical_loss": 3.6217555432860844, + "tokens_seen": 1179058176 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032462387161484455, + "loss": 1.5285, + "theoretical_loss": 3.621737110876335, + "tokens_seen": 1179123712 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032461384152457373, + "loss": 1.7162, + "theoretical_loss": 3.6217186797778713, + "tokens_seen": 1179189248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003246038114343029, + "loss": 1.4455, + "theoretical_loss": 3.6217002499905266, + "tokens_seen": 1179254784 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003245937813440321, + "loss": 1.4774, + "theoretical_loss": 3.6216818215141346, + "tokens_seen": 1179320320 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032458375125376127, + "loss": 1.5161, + "theoretical_loss": 3.62166339434853, + "tokens_seen": 1179385856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003245737211634905, + "loss": 1.4679, + "theoretical_loss": 3.6216449684935466, + "tokens_seen": 1179451392 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032456369107321963, + "loss": 1.4183, + "theoretical_loss": 3.6216265439490174, + "tokens_seen": 1179516928 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032455366098294887, + "loss": 1.3582, + "theoretical_loss": 3.621608120714777, + "tokens_seen": 1179582464 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 431781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.442383050918579, + "objective/train/theoretical_loss": 3.6215896987906597, + "objective/train/tokens_used": 1200108000, + "theoretical_loss": 3.6215896987906597, + "tokens_seen": 1179648000 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324543630892678, + "loss": 1.55, + "theoretical_loss": 3.6215896987906597, + "tokens_seen": 1179648000 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032453360080240723, + "loss": 1.4066, + "theoretical_loss": 3.621571278176499, + "tokens_seen": 1179713536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003245235707121364, + "loss": 1.4231, + "theoretical_loss": 3.6215528588721297, + "tokens_seen": 1179779072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003245135406218656, + "loss": 1.5092, + "theoretical_loss": 3.6215344408773857, + "tokens_seen": 1179844608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003245035105315948, + "loss": 1.4451, + "theoretical_loss": 3.6215160241921005, + "tokens_seen": 1179910144 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324493480441324, + "loss": 1.3747, + "theoretical_loss": 3.621497608816109, + "tokens_seen": 1179975680 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032448345035105314, + "loss": 1.5229, + "theoretical_loss": 3.621479194749245, + "tokens_seen": 1180041216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003244734202607824, + "loss": 1.4684, + "theoretical_loss": 3.6214607819913427, + "tokens_seen": 1180106752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003244633901705115, + "loss": 1.4155, + "theoretical_loss": 3.621442370542237, + "tokens_seen": 1180172288 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032445336008024074, + "loss": 1.4038, + "theoretical_loss": 3.6214239604017617, + "tokens_seen": 1180237824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003244433299899699, + "loss": 1.4379, + "theoretical_loss": 3.621405551569751, + "tokens_seen": 1180303360 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003244332998996991, + "loss": 1.4972, + "theoretical_loss": 3.6213871440460395, + "tokens_seen": 1180368896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003244232698094283, + "loss": 1.4289, + "theoretical_loss": 3.6213687378304615, + "tokens_seen": 1180434432 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032441323971915746, + "loss": 1.3008, + "theoretical_loss": 3.6213503329228516, + "tokens_seen": 1180499968 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032440320962888664, + "loss": 1.442, + "theoretical_loss": 3.621331929323044, + "tokens_seen": 1180565504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003243931795386159, + "loss": 1.4111, + "theoretical_loss": 3.621313527030873, + "tokens_seen": 1180631040 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324383149448345, + "loss": 1.4907, + "theoretical_loss": 3.621295126046174, + "tokens_seen": 1180696576 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032437311935807424, + "loss": 1.4427, + "theoretical_loss": 3.6212767263687806, + "tokens_seen": 1180762112 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032436308926780337, + "loss": 1.3906, + "theoretical_loss": 3.621258327998528, + "tokens_seen": 1180827648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003243530591775326, + "loss": 1.433, + "theoretical_loss": 3.62123993093525, + "tokens_seen": 1180893184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003243430290872618, + "loss": 1.5047, + "theoretical_loss": 3.621221535178782, + "tokens_seen": 1180958720 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032433299899699096, + "loss": 1.5675, + "theoretical_loss": 3.6212031407289587, + "tokens_seen": 1181024256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003243229689067202, + "loss": 1.4067, + "theoretical_loss": 3.621184747585614, + "tokens_seen": 1181089792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003243129388164494, + "loss": 1.5064, + "theoretical_loss": 3.621166355748583, + "tokens_seen": 1181155328 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032430290872617856, + "loss": 1.3352, + "theoretical_loss": 3.621147965217701, + "tokens_seen": 1181220864 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032429287863590774, + "loss": 1.5083, + "theoretical_loss": 3.621129575992802, + "tokens_seen": 1181286400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003242828485456369, + "loss": 1.3876, + "theoretical_loss": 3.6211111880737215, + "tokens_seen": 1181351936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003242728184553661, + "loss": 1.4568, + "theoretical_loss": 3.6210928014602937, + "tokens_seen": 1181417472 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032426278836509534, + "loss": 1.3993, + "theoretical_loss": 3.621074416152354, + "tokens_seen": 1181483008 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032425275827482447, + "loss": 1.5202, + "theoretical_loss": 3.621056032149737, + "tokens_seen": 1181548544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003242427281845537, + "loss": 1.4626, + "theoretical_loss": 3.6210376494522776, + "tokens_seen": 1181614080 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032423269809428283, + "loss": 1.353, + "theoretical_loss": 3.621019268059811, + "tokens_seen": 1181679616 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032422266800401206, + "loss": 1.3818, + "theoretical_loss": 3.621000887972172, + "tokens_seen": 1181745152 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032421263791374125, + "loss": 1.5311, + "theoretical_loss": 3.6209825091891954, + "tokens_seen": 1181810688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003242026078234704, + "loss": 1.4148, + "theoretical_loss": 3.620964131710717, + "tokens_seen": 1181876224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003241925777331996, + "loss": 1.4857, + "theoretical_loss": 3.620945755536571, + "tokens_seen": 1181941760 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032418254764292884, + "loss": 1.4479, + "theoretical_loss": 3.620927380666593, + "tokens_seen": 1182007296 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032417251755265797, + "loss": 1.4789, + "theoretical_loss": 3.620909007100618, + "tokens_seen": 1182072832 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003241624874623872, + "loss": 1.5361, + "theoretical_loss": 3.6208906348384815, + "tokens_seen": 1182138368 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032415245737211633, + "loss": 1.361, + "theoretical_loss": 3.6208722638800177, + "tokens_seen": 1182203904 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032414242728184557, + "loss": 1.3945, + "theoretical_loss": 3.620853894225063, + "tokens_seen": 1182269440 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032413239719157475, + "loss": 1.4932, + "theoretical_loss": 3.620835525873453, + "tokens_seen": 1182334976 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032412236710130393, + "loss": 1.3508, + "theoretical_loss": 3.6208171588250213, + "tokens_seen": 1182400512 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003241123370110331, + "loss": 1.4161, + "theoretical_loss": 3.6207987930796044, + "tokens_seen": 1182466048 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003241023069207623, + "loss": 1.5152, + "theoretical_loss": 3.620780428637037, + "tokens_seen": 1182531584 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032409227683049147, + "loss": 1.4494, + "theoretical_loss": 3.6207620654971553, + "tokens_seen": 1182597120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003240822467402207, + "loss": 1.369, + "theoretical_loss": 3.620743703659794, + "tokens_seen": 1182662656 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032407221664994984, + "loss": 1.3488, + "theoretical_loss": 3.620725343124789, + "tokens_seen": 1182728192 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032406218655967907, + "loss": 1.3576, + "theoretical_loss": 3.620706983891975, + "tokens_seen": 1182793728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003240521564694082, + "loss": 1.4193, + "theoretical_loss": 3.620688625961189, + "tokens_seen": 1182859264 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 433029, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4132041931152344, + "objective/train/theoretical_loss": 3.6206702693322654, + "objective/train/tokens_used": 1203384800, + "theoretical_loss": 3.6206702693322654, + "tokens_seen": 1182924800 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032404212637913743, + "loss": 1.3739, + "theoretical_loss": 3.6206702693322654, + "tokens_seen": 1182924800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003240320962888666, + "loss": 1.4216, + "theoretical_loss": 3.6206519140050393, + "tokens_seen": 1182990336 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003240220661985958, + "loss": 1.5019, + "theoretical_loss": 3.620633559979348, + "tokens_seen": 1183055872 + }, + { + "epoch": 0.36, + "learning_rate": 0.000324012036108325, + "loss": 1.4402, + "theoretical_loss": 3.6206152072550255, + "tokens_seen": 1183121408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003240020060180542, + "loss": 1.5077, + "theoretical_loss": 3.620596855831908, + "tokens_seen": 1183186944 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032399197592778334, + "loss": 1.3466, + "theoretical_loss": 3.620578505709831, + "tokens_seen": 1183252480 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239819458375126, + "loss": 1.562, + "theoretical_loss": 3.6205601568886308, + "tokens_seen": 1183318016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239719157472417, + "loss": 1.4736, + "theoretical_loss": 3.6205418093681425, + "tokens_seen": 1183383552 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032396188565697094, + "loss": 1.3572, + "theoretical_loss": 3.6205234631482024, + "tokens_seen": 1183449088 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239518555667001, + "loss": 1.4213, + "theoretical_loss": 3.6205051182286456, + "tokens_seen": 1183514624 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239418254764293, + "loss": 1.5854, + "theoretical_loss": 3.620486774609309, + "tokens_seen": 1183580160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239317953861585, + "loss": 1.385, + "theoretical_loss": 3.620468432290027, + "tokens_seen": 1183645696 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032392176529588766, + "loss": 1.3901, + "theoretical_loss": 3.620450091270637, + "tokens_seen": 1183711232 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032391173520561684, + "loss": 1.3314, + "theoretical_loss": 3.620431751550974, + "tokens_seen": 1183776768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003239017051153461, + "loss": 1.3544, + "theoretical_loss": 3.6204134131308736, + "tokens_seen": 1183842304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003238916750250752, + "loss": 1.5762, + "theoretical_loss": 3.620395076010173, + "tokens_seen": 1183907840 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032388164493480444, + "loss": 1.4111, + "theoretical_loss": 3.6203767401887075, + "tokens_seen": 1183973376 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032387161484453357, + "loss": 1.3585, + "theoretical_loss": 3.6203584056663134, + "tokens_seen": 1184038912 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003238615847542628, + "loss": 1.4694, + "theoretical_loss": 3.620340072442826, + "tokens_seen": 1184104448 + }, + { + "epoch": 0.36, + "learning_rate": 0.000323851554663992, + "loss": 1.4763, + "theoretical_loss": 3.620321740518082, + "tokens_seen": 1184169984 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032384152457372116, + "loss": 1.4091, + "theoretical_loss": 3.6203034098919176, + "tokens_seen": 1184235520 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032383149448345034, + "loss": 1.479, + "theoretical_loss": 3.6202850805641695, + "tokens_seen": 1184301056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003238214643931796, + "loss": 1.4195, + "theoretical_loss": 3.6202667525346723, + "tokens_seen": 1184366592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003238114343029087, + "loss": 1.4263, + "theoretical_loss": 3.6202484258032634, + "tokens_seen": 1184432128 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032380140421263794, + "loss": 1.4037, + "theoretical_loss": 3.6202301003697785, + "tokens_seen": 1184497664 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032379137412236707, + "loss": 1.3595, + "theoretical_loss": 3.620211776234055, + "tokens_seen": 1184563200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003237813440320963, + "loss": 1.4554, + "theoretical_loss": 3.6201934533959275, + "tokens_seen": 1184628736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003237713139418255, + "loss": 1.3774, + "theoretical_loss": 3.620175131855234, + "tokens_seen": 1184694272 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032376128385155467, + "loss": 1.4392, + "theoretical_loss": 3.6201568116118095, + "tokens_seen": 1184759808 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032375125376128385, + "loss": 1.4892, + "theoretical_loss": 3.6201384926654914, + "tokens_seen": 1184825344 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032374122367101303, + "loss": 1.3494, + "theoretical_loss": 3.620120175016115, + "tokens_seen": 1184890880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003237311935807422, + "loss": 1.4365, + "theoretical_loss": 3.620101858663518, + "tokens_seen": 1184956416 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032372116349047145, + "loss": 1.4836, + "theoretical_loss": 3.620083543607536, + "tokens_seen": 1185021952 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032371113340020057, + "loss": 1.4055, + "theoretical_loss": 3.620065229848006, + "tokens_seen": 1185087488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003237011033099298, + "loss": 1.4002, + "theoretical_loss": 3.620046917384764, + "tokens_seen": 1185153024 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032369107321965893, + "loss": 1.4031, + "theoretical_loss": 3.620028606217647, + "tokens_seen": 1185218560 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032368104312938817, + "loss": 1.4334, + "theoretical_loss": 3.620010296346492, + "tokens_seen": 1185284096 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032367101303911735, + "loss": 1.509, + "theoretical_loss": 3.619991987771135, + "tokens_seen": 1185349632 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032366098294884653, + "loss": 1.441, + "theoretical_loss": 3.619973680491413, + "tokens_seen": 1185415168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003236509528585757, + "loss": 1.3249, + "theoretical_loss": 3.619955374507162, + "tokens_seen": 1185480704 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032364092276830495, + "loss": 1.4566, + "theoretical_loss": 3.61993706981822, + "tokens_seen": 1185546240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003236308926780341, + "loss": 1.4371, + "theoretical_loss": 3.6199187664244223, + "tokens_seen": 1185611776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003236208625877633, + "loss": 1.338, + "theoretical_loss": 3.6199004643256067, + "tokens_seen": 1185677312 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032361083249749244, + "loss": 1.4424, + "theoretical_loss": 3.6198821635216096, + "tokens_seen": 1185742848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003236008024072217, + "loss": 1.5759, + "theoretical_loss": 3.6198638640122676, + "tokens_seen": 1185808384 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032359077231695085, + "loss": 1.4307, + "theoretical_loss": 3.6198455657974185, + "tokens_seen": 1185873920 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032358074222668004, + "loss": 1.4556, + "theoretical_loss": 3.6198272688768984, + "tokens_seen": 1185939456 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032357071213640927, + "loss": 1.4337, + "theoretical_loss": 3.619808973250544, + "tokens_seen": 1186004992 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003235606820461384, + "loss": 1.5102, + "theoretical_loss": 3.6197906789181933, + "tokens_seen": 1186070528 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032355065195586763, + "loss": 1.4613, + "theoretical_loss": 3.619772385879682, + "tokens_seen": 1186136064 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 434243, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5193933248519897, + "objective/train/theoretical_loss": 3.6197540941348487, + "objective/train/tokens_used": 1206661600, + "theoretical_loss": 3.6197540941348487, + "tokens_seen": 1186201600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003235406218655968, + "loss": 1.4754, + "theoretical_loss": 3.6197540941348487, + "tokens_seen": 1186201600 + }, + { + "epoch": 0.36, + "learning_rate": 0.000323530591775326, + "loss": 1.4218, + "theoretical_loss": 3.619735803683529, + "tokens_seen": 1186267136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003235205616850552, + "loss": 1.3541, + "theoretical_loss": 3.61971751452556, + "tokens_seen": 1186332672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003235105315947844, + "loss": 1.3802, + "theoretical_loss": 3.6196992266607797, + "tokens_seen": 1186398208 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032350050150451354, + "loss": 1.4975, + "theoretical_loss": 3.619680940089025, + "tokens_seen": 1186463744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234904714142428, + "loss": 1.4458, + "theoretical_loss": 3.6196626548101327, + "tokens_seen": 1186529280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234804413239719, + "loss": 1.5161, + "theoretical_loss": 3.6196443708239405, + "tokens_seen": 1186594816 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032347041123370114, + "loss": 1.5713, + "theoretical_loss": 3.6196260881302846, + "tokens_seen": 1186660352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234603811434303, + "loss": 1.434, + "theoretical_loss": 3.619607806729004, + "tokens_seen": 1186725888 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234503510531595, + "loss": 1.3948, + "theoretical_loss": 3.619589526619934, + "tokens_seen": 1186791424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234403209628887, + "loss": 1.3945, + "theoretical_loss": 3.6195712478029134, + "tokens_seen": 1186856960 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032343029087261786, + "loss": 1.6021, + "theoretical_loss": 3.6195529702777787, + "tokens_seen": 1186922496 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032342026078234704, + "loss": 1.4675, + "theoretical_loss": 3.6195346940443676, + "tokens_seen": 1186988032 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234102306920763, + "loss": 1.4979, + "theoretical_loss": 3.619516419102518, + "tokens_seen": 1187053568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003234002006018054, + "loss": 1.4345, + "theoretical_loss": 3.619498145452066, + "tokens_seen": 1187119104 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032339017051153464, + "loss": 1.3891, + "theoretical_loss": 3.6194798730928506, + "tokens_seen": 1187184640 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032338014042126377, + "loss": 1.4727, + "theoretical_loss": 3.6194616020247077, + "tokens_seen": 1187250176 + }, + { + "epoch": 0.36, + "learning_rate": 0.000323370110330993, + "loss": 1.4227, + "theoretical_loss": 3.619443332247476, + "tokens_seen": 1187315712 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003233600802407222, + "loss": 1.3271, + "theoretical_loss": 3.6194250637609926, + "tokens_seen": 1187381248 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032335005015045136, + "loss": 1.3833, + "theoretical_loss": 3.6194067965650953, + "tokens_seen": 1187446784 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032334002006018054, + "loss": 1.4875, + "theoretical_loss": 3.619388530659622, + "tokens_seen": 1187512320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003233299899699098, + "loss": 1.4351, + "theoretical_loss": 3.6193702660444096, + "tokens_seen": 1187577856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003233199598796389, + "loss": 1.5731, + "theoretical_loss": 3.619352002719296, + "tokens_seen": 1187643392 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032330992978936814, + "loss": 1.4799, + "theoretical_loss": 3.619333740684119, + "tokens_seen": 1187708928 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032329989969909727, + "loss": 1.5074, + "theoretical_loss": 3.6193154799387166, + "tokens_seen": 1187774464 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003232898696088265, + "loss": 1.4749, + "theoretical_loss": 3.6192972204829257, + "tokens_seen": 1187840000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003232798395185557, + "loss": 1.3904, + "theoretical_loss": 3.619278962316585, + "tokens_seen": 1187905536 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032326980942828487, + "loss": 1.5378, + "theoretical_loss": 3.619260705439532, + "tokens_seen": 1187971072 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032325977933801405, + "loss": 1.5134, + "theoretical_loss": 3.619242449851604, + "tokens_seen": 1188036608 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032324974924774323, + "loss": 1.3924, + "theoretical_loss": 3.61922419555264, + "tokens_seen": 1188102144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003232397191574724, + "loss": 1.3014, + "theoretical_loss": 3.619205942542477, + "tokens_seen": 1188167680 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032322968906720165, + "loss": 1.3702, + "theoretical_loss": 3.6191876908209526, + "tokens_seen": 1188233216 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032321965897693077, + "loss": 1.4487, + "theoretical_loss": 3.619169440387906, + "tokens_seen": 1188298752 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032320962888666, + "loss": 1.4328, + "theoretical_loss": 3.6191511912431746, + "tokens_seen": 1188364288 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032319959879638913, + "loss": 1.5039, + "theoretical_loss": 3.619132943386596, + "tokens_seen": 1188429824 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032318956870611837, + "loss": 1.5124, + "theoretical_loss": 3.6191146968180083, + "tokens_seen": 1188495360 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032317953861584755, + "loss": 1.5414, + "theoretical_loss": 3.61909645153725, + "tokens_seen": 1188560896 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032316950852557673, + "loss": 1.5637, + "theoretical_loss": 3.6190782075441597, + "tokens_seen": 1188626432 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003231594784353059, + "loss": 1.4071, + "theoretical_loss": 3.6190599648385744, + "tokens_seen": 1188691968 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032314944834503515, + "loss": 1.4181, + "theoretical_loss": 3.6190417234203327, + "tokens_seen": 1188757504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003231394182547643, + "loss": 1.2965, + "theoretical_loss": 3.6190234832892734, + "tokens_seen": 1188823040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003231293881644935, + "loss": 1.4363, + "theoretical_loss": 3.6190052444452334, + "tokens_seen": 1188888576 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032311935807422264, + "loss": 1.4499, + "theoretical_loss": 3.6189870068880525, + "tokens_seen": 1188954112 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003231093279839519, + "loss": 1.6235, + "theoretical_loss": 3.6189687706175677, + "tokens_seen": 1189019648 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032309929789368105, + "loss": 1.3079, + "theoretical_loss": 3.618950535633618, + "tokens_seen": 1189085184 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032308926780341024, + "loss": 1.366, + "theoretical_loss": 3.6189323019360415, + "tokens_seen": 1189150720 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003230792377131394, + "loss": 1.3332, + "theoretical_loss": 3.6189140695246764, + "tokens_seen": 1189216256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003230692076228686, + "loss": 1.4468, + "theoretical_loss": 3.6188958383993617, + "tokens_seen": 1189281792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003230591775325978, + "loss": 1.3545, + "theoretical_loss": 3.618877608559935, + "tokens_seen": 1189347328 + }, + { + "epoch": 0.36, + "learning_rate": 0.000323049147442327, + "loss": 1.4471, + "theoretical_loss": 3.6188593800062354, + "tokens_seen": 1189412864 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 435462, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4318820238113403, + "objective/train/theoretical_loss": 3.618841152738101, + "objective/train/tokens_used": 1209938400, + "theoretical_loss": 3.618841152738101, + "tokens_seen": 1189478400 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032303911735205614, + "loss": 1.4061, + "theoretical_loss": 3.618841152738101, + "tokens_seen": 1189478400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003230290872617854, + "loss": 1.5479, + "theoretical_loss": 3.618822926755371, + "tokens_seen": 1189543936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003230190571715145, + "loss": 1.5732, + "theoretical_loss": 3.618804702057883, + "tokens_seen": 1189609472 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032300902708124374, + "loss": 1.4953, + "theoretical_loss": 3.6187864786454758, + "tokens_seen": 1189675008 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229989969909729, + "loss": 1.5431, + "theoretical_loss": 3.6187682565179884, + "tokens_seen": 1189740544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229889669007021, + "loss": 1.432, + "theoretical_loss": 3.6187500356752595, + "tokens_seen": 1189806080 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229789368104313, + "loss": 1.3076, + "theoretical_loss": 3.618731816117127, + "tokens_seen": 1189871616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229689067201605, + "loss": 1.4139, + "theoretical_loss": 3.6187135978434304, + "tokens_seen": 1189937152 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032295887662988964, + "loss": 1.4707, + "theoretical_loss": 3.6186953808540085, + "tokens_seen": 1190002688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229488465396189, + "loss": 1.4577, + "theoretical_loss": 3.618677165148699, + "tokens_seen": 1190068224 + }, + { + "epoch": 0.36, + "learning_rate": 0.000322938816449348, + "loss": 1.4555, + "theoretical_loss": 3.6186589507273412, + "tokens_seen": 1190133760 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032292878635907724, + "loss": 1.4921, + "theoretical_loss": 3.618640737589775, + "tokens_seen": 1190199296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229187562688064, + "loss": 1.5013, + "theoretical_loss": 3.6186225257358373, + "tokens_seen": 1190264832 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003229087261785356, + "loss": 1.312, + "theoretical_loss": 3.618604315165368, + "tokens_seen": 1190330368 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003228986960882648, + "loss": 1.4228, + "theoretical_loss": 3.6185861058782063, + "tokens_seen": 1190395904 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032288866599799397, + "loss": 1.466, + "theoretical_loss": 3.6185678978741906, + "tokens_seen": 1190461440 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032287863590772315, + "loss": 1.4255, + "theoretical_loss": 3.6185496911531603, + "tokens_seen": 1190526976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003228686058174524, + "loss": 1.3853, + "theoretical_loss": 3.6185314857149535, + "tokens_seen": 1190592512 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003228585757271815, + "loss": 1.4563, + "theoretical_loss": 3.61851328155941, + "tokens_seen": 1190658048 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032284854563691074, + "loss": 1.5336, + "theoretical_loss": 3.6184950786863688, + "tokens_seen": 1190723584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003228385155466399, + "loss": 1.3904, + "theoretical_loss": 3.6184768770956683, + "tokens_seen": 1190789120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003228284854563691, + "loss": 1.3721, + "theoretical_loss": 3.6184586767871485, + "tokens_seen": 1190854656 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032281845536609834, + "loss": 1.3729, + "theoretical_loss": 3.6184404777606485, + "tokens_seen": 1190920192 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032280842527582747, + "loss": 1.5414, + "theoretical_loss": 3.6184222800160057, + "tokens_seen": 1190985728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003227983951855567, + "loss": 1.5524, + "theoretical_loss": 3.6184040835530618, + "tokens_seen": 1191051264 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003227883650952859, + "loss": 1.3748, + "theoretical_loss": 3.6183858883716544, + "tokens_seen": 1191116800 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032277833500501507, + "loss": 1.5792, + "theoretical_loss": 3.6183676944716234, + "tokens_seen": 1191182336 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032276830491474425, + "loss": 1.4138, + "theoretical_loss": 3.6183495018528076, + "tokens_seen": 1191247872 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032275827482447343, + "loss": 1.4008, + "theoretical_loss": 3.618331310515047, + "tokens_seen": 1191313408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003227482447342026, + "loss": 1.44, + "theoretical_loss": 3.61831312045818, + "tokens_seen": 1191378944 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032273821464393185, + "loss": 1.5282, + "theoretical_loss": 3.6182949316820467, + "tokens_seen": 1191444480 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032272818455366097, + "loss": 1.4035, + "theoretical_loss": 3.6182767441864856, + "tokens_seen": 1191510016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003227181544633902, + "loss": 1.3778, + "theoretical_loss": 3.6182585579713376, + "tokens_seen": 1191575552 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032270812437311933, + "loss": 1.4646, + "theoretical_loss": 3.618240373036441, + "tokens_seen": 1191641088 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032269809428284857, + "loss": 1.3903, + "theoretical_loss": 3.618222189381635, + "tokens_seen": 1191706624 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032268806419257775, + "loss": 1.2838, + "theoretical_loss": 3.6182040070067605, + "tokens_seen": 1191772160 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032267803410230693, + "loss": 1.3682, + "theoretical_loss": 3.618185825911656, + "tokens_seen": 1191837696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003226680040120361, + "loss": 1.2686, + "theoretical_loss": 3.6181676460961607, + "tokens_seen": 1191903232 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032265797392176535, + "loss": 1.5574, + "theoretical_loss": 3.618149467560115, + "tokens_seen": 1191968768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003226479438314945, + "loss": 1.4962, + "theoretical_loss": 3.6181312903033587, + "tokens_seen": 1192034304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003226379137412237, + "loss": 1.4117, + "theoretical_loss": 3.6181131143257304, + "tokens_seen": 1192099840 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032262788365095284, + "loss": 1.4013, + "theoretical_loss": 3.6180949396270705, + "tokens_seen": 1192165376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003226178535606821, + "loss": 1.4439, + "theoretical_loss": 3.618076766207219, + "tokens_seen": 1192230912 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032260782347041125, + "loss": 1.3617, + "theoretical_loss": 3.6180585940660146, + "tokens_seen": 1192296448 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032259779338014044, + "loss": 1.5262, + "theoretical_loss": 3.618040423203298, + "tokens_seen": 1192361984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225877632898696, + "loss": 1.3685, + "theoretical_loss": 3.618022253618909, + "tokens_seen": 1192427520 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225777331995988, + "loss": 1.3123, + "theoretical_loss": 3.6180040853126862, + "tokens_seen": 1192493056 + }, + { + "epoch": 0.36, + "learning_rate": 0.000322567703109328, + "loss": 1.3366, + "theoretical_loss": 3.6179859182844707, + "tokens_seen": 1192558592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225576730190572, + "loss": 1.6602, + "theoretical_loss": 3.6179677525341027, + "tokens_seen": 1192624128 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032254764292878634, + "loss": 1.4466, + "theoretical_loss": 3.6179495880614203, + "tokens_seen": 1192689664 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 436657, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5091347694396973, + "objective/train/theoretical_loss": 3.6179314248662653, + "objective/train/tokens_used": 1213215200, + "theoretical_loss": 3.6179314248662653, + "tokens_seen": 1192755200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225376128385156, + "loss": 1.4266, + "theoretical_loss": 3.6179314248662653, + "tokens_seen": 1192755200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225275827482447, + "loss": 1.4381, + "theoretical_loss": 3.617913262948477, + "tokens_seen": 1192820736 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032251755265797394, + "loss": 1.4607, + "theoretical_loss": 3.6178951023078945, + "tokens_seen": 1192886272 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003225075225677031, + "loss": 1.3491, + "theoretical_loss": 3.6178769429443594, + "tokens_seen": 1192951808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224974924774323, + "loss": 1.4811, + "theoretical_loss": 3.6178587848577104, + "tokens_seen": 1193017344 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224874623871615, + "loss": 1.3902, + "theoretical_loss": 3.6178406280477886, + "tokens_seen": 1193082880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224774322968907, + "loss": 1.3626, + "theoretical_loss": 3.6178224725144332, + "tokens_seen": 1193148416 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032246740220661984, + "loss": 1.4609, + "theoretical_loss": 3.6178043182574857, + "tokens_seen": 1193213952 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224573721163491, + "loss": 1.503, + "theoretical_loss": 3.617786165276785, + "tokens_seen": 1193279488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224473420260782, + "loss": 1.5118, + "theoretical_loss": 3.617768013572171, + "tokens_seen": 1193345024 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032243731193580744, + "loss": 1.3995, + "theoretical_loss": 3.6177498631434846, + "tokens_seen": 1193410560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224272818455366, + "loss": 1.41, + "theoretical_loss": 3.6177317139905667, + "tokens_seen": 1193476096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003224172517552658, + "loss": 1.4997, + "theoretical_loss": 3.617713566113257, + "tokens_seen": 1193541632 + }, + { + "epoch": 0.36, + "learning_rate": 0.000322407221664995, + "loss": 1.5077, + "theoretical_loss": 3.617695419511395, + "tokens_seen": 1193607168 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032239719157472417, + "loss": 1.4872, + "theoretical_loss": 3.617677274184822, + "tokens_seen": 1193672704 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032238716148445335, + "loss": 1.4725, + "theoretical_loss": 3.6176591301333785, + "tokens_seen": 1193738240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003223771313941826, + "loss": 1.4013, + "theoretical_loss": 3.6176409873569044, + "tokens_seen": 1193803776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003223671013039117, + "loss": 1.4904, + "theoretical_loss": 3.61762284585524, + "tokens_seen": 1193869312 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032235707121364095, + "loss": 1.54, + "theoretical_loss": 3.617604705628226, + "tokens_seen": 1193934848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003223470411233701, + "loss": 1.3887, + "theoretical_loss": 3.617586566675703, + "tokens_seen": 1194000384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003223370110330993, + "loss": 1.4222, + "theoretical_loss": 3.6175684289975116, + "tokens_seen": 1194065920 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003223269809428285, + "loss": 1.3895, + "theoretical_loss": 3.6175502925934917, + "tokens_seen": 1194131456 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032231695085255767, + "loss": 1.512, + "theoretical_loss": 3.6175321574634847, + "tokens_seen": 1194196992 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032230692076228685, + "loss": 1.4607, + "theoretical_loss": 3.6175140236073307, + "tokens_seen": 1194262528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003222968906720161, + "loss": 1.351, + "theoretical_loss": 3.6174958910248707, + "tokens_seen": 1194328064 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003222868605817452, + "loss": 1.4971, + "theoretical_loss": 3.6174777597159444, + "tokens_seen": 1194393600 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032227683049147445, + "loss": 1.4924, + "theoretical_loss": 3.6174596296803934, + "tokens_seen": 1194459136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003222668004012036, + "loss": 1.3887, + "theoretical_loss": 3.6174415009180585, + "tokens_seen": 1194524672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003222567703109328, + "loss": 1.4783, + "theoretical_loss": 3.61742337342878, + "tokens_seen": 1194590208 + }, + { + "epoch": 0.36, + "learning_rate": 0.000322246740220662, + "loss": 1.5129, + "theoretical_loss": 3.6174052472123983, + "tokens_seen": 1194655744 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032223671013039117, + "loss": 1.4943, + "theoretical_loss": 3.617387122268755, + "tokens_seen": 1194721280 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032222668004012035, + "loss": 1.4532, + "theoretical_loss": 3.6173689985976907, + "tokens_seen": 1194786816 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032221664994984953, + "loss": 1.591, + "theoretical_loss": 3.617350876199046, + "tokens_seen": 1194852352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003222066198595787, + "loss": 1.3639, + "theoretical_loss": 3.6173327550726624, + "tokens_seen": 1194917888 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032219658976930795, + "loss": 1.4526, + "theoretical_loss": 3.6173146352183796, + "tokens_seen": 1194983424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221865596790371, + "loss": 1.4266, + "theoretical_loss": 3.6172965166360402, + "tokens_seen": 1195048960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221765295887663, + "loss": 1.4585, + "theoretical_loss": 3.6172783993254836, + "tokens_seen": 1195114496 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221664994984955, + "loss": 1.2478, + "theoretical_loss": 3.6172602832865515, + "tokens_seen": 1195180032 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221564694082247, + "loss": 1.3029, + "theoretical_loss": 3.617242168519085, + "tokens_seen": 1195245568 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032214643931795386, + "loss": 1.5606, + "theoretical_loss": 3.617224055022925, + "tokens_seen": 1195311104 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032213640922768304, + "loss": 1.3564, + "theoretical_loss": 3.617205942797913, + "tokens_seen": 1195376640 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221263791374122, + "loss": 1.4377, + "theoretical_loss": 3.6171878318438893, + "tokens_seen": 1195442176 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032211634904714145, + "loss": 1.4731, + "theoretical_loss": 3.6171697221606953, + "tokens_seen": 1195507712 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003221063189568706, + "loss": 1.5479, + "theoretical_loss": 3.617151613748173, + "tokens_seen": 1195573248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220962888665998, + "loss": 1.3075, + "theoretical_loss": 3.6171335066061623, + "tokens_seen": 1195638784 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032208625877632894, + "loss": 1.4986, + "theoretical_loss": 3.6171154007345057, + "tokens_seen": 1195704320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220762286860582, + "loss": 1.4954, + "theoretical_loss": 3.617097296133043, + "tokens_seen": 1195769856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220661985957874, + "loss": 1.3696, + "theoretical_loss": 3.617079192801617, + "tokens_seen": 1195835392 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032205616850551654, + "loss": 1.4345, + "theoretical_loss": 3.6170610907400684, + "tokens_seen": 1195900928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220461384152458, + "loss": 1.365, + "theoretical_loss": 3.617042989948238, + "tokens_seen": 1195966464 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 437781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.500037670135498, + "objective/train/theoretical_loss": 3.617024890425968, + "objective/train/tokens_used": 1216492000, + "theoretical_loss": 3.617024890425968, + "tokens_seen": 1196032000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220361083249749, + "loss": 1.5281, + "theoretical_loss": 3.617024890425968, + "tokens_seen": 1196032000 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032202607823470414, + "loss": 1.4224, + "theoretical_loss": 3.617006792173099, + "tokens_seen": 1196097536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220160481444333, + "loss": 1.34, + "theoretical_loss": 3.616988695189473, + "tokens_seen": 1196163072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003220060180541625, + "loss": 1.4443, + "theoretical_loss": 3.6169705994749317, + "tokens_seen": 1196228608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219959879638917, + "loss": 1.4107, + "theoretical_loss": 3.6169525050293156, + "tokens_seen": 1196294144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219859578736209, + "loss": 1.5503, + "theoretical_loss": 3.6169344118524673, + "tokens_seen": 1196359680 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032197592778335004, + "loss": 1.407, + "theoretical_loss": 3.616916319944228, + "tokens_seen": 1196425216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219658976930793, + "loss": 1.4347, + "theoretical_loss": 3.6168982293044385, + "tokens_seen": 1196490752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219558676028084, + "loss": 1.4547, + "theoretical_loss": 3.6168801399329413, + "tokens_seen": 1196556288 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032194583751253764, + "loss": 1.5175, + "theoretical_loss": 3.6168620518295778, + "tokens_seen": 1196621824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219358074222668, + "loss": 1.3767, + "theoretical_loss": 3.6168439649941897, + "tokens_seen": 1196687360 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321925777331996, + "loss": 1.4385, + "theoretical_loss": 3.616825879426618, + "tokens_seen": 1196752896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003219157472417252, + "loss": 1.3554, + "theoretical_loss": 3.6168077951267055, + "tokens_seen": 1196818432 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032190571715145437, + "loss": 1.4965, + "theoretical_loss": 3.6167897120942936, + "tokens_seen": 1196883968 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032189568706118355, + "loss": 1.3706, + "theoretical_loss": 3.6167716303292234, + "tokens_seen": 1196949504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218856569709128, + "loss": 1.3828, + "theoretical_loss": 3.616753549831337, + "tokens_seen": 1197015040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218756268806419, + "loss": 1.5686, + "theoretical_loss": 3.616735470600477, + "tokens_seen": 1197080576 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032186559679037115, + "loss": 1.4191, + "theoretical_loss": 3.6167173926364846, + "tokens_seen": 1197146112 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218555667001003, + "loss": 1.5573, + "theoretical_loss": 3.6166993159392016, + "tokens_seen": 1197211648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218455366098295, + "loss": 1.3925, + "theoretical_loss": 3.61668124050847, + "tokens_seen": 1197277184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218355065195587, + "loss": 1.4167, + "theoretical_loss": 3.6166631663441318, + "tokens_seen": 1197342720 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032182547642928787, + "loss": 1.483, + "theoretical_loss": 3.616645093446029, + "tokens_seen": 1197408256 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032181544633901705, + "loss": 1.3882, + "theoretical_loss": 3.6166270218140033, + "tokens_seen": 1197473792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003218054162487463, + "loss": 1.3791, + "theoretical_loss": 3.6166089514478976, + "tokens_seen": 1197539328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003217953861584754, + "loss": 1.5064, + "theoretical_loss": 3.6165908823475528, + "tokens_seen": 1197604864 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032178535606820465, + "loss": 1.3621, + "theoretical_loss": 3.6165728145128115, + "tokens_seen": 1197670400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003217753259779338, + "loss": 1.5574, + "theoretical_loss": 3.616554747943516, + "tokens_seen": 1197735936 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321765295887663, + "loss": 1.4756, + "theoretical_loss": 3.616536682639508, + "tokens_seen": 1197801472 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003217552657973922, + "loss": 1.3807, + "theoretical_loss": 3.6165186186006304, + "tokens_seen": 1197867008 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032174523570712137, + "loss": 1.4876, + "theoretical_loss": 3.6165005558267245, + "tokens_seen": 1197932544 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032173520561685055, + "loss": 1.4627, + "theoretical_loss": 3.616482494317633, + "tokens_seen": 1197998080 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032172517552657974, + "loss": 1.3857, + "theoretical_loss": 3.6164644340731984, + "tokens_seen": 1198063616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003217151454363089, + "loss": 1.4171, + "theoretical_loss": 3.616446375093262, + "tokens_seen": 1198129152 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032170511534603815, + "loss": 1.3165, + "theoretical_loss": 3.6164283173776672, + "tokens_seen": 1198194688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216950852557673, + "loss": 1.4046, + "theoretical_loss": 3.616410260926256, + "tokens_seen": 1198260224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216850551654965, + "loss": 1.3697, + "theoretical_loss": 3.6163922057388707, + "tokens_seen": 1198325760 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216750250752257, + "loss": 1.4018, + "theoretical_loss": 3.6163741518153536, + "tokens_seen": 1198391296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216649949849549, + "loss": 1.3605, + "theoretical_loss": 3.6163560991555466, + "tokens_seen": 1198456832 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032165496489468406, + "loss": 1.3235, + "theoretical_loss": 3.6163380477592932, + "tokens_seen": 1198522368 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032164493480441324, + "loss": 1.4474, + "theoretical_loss": 3.6163199976264355, + "tokens_seen": 1198587904 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216349047141424, + "loss": 1.3494, + "theoretical_loss": 3.6163019487568158, + "tokens_seen": 1198653440 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032162487462387165, + "loss": 1.4373, + "theoretical_loss": 3.6162839011502763, + "tokens_seen": 1198718976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003216148445336008, + "loss": 1.2959, + "theoretical_loss": 3.6162658548066604, + "tokens_seen": 1198784512 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032160481444333, + "loss": 1.3715, + "theoretical_loss": 3.6162478097258104, + "tokens_seen": 1198850048 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032159478435305914, + "loss": 1.534, + "theoretical_loss": 3.6162297659075686, + "tokens_seen": 1198915584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215847542627884, + "loss": 1.3638, + "theoretical_loss": 3.616211723351778, + "tokens_seen": 1198981120 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032157472417251756, + "loss": 1.3713, + "theoretical_loss": 3.616193682058281, + "tokens_seen": 1199046656 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032156469408224674, + "loss": 1.3967, + "theoretical_loss": 3.6161756420269207, + "tokens_seen": 1199112192 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215546639919759, + "loss": 1.457, + "theoretical_loss": 3.6161576032575393, + "tokens_seen": 1199177728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215446339017051, + "loss": 1.2859, + "theoretical_loss": 3.6161395657499797, + "tokens_seen": 1199243264 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 438488, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.525244951248169, + "objective/train/theoretical_loss": 3.6161215295040847, + "objective/train/tokens_used": 1219768800, + "theoretical_loss": 3.6161215295040847, + "tokens_seen": 1199308800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215346038114343, + "loss": 1.468, + "theoretical_loss": 3.6161215295040847, + "tokens_seen": 1199308800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215245737211635, + "loss": 1.3842, + "theoretical_loss": 3.6161034945196975, + "tokens_seen": 1199374336 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032151454363089265, + "loss": 1.3922, + "theoretical_loss": 3.6160854607966604, + "tokens_seen": 1199439872 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003215045135406219, + "loss": 1.4616, + "theoretical_loss": 3.6160674283348166, + "tokens_seen": 1199505408 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032149448345035106, + "loss": 1.3296, + "theoretical_loss": 3.616049397134009, + "tokens_seen": 1199570944 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032148445336008024, + "loss": 1.4493, + "theoretical_loss": 3.6160313671940805, + "tokens_seen": 1199636480 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003214744232698094, + "loss": 1.4708, + "theoretical_loss": 3.6160133385148736, + "tokens_seen": 1199702016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003214643931795386, + "loss": 1.3694, + "theoretical_loss": 3.615995311096232, + "tokens_seen": 1199767552 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003214543630892678, + "loss": 1.2855, + "theoretical_loss": 3.6159772849379985, + "tokens_seen": 1199833088 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321444332998997, + "loss": 1.4834, + "theoretical_loss": 3.6159592600400163, + "tokens_seen": 1199898624 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032143430290872615, + "loss": 1.3363, + "theoretical_loss": 3.6159412364021275, + "tokens_seen": 1199964160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003214242728184554, + "loss": 1.436, + "theoretical_loss": 3.6159232140241766, + "tokens_seen": 1200029696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003214142427281845, + "loss": 1.4829, + "theoretical_loss": 3.6159051929060055, + "tokens_seen": 1200095232 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032140421263791375, + "loss": 1.5417, + "theoretical_loss": 3.615887173047458, + "tokens_seen": 1200160768 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032139418254764293, + "loss": 1.5076, + "theoretical_loss": 3.6158691544483776, + "tokens_seen": 1200226304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213841524573721, + "loss": 1.4147, + "theoretical_loss": 3.6158511371086064, + "tokens_seen": 1200291840 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213741223671013, + "loss": 1.4556, + "theoretical_loss": 3.615833121027989, + "tokens_seen": 1200357376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213640922768305, + "loss": 1.3912, + "theoretical_loss": 3.6158151062063673, + "tokens_seen": 1200422912 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032135406218655965, + "loss": 1.4251, + "theoretical_loss": 3.6157970926435854, + "tokens_seen": 1200488448 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213440320962889, + "loss": 1.3148, + "theoretical_loss": 3.615779080339487, + "tokens_seen": 1200553984 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321334002006018, + "loss": 1.3841, + "theoretical_loss": 3.615761069293914, + "tokens_seen": 1200619520 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032132397191574725, + "loss": 1.3146, + "theoretical_loss": 3.6157430595067113, + "tokens_seen": 1200685056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213139418254765, + "loss": 1.4279, + "theoretical_loss": 3.615725050977722, + "tokens_seen": 1200750592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003213039117352056, + "loss": 1.4135, + "theoretical_loss": 3.6157070437067884, + "tokens_seen": 1200816128 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032129388164493485, + "loss": 1.4063, + "theoretical_loss": 3.6156890376937554, + "tokens_seen": 1200881664 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321283851554664, + "loss": 1.5496, + "theoretical_loss": 3.615671032938466, + "tokens_seen": 1200947200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003212738214643932, + "loss": 1.5306, + "theoretical_loss": 3.615653029440763, + "tokens_seen": 1201012736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003212637913741224, + "loss": 1.3624, + "theoretical_loss": 3.615635027200491, + "tokens_seen": 1201078272 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003212537612838516, + "loss": 1.3509, + "theoretical_loss": 3.6156170262174934, + "tokens_seen": 1201143808 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032124373119358075, + "loss": 1.4739, + "theoretical_loss": 3.615599026491613, + "tokens_seen": 1201209344 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032123370110330994, + "loss": 1.3711, + "theoretical_loss": 3.615581028022694, + "tokens_seen": 1201274880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003212236710130391, + "loss": 1.4525, + "theoretical_loss": 3.61556303081058, + "tokens_seen": 1201340416 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032121364092276835, + "loss": 1.5649, + "theoretical_loss": 3.6155450348551152, + "tokens_seen": 1201405952 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003212036108324975, + "loss": 1.4124, + "theoretical_loss": 3.6155270401561426, + "tokens_seen": 1201471488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003211935807422267, + "loss": 1.4667, + "theoretical_loss": 3.615509046713506, + "tokens_seen": 1201537024 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003211835506519559, + "loss": 1.3953, + "theoretical_loss": 3.615491054527049, + "tokens_seen": 1201602560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003211735205616851, + "loss": 1.3811, + "theoretical_loss": 3.615473063596616, + "tokens_seen": 1201668096 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032116349047141426, + "loss": 1.4429, + "theoretical_loss": 3.6154550739220506, + "tokens_seen": 1201733632 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032115346038114344, + "loss": 1.4809, + "theoretical_loss": 3.6154370855031965, + "tokens_seen": 1201799168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003211434302908726, + "loss": 1.5116, + "theoretical_loss": 3.6154190983398977, + "tokens_seen": 1201864704 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032113340020060185, + "loss": 1.3034, + "theoretical_loss": 3.6154011124319982, + "tokens_seen": 1201930240 + }, + { + "epoch": 0.36, + "learning_rate": 0.000321123370110331, + "loss": 1.3944, + "theoretical_loss": 3.615383127779342, + "tokens_seen": 1201995776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003211133400200602, + "loss": 1.4313, + "theoretical_loss": 3.6153651443817725, + "tokens_seen": 1202061312 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032110330992978934, + "loss": 1.3752, + "theoretical_loss": 3.615347162239134, + "tokens_seen": 1202126848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210932798395186, + "loss": 1.4963, + "theoretical_loss": 3.615329181351271, + "tokens_seen": 1202192384 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032108324974924776, + "loss": 1.3513, + "theoretical_loss": 3.6153112017180264, + "tokens_seen": 1202257920 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032107321965897694, + "loss": 1.3347, + "theoretical_loss": 3.615293223339246, + "tokens_seen": 1202323456 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210631895687061, + "loss": 1.4025, + "theoretical_loss": 3.6152752462147726, + "tokens_seen": 1202388992 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210531594784353, + "loss": 1.4303, + "theoretical_loss": 3.6152572703444505, + "tokens_seen": 1202454528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210431293881645, + "loss": 1.3758, + "theoretical_loss": 3.6152392957281236, + "tokens_seen": 1202520064 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 439824, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.47721529006958, + "objective/train/theoretical_loss": 3.6152213223656373, + "objective/train/tokens_used": 1223045600, + "theoretical_loss": 3.6152213223656373, + "tokens_seen": 1202585600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210330992978937, + "loss": 1.3196, + "theoretical_loss": 3.6152213223656373, + "tokens_seen": 1202585600 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032102306920762285, + "loss": 1.3268, + "theoretical_loss": 3.6152033502568344, + "tokens_seen": 1202651136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003210130391173521, + "loss": 1.3118, + "theoretical_loss": 3.6151853794015603, + "tokens_seen": 1202716672 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032100300902708126, + "loss": 1.3903, + "theoretical_loss": 3.615167409799658, + "tokens_seen": 1202782208 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032099297893681044, + "loss": 1.4349, + "theoretical_loss": 3.6151494414509733, + "tokens_seen": 1202847744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003209829488465396, + "loss": 1.5022, + "theoretical_loss": 3.6151314743553495, + "tokens_seen": 1202913280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003209729187562688, + "loss": 1.406, + "theoretical_loss": 3.615113508512631, + "tokens_seen": 1202978816 + }, + { + "epoch": 0.36, + "learning_rate": 0.000320962888665998, + "loss": 1.386, + "theoretical_loss": 3.615095543922662, + "tokens_seen": 1203044352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003209528585757272, + "loss": 1.3988, + "theoretical_loss": 3.6150775805852877, + "tokens_seen": 1203109888 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032094282848545635, + "loss": 1.3578, + "theoretical_loss": 3.6150596185003527, + "tokens_seen": 1203175424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003209327983951856, + "loss": 1.3972, + "theoretical_loss": 3.6150416576677005, + "tokens_seen": 1203240960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003209227683049147, + "loss": 1.4246, + "theoretical_loss": 3.6150236980871755, + "tokens_seen": 1203306496 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032091273821464395, + "loss": 1.3525, + "theoretical_loss": 3.6150057397586233, + "tokens_seen": 1203372032 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032090270812437313, + "loss": 1.5107, + "theoretical_loss": 3.6149877826818875, + "tokens_seen": 1203437568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208926780341023, + "loss": 1.4333, + "theoretical_loss": 3.6149698268568136, + "tokens_seen": 1203503104 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208826479438315, + "loss": 1.4719, + "theoretical_loss": 3.614951872283245, + "tokens_seen": 1203568640 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208726178535607, + "loss": 1.3509, + "theoretical_loss": 3.614933918961027, + "tokens_seen": 1203634176 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032086258776328985, + "loss": 1.5135, + "theoretical_loss": 3.614915966890005, + "tokens_seen": 1203699712 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208525576730191, + "loss": 1.3891, + "theoretical_loss": 3.614898016070022, + "tokens_seen": 1203765248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208425275827482, + "loss": 1.3609, + "theoretical_loss": 3.614880066500924, + "tokens_seen": 1203830784 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032083249749247745, + "loss": 1.3756, + "theoretical_loss": 3.6148621181825558, + "tokens_seen": 1203896320 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032082246740220663, + "loss": 1.3356, + "theoretical_loss": 3.614844171114761, + "tokens_seen": 1203961856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003208124373119358, + "loss": 1.4502, + "theoretical_loss": 3.614826225297386, + "tokens_seen": 1204027392 + }, + { + "epoch": 0.36, + "learning_rate": 0.000320802407221665, + "loss": 1.335, + "theoretical_loss": 3.614808280730274, + "tokens_seen": 1204092928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003207923771313942, + "loss": 1.4411, + "theoretical_loss": 3.614790337413271, + "tokens_seen": 1204158464 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032078234704112336, + "loss": 1.4544, + "theoretical_loss": 3.6147723953462214, + "tokens_seen": 1204224000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003207723169508526, + "loss": 1.566, + "theoretical_loss": 3.61475445452897, + "tokens_seen": 1204289536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003207622868605817, + "loss": 1.4203, + "theoretical_loss": 3.614736514961362, + "tokens_seen": 1204355072 + }, + { + "epoch": 0.36, + "learning_rate": 0.00032075225677031095, + "loss": 1.4779, + "theoretical_loss": 3.6147185766432433, + "tokens_seen": 1204420608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0003207422266800401, + "loss": 1.4791, + "theoretical_loss": 3.6147006395744565, + "tokens_seen": 1204486144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003207321965897693, + "loss": 1.3362, + "theoretical_loss": 3.6146827037548492, + "tokens_seen": 1204551680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003207221664994985, + "loss": 1.3958, + "theoretical_loss": 3.6146647691842646, + "tokens_seen": 1204617216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003207121364092277, + "loss": 1.5641, + "theoretical_loss": 3.6146468358625485, + "tokens_seen": 1204682752 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032070210631895686, + "loss": 1.4184, + "theoretical_loss": 3.614628903789546, + "tokens_seen": 1204748288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206920762286861, + "loss": 1.282, + "theoretical_loss": 3.614610972965102, + "tokens_seen": 1204813824 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206820461384152, + "loss": 1.5361, + "theoretical_loss": 3.6145930433890623, + "tokens_seen": 1204879360 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032067201604814446, + "loss": 1.5139, + "theoretical_loss": 3.6145751150612715, + "tokens_seen": 1204944896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206619859578736, + "loss": 1.3897, + "theoretical_loss": 3.6145571879815748, + "tokens_seen": 1205010432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206519558676028, + "loss": 1.4105, + "theoretical_loss": 3.614539262149818, + "tokens_seen": 1205075968 + }, + { + "epoch": 0.37, + "learning_rate": 0.000320641925777332, + "loss": 1.5152, + "theoretical_loss": 3.6145213375658454, + "tokens_seen": 1205141504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206318956870612, + "loss": 1.3484, + "theoretical_loss": 3.614503414229503, + "tokens_seen": 1205207040 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032062186559679036, + "loss": 1.4282, + "theoretical_loss": 3.614485492140636, + "tokens_seen": 1205272576 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032061183550651954, + "loss": 1.4729, + "theoretical_loss": 3.61446757129909, + "tokens_seen": 1205338112 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003206018054162487, + "loss": 1.4465, + "theoretical_loss": 3.6144496517047098, + "tokens_seen": 1205403648 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032059177532597796, + "loss": 1.4428, + "theoretical_loss": 3.6144317333573412, + "tokens_seen": 1205469184 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205817452357071, + "loss": 1.3654, + "theoretical_loss": 3.6144138162568296, + "tokens_seen": 1205534720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205717151454363, + "loss": 1.4666, + "theoretical_loss": 3.6143959004030206, + "tokens_seen": 1205600256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205616850551655, + "loss": 1.3839, + "theoretical_loss": 3.614377985795759, + "tokens_seen": 1205665792 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205516549648947, + "loss": 1.3326, + "theoretical_loss": 3.6143600724348914, + "tokens_seen": 1205731328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205416248746239, + "loss": 1.497, + "theoretical_loss": 3.614342160320262, + "tokens_seen": 1205796864 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 441116, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4223437309265137, + "objective/train/theoretical_loss": 3.614324249451718, + "objective/train/tokens_used": 1226322400, + "theoretical_loss": 3.614324249451718, + "tokens_seen": 1205862400 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032053159478435305, + "loss": 1.4405, + "theoretical_loss": 3.614324249451718, + "tokens_seen": 1205862400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003205215646940823, + "loss": 1.5407, + "theoretical_loss": 3.6143063398291035, + "tokens_seen": 1205927936 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032051153460381146, + "loss": 1.5143, + "theoretical_loss": 3.6142884314522647, + "tokens_seen": 1205993472 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032050150451354064, + "loss": 1.4704, + "theoretical_loss": 3.6142705243210473, + "tokens_seen": 1206059008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204914744232698, + "loss": 1.3984, + "theoretical_loss": 3.614252618435297, + "tokens_seen": 1206124544 + }, + { + "epoch": 0.37, + "learning_rate": 0.000320481444332999, + "loss": 1.4498, + "theoretical_loss": 3.6142347137948603, + "tokens_seen": 1206190080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204714142427282, + "loss": 1.4868, + "theoretical_loss": 3.614216810399581, + "tokens_seen": 1206255616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204613841524574, + "loss": 1.4433, + "theoretical_loss": 3.614198908249307, + "tokens_seen": 1206321152 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032045135406218655, + "loss": 1.4467, + "theoretical_loss": 3.6141810073438823, + "tokens_seen": 1206386688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204413239719158, + "loss": 1.4638, + "theoretical_loss": 3.6141631076831544, + "tokens_seen": 1206452224 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204312938816449, + "loss": 1.3702, + "theoretical_loss": 3.6141452092669675, + "tokens_seen": 1206517760 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032042126379137415, + "loss": 1.272, + "theoretical_loss": 3.6141273120951682, + "tokens_seen": 1206583296 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032041123370110333, + "loss": 1.4153, + "theoretical_loss": 3.6141094161676026, + "tokens_seen": 1206648832 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003204012036108325, + "loss": 1.598, + "theoretical_loss": 3.614091521484117, + "tokens_seen": 1206714368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203911735205617, + "loss": 1.2857, + "theoretical_loss": 3.6140736280445562, + "tokens_seen": 1206779904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203811434302909, + "loss": 1.2981, + "theoretical_loss": 3.6140557358487673, + "tokens_seen": 1206845440 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032037111334002005, + "loss": 1.4052, + "theoretical_loss": 3.614037844896595, + "tokens_seen": 1206910976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203610832497493, + "loss": 1.4671, + "theoretical_loss": 3.614019955187887, + "tokens_seen": 1206976512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203510531594784, + "loss": 1.3816, + "theoretical_loss": 3.614002066722488, + "tokens_seen": 1207042048 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032034102306920765, + "loss": 1.5269, + "theoretical_loss": 3.6139841795002448, + "tokens_seen": 1207107584 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032033099297893683, + "loss": 1.4633, + "theoretical_loss": 3.6139662935210035, + "tokens_seen": 1207173120 + }, + { + "epoch": 0.37, + "learning_rate": 0.000320320962888666, + "loss": 1.4799, + "theoretical_loss": 3.6139484087846094, + "tokens_seen": 1207238656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203109327983952, + "loss": 1.4602, + "theoretical_loss": 3.61393052529091, + "tokens_seen": 1207304192 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003203009027081244, + "loss": 1.4823, + "theoretical_loss": 3.6139126430397512, + "tokens_seen": 1207369728 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032029087261785356, + "loss": 1.5107, + "theoretical_loss": 3.613894762030978, + "tokens_seen": 1207435264 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202808425275828, + "loss": 1.3855, + "theoretical_loss": 3.613876882264438, + "tokens_seen": 1207500800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202708124373119, + "loss": 1.4072, + "theoretical_loss": 3.613859003739977, + "tokens_seen": 1207566336 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032026078234704115, + "loss": 1.3897, + "theoretical_loss": 3.6138411264574413, + "tokens_seen": 1207631872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202507522567703, + "loss": 1.479, + "theoretical_loss": 3.6138232504166776, + "tokens_seen": 1207697408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202407221664995, + "loss": 1.5396, + "theoretical_loss": 3.613805375617532, + "tokens_seen": 1207762944 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202306920762287, + "loss": 1.5472, + "theoretical_loss": 3.61378750205985, + "tokens_seen": 1207828480 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202206619859579, + "loss": 1.4564, + "theoretical_loss": 3.613769629743479, + "tokens_seen": 1207894016 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032021063189568706, + "loss": 1.464, + "theoretical_loss": 3.6137517586682666, + "tokens_seen": 1207959552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003202006018054163, + "loss": 1.4278, + "theoretical_loss": 3.613733888834057, + "tokens_seen": 1208025088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003201905717151454, + "loss": 1.4686, + "theoretical_loss": 3.6137160202406977, + "tokens_seen": 1208090624 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032018054162487466, + "loss": 1.375, + "theoretical_loss": 3.6136981528880354, + "tokens_seen": 1208156160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003201705115346038, + "loss": 1.4978, + "theoretical_loss": 3.6136802867759164, + "tokens_seen": 1208221696 + }, + { + "epoch": 0.37, + "learning_rate": 0.000320160481444333, + "loss": 1.4581, + "theoretical_loss": 3.6136624219041877, + "tokens_seen": 1208287232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003201504513540622, + "loss": 1.2934, + "theoretical_loss": 3.613644558272695, + "tokens_seen": 1208352768 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003201404212637914, + "loss": 1.4587, + "theoretical_loss": 3.613626695881286, + "tokens_seen": 1208418304 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032013039117352056, + "loss": 1.5506, + "theoretical_loss": 3.6136088347298068, + "tokens_seen": 1208483840 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032012036108324974, + "loss": 1.4398, + "theoretical_loss": 3.613590974818104, + "tokens_seen": 1208549376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003201103309929789, + "loss": 1.4411, + "theoretical_loss": 3.613573116146025, + "tokens_seen": 1208614912 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032010030090270816, + "loss": 1.4391, + "theoretical_loss": 3.613555258713416, + "tokens_seen": 1208680448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003200902708124373, + "loss": 1.4812, + "theoretical_loss": 3.6135374025201235, + "tokens_seen": 1208745984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003200802407221665, + "loss": 1.3825, + "theoretical_loss": 3.6135195475659945, + "tokens_seen": 1208811520 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032007021063189565, + "loss": 1.2836, + "theoretical_loss": 3.6135016938508766, + "tokens_seen": 1208877056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003200601805416249, + "loss": 1.3073, + "theoretical_loss": 3.6134838413746158, + "tokens_seen": 1208942592 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032005015045135407, + "loss": 1.4137, + "theoretical_loss": 3.6134659901370587, + "tokens_seen": 1209008128 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032004012036108325, + "loss": 1.3786, + "theoretical_loss": 3.613448140138053, + "tokens_seen": 1209073664 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 442331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2707513570785522, + "objective/train/theoretical_loss": 3.613430291377446, + "objective/train/tokens_used": 1229599200, + "theoretical_loss": 3.613430291377446, + "tokens_seen": 1209139200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032003009027081243, + "loss": 1.3175, + "theoretical_loss": 3.613430291377446, + "tokens_seen": 1209139200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032002006018054166, + "loss": 1.3805, + "theoretical_loss": 3.613412443855083, + "tokens_seen": 1209204736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003200100300902708, + "loss": 1.4321, + "theoretical_loss": 3.613394597570813, + "tokens_seen": 1209270272 + }, + { + "epoch": 0.37, + "learning_rate": 0.00032, + "loss": 1.4429, + "theoretical_loss": 3.6133767525244815, + "tokens_seen": 1209335808 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031998996990972915, + "loss": 1.4095, + "theoretical_loss": 3.613358908715936, + "tokens_seen": 1209401344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003199799398194584, + "loss": 1.4755, + "theoretical_loss": 3.6133410661450243, + "tokens_seen": 1209466880 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031996990972918757, + "loss": 1.4625, + "theoretical_loss": 3.6133232248115927, + "tokens_seen": 1209532416 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031995987963891675, + "loss": 1.4072, + "theoretical_loss": 3.6133053847154883, + "tokens_seen": 1209597952 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031994984954864593, + "loss": 1.4196, + "theoretical_loss": 3.6132875458565588, + "tokens_seen": 1209663488 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003199398194583751, + "loss": 1.3377, + "theoretical_loss": 3.6132697082346508, + "tokens_seen": 1209729024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003199297893681043, + "loss": 1.3392, + "theoretical_loss": 3.613251871849612, + "tokens_seen": 1209794560 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031991975927783353, + "loss": 1.3853, + "theoretical_loss": 3.613234036701289, + "tokens_seen": 1209860096 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031990972918756266, + "loss": 1.468, + "theoretical_loss": 3.6132162027895296, + "tokens_seen": 1209925632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003198996990972919, + "loss": 1.3663, + "theoretical_loss": 3.613198370114181, + "tokens_seen": 1209991168 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319889669007021, + "loss": 1.4218, + "theoretical_loss": 3.613180538675091, + "tokens_seen": 1210056704 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031987963891675025, + "loss": 1.3738, + "theoretical_loss": 3.613162708472106, + "tokens_seen": 1210122240 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031986960882647943, + "loss": 1.4186, + "theoretical_loss": 3.6131448795050733, + "tokens_seen": 1210187776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003198595787362086, + "loss": 1.479, + "theoretical_loss": 3.6131270517738416, + "tokens_seen": 1210253312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003198495486459378, + "loss": 1.3079, + "theoretical_loss": 3.6131092252782575, + "tokens_seen": 1210318848 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031983951855566703, + "loss": 1.3811, + "theoretical_loss": 3.6130914000181686, + "tokens_seen": 1210384384 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031982948846539616, + "loss": 1.402, + "theoretical_loss": 3.613073575993422, + "tokens_seen": 1210449920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003198194583751254, + "loss": 1.3784, + "theoretical_loss": 3.6130557532038656, + "tokens_seen": 1210515456 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003198094282848546, + "loss": 1.4113, + "theoretical_loss": 3.6130379316493464, + "tokens_seen": 1210580992 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031979939819458376, + "loss": 1.3899, + "theoretical_loss": 3.613020111329713, + "tokens_seen": 1210646528 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319789368104313, + "loss": 1.5037, + "theoretical_loss": 3.613002292244812, + "tokens_seen": 1210712064 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197793380140421, + "loss": 1.3986, + "theoretical_loss": 3.6129844743944917, + "tokens_seen": 1210777600 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031976930792377135, + "loss": 1.4725, + "theoretical_loss": 3.6129666577785993, + "tokens_seen": 1210843136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197592778335005, + "loss": 1.4517, + "theoretical_loss": 3.6129488423969827, + "tokens_seen": 1210908672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197492477432297, + "loss": 1.392, + "theoretical_loss": 3.6129310282494895, + "tokens_seen": 1210974208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197392176529589, + "loss": 1.3307, + "theoretical_loss": 3.6129132153359675, + "tokens_seen": 1211039744 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197291875626881, + "loss": 1.4465, + "theoretical_loss": 3.6128954036562644, + "tokens_seen": 1211105280 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031971915747241726, + "loss": 1.4515, + "theoretical_loss": 3.6128775932102277, + "tokens_seen": 1211170816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003197091273821465, + "loss": 1.4457, + "theoretical_loss": 3.612859783997706, + "tokens_seen": 1211236352 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003196990972918756, + "loss": 1.3811, + "theoretical_loss": 3.6128419760185464, + "tokens_seen": 1211301888 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031968906720160486, + "loss": 1.4068, + "theoretical_loss": 3.612824169272597, + "tokens_seen": 1211367424 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319679037111334, + "loss": 1.514, + "theoretical_loss": 3.612806363759706, + "tokens_seen": 1211432960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003196690070210632, + "loss": 1.4875, + "theoretical_loss": 3.6127885594797204, + "tokens_seen": 1211498496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003196589769307924, + "loss": 1.4764, + "theoretical_loss": 3.612770756432489, + "tokens_seen": 1211564032 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003196489468405216, + "loss": 1.4037, + "theoretical_loss": 3.61275295461786, + "tokens_seen": 1211629568 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031963891675025076, + "loss": 1.4012, + "theoretical_loss": 3.61273515403568, + "tokens_seen": 1211695104 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031962888665997994, + "loss": 1.429, + "theoretical_loss": 3.612717354685798, + "tokens_seen": 1211760640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003196188565697091, + "loss": 1.4262, + "theoretical_loss": 3.612699556568063, + "tokens_seen": 1211826176 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031960882647943836, + "loss": 1.3754, + "theoretical_loss": 3.612681759682321, + "tokens_seen": 1211891712 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003195987963891675, + "loss": 1.4081, + "theoretical_loss": 3.612663964028422, + "tokens_seen": 1211957248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003195887662988967, + "loss": 1.3957, + "theoretical_loss": 3.6126461696062124, + "tokens_seen": 1212022784 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031957873620862585, + "loss": 1.5157, + "theoretical_loss": 3.6126283764155422, + "tokens_seen": 1212088320 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003195687061183551, + "loss": 1.4646, + "theoretical_loss": 3.612610584456258, + "tokens_seen": 1212153856 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031955867602808427, + "loss": 1.3218, + "theoretical_loss": 3.6125927937282087, + "tokens_seen": 1212219392 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031954864593781345, + "loss": 1.5498, + "theoretical_loss": 3.612575004231243, + "tokens_seen": 1212284928 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031953861584754263, + "loss": 1.4218, + "theoretical_loss": 3.612557215965208, + "tokens_seen": 1212350464 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 443456, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.436990737915039, + "objective/train/theoretical_loss": 3.6125394289299533, + "objective/train/tokens_used": 1232876000, + "theoretical_loss": 3.6125394289299533, + "tokens_seen": 1212416000 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031952858575727186, + "loss": 1.4625, + "theoretical_loss": 3.6125394289299533, + "tokens_seen": 1212416000 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319518555667001, + "loss": 1.3826, + "theoretical_loss": 3.6125216431253264, + "tokens_seen": 1212481536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003195085255767302, + "loss": 1.281, + "theoretical_loss": 3.6125038585511753, + "tokens_seen": 1212547072 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031949849548645935, + "loss": 1.3709, + "theoretical_loss": 3.6124860752073493, + "tokens_seen": 1212612608 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003194884653961886, + "loss": 1.2904, + "theoretical_loss": 3.6124682930936967, + "tokens_seen": 1212678144 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031947843530591777, + "loss": 1.3596, + "theoretical_loss": 3.6124505122100654, + "tokens_seen": 1212743680 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031946840521564695, + "loss": 1.4657, + "theoretical_loss": 3.6124327325563037, + "tokens_seen": 1212809216 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031945837512537613, + "loss": 1.4275, + "theoretical_loss": 3.6124149541322605, + "tokens_seen": 1212874752 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003194483450351053, + "loss": 1.5051, + "theoretical_loss": 3.612397176937785, + "tokens_seen": 1212940288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003194383149448345, + "loss": 1.4124, + "theoretical_loss": 3.612379400972724, + "tokens_seen": 1213005824 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031942828485456373, + "loss": 1.5059, + "theoretical_loss": 3.612361626236928, + "tokens_seen": 1213071360 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031941825476429286, + "loss": 1.3864, + "theoretical_loss": 3.612343852730244, + "tokens_seen": 1213136896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003194082246740221, + "loss": 1.4851, + "theoretical_loss": 3.6123260804525215, + "tokens_seen": 1213202432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003193981945837512, + "loss": 1.324, + "theoretical_loss": 3.612308309403609, + "tokens_seen": 1213267968 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031938816449348045, + "loss": 1.4159, + "theoretical_loss": 3.6122905395833547, + "tokens_seen": 1213333504 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031937813440320964, + "loss": 1.4738, + "theoretical_loss": 3.6122727709916083, + "tokens_seen": 1213399040 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003193681043129388, + "loss": 1.381, + "theoretical_loss": 3.612255003628217, + "tokens_seen": 1213464576 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319358074222668, + "loss": 1.4359, + "theoretical_loss": 3.6122372374930314, + "tokens_seen": 1213530112 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031934804413239723, + "loss": 1.389, + "theoretical_loss": 3.612219472585899, + "tokens_seen": 1213595648 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031933801404212636, + "loss": 1.3439, + "theoretical_loss": 3.6122017089066687, + "tokens_seen": 1213661184 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003193279839518556, + "loss": 1.4144, + "theoretical_loss": 3.6121839464551897, + "tokens_seen": 1213726720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003193179538615847, + "loss": 1.4427, + "theoretical_loss": 3.6121661852313105, + "tokens_seen": 1213792256 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031930792377131396, + "loss": 1.4103, + "theoretical_loss": 3.6121484252348806, + "tokens_seen": 1213857792 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031929789368104314, + "loss": 1.4666, + "theoretical_loss": 3.6121306664657484, + "tokens_seen": 1213923328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192878635907723, + "loss": 1.4672, + "theoretical_loss": 3.6121129089237627, + "tokens_seen": 1213988864 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192778335005015, + "loss": 1.4886, + "theoretical_loss": 3.612095152608773, + "tokens_seen": 1214054400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192678034102307, + "loss": 1.5308, + "theoretical_loss": 3.612077397520628, + "tokens_seen": 1214119936 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031925777331995986, + "loss": 1.399, + "theoretical_loss": 3.6120596436591765, + "tokens_seen": 1214185472 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192477432296891, + "loss": 1.4285, + "theoretical_loss": 3.6120418910242678, + "tokens_seen": 1214251008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192377131394182, + "loss": 1.3006, + "theoretical_loss": 3.612024139615751, + "tokens_seen": 1214316544 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031922768304914746, + "loss": 1.4443, + "theoretical_loss": 3.612006389433475, + "tokens_seen": 1214382080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192176529588766, + "loss": 1.4459, + "theoretical_loss": 3.611988640477289, + "tokens_seen": 1214447616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003192076228686058, + "loss": 1.4203, + "theoretical_loss": 3.6119708927470424, + "tokens_seen": 1214513152 + }, + { + "epoch": 0.37, + "learning_rate": 0.000319197592778335, + "loss": 1.454, + "theoretical_loss": 3.611953146242584, + "tokens_seen": 1214578688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191875626880642, + "loss": 1.4646, + "theoretical_loss": 3.6119354009637634, + "tokens_seen": 1214644224 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031917753259779337, + "loss": 1.506, + "theoretical_loss": 3.6119176569104297, + "tokens_seen": 1214709760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191675025075226, + "loss": 1.3414, + "theoretical_loss": 3.611899914082432, + "tokens_seen": 1214775296 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031915747241725173, + "loss": 1.4881, + "theoretical_loss": 3.6118821724796195, + "tokens_seen": 1214840832 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031914744232698096, + "loss": 1.4455, + "theoretical_loss": 3.6118644321018416, + "tokens_seen": 1214906368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191374122367101, + "loss": 1.4641, + "theoretical_loss": 3.611846692948948, + "tokens_seen": 1214971904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191273821464393, + "loss": 1.3342, + "theoretical_loss": 3.6118289550207874, + "tokens_seen": 1215037440 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191173520561685, + "loss": 1.2917, + "theoretical_loss": 3.61181121831721, + "tokens_seen": 1215102976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003191073219658977, + "loss": 1.3132, + "theoretical_loss": 3.6117934828380642, + "tokens_seen": 1215168512 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031909729187562687, + "loss": 1.5961, + "theoretical_loss": 3.6117757485832005, + "tokens_seen": 1215234048 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031908726178535605, + "loss": 1.4254, + "theoretical_loss": 3.611758015552468, + "tokens_seen": 1215299584 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031907723169508523, + "loss": 1.4815, + "theoretical_loss": 3.6117402837457155, + "tokens_seen": 1215365120 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031906720160481447, + "loss": 1.3477, + "theoretical_loss": 3.6117225531627932, + "tokens_seen": 1215430656 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031905717151454365, + "loss": 1.4014, + "theoretical_loss": 3.6117048238035507, + "tokens_seen": 1215496192 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031904714142427283, + "loss": 1.4456, + "theoretical_loss": 3.611687095667838, + "tokens_seen": 1215561728 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031903711133400206, + "loss": 1.4678, + "theoretical_loss": 3.6116693687555035, + "tokens_seen": 1215627264 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 444688, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3441038131713867, + "objective/train/theoretical_loss": 3.6116516430663976, + "objective/train/tokens_used": 1236152800, + "theoretical_loss": 3.6116516430663976, + "tokens_seen": 1215692800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003190270812437312, + "loss": 1.389, + "theoretical_loss": 3.6116516430663976, + "tokens_seen": 1215692800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003190170511534604, + "loss": 1.343, + "theoretical_loss": 3.6116339186003694, + "tokens_seen": 1215758336 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031900702106318955, + "loss": 1.4108, + "theoretical_loss": 3.6116161953572696, + "tokens_seen": 1215823872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003189969909729188, + "loss": 1.507, + "theoretical_loss": 3.6115984733369473, + "tokens_seen": 1215889408 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031898696088264797, + "loss": 1.4121, + "theoretical_loss": 3.6115807525392523, + "tokens_seen": 1215954944 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031897693079237715, + "loss": 1.4198, + "theoretical_loss": 3.611563032964034, + "tokens_seen": 1216020480 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031896690070210633, + "loss": 1.3463, + "theoretical_loss": 3.611545314611143, + "tokens_seen": 1216086016 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003189568706118355, + "loss": 1.4012, + "theoretical_loss": 3.6115275974804284, + "tokens_seen": 1216151552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003189468405215647, + "loss": 1.3791, + "theoretical_loss": 3.61150988157174, + "tokens_seen": 1216217088 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031893681043129393, + "loss": 1.5633, + "theoretical_loss": 3.6114921668849282, + "tokens_seen": 1216282624 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031892678034102306, + "loss": 1.3368, + "theoretical_loss": 3.6114744534198424, + "tokens_seen": 1216348160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003189167502507523, + "loss": 1.3624, + "theoretical_loss": 3.6114567411763328, + "tokens_seen": 1216413696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003189067201604814, + "loss": 1.372, + "theoretical_loss": 3.6114390301542496, + "tokens_seen": 1216479232 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031889669007021065, + "loss": 1.5161, + "theoretical_loss": 3.6114213203534424, + "tokens_seen": 1216544768 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031888665997993984, + "loss": 1.3982, + "theoretical_loss": 3.611403611773761, + "tokens_seen": 1216610304 + }, + { + "epoch": 0.37, + "learning_rate": 0.000318876629889669, + "loss": 1.4174, + "theoretical_loss": 3.611385904415056, + "tokens_seen": 1216675840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003188665997993982, + "loss": 1.4791, + "theoretical_loss": 3.611368198277177, + "tokens_seen": 1216741376 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031885656970912743, + "loss": 1.3581, + "theoretical_loss": 3.611350493359974, + "tokens_seen": 1216806912 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031884653961885656, + "loss": 1.3808, + "theoretical_loss": 3.6113327896632983, + "tokens_seen": 1216872448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003188365095285858, + "loss": 1.3357, + "theoretical_loss": 3.6113150871869983, + "tokens_seen": 1216937984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003188264794383149, + "loss": 1.3813, + "theoretical_loss": 3.611297385930925, + "tokens_seen": 1217003520 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031881644934804416, + "loss": 1.2852, + "theoretical_loss": 3.611279685894929, + "tokens_seen": 1217069056 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031880641925777334, + "loss": 1.3744, + "theoretical_loss": 3.6112619870788594, + "tokens_seen": 1217134592 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187963891675025, + "loss": 1.4028, + "theoretical_loss": 3.611244289482568, + "tokens_seen": 1217200128 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187863590772317, + "loss": 1.3985, + "theoretical_loss": 3.6112265931059033, + "tokens_seen": 1217265664 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187763289869609, + "loss": 1.4465, + "theoretical_loss": 3.6112088979487167, + "tokens_seen": 1217331200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031876629889669006, + "loss": 1.3351, + "theoretical_loss": 3.611191204010858, + "tokens_seen": 1217396736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187562688064193, + "loss": 1.277, + "theoretical_loss": 3.6111735112921783, + "tokens_seen": 1217462272 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187462387161484, + "loss": 1.4217, + "theoretical_loss": 3.6111558197925273, + "tokens_seen": 1217527808 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031873620862587766, + "loss": 1.434, + "theoretical_loss": 3.6111381295117555, + "tokens_seen": 1217593344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187261785356068, + "loss": 1.5178, + "theoretical_loss": 3.6111204404497137, + "tokens_seen": 1217658880 + }, + { + "epoch": 0.37, + "learning_rate": 0.000318716148445336, + "loss": 1.397, + "theoretical_loss": 3.611102752606252, + "tokens_seen": 1217724416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003187061183550652, + "loss": 1.3856, + "theoretical_loss": 3.6110850659812206, + "tokens_seen": 1217789952 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186960882647944, + "loss": 1.4574, + "theoretical_loss": 3.611067380574471, + "tokens_seen": 1217855488 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031868605817452357, + "loss": 1.5184, + "theoretical_loss": 3.6110496963858525, + "tokens_seen": 1217921024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186760280842528, + "loss": 1.3706, + "theoretical_loss": 3.6110320134152163, + "tokens_seen": 1217986560 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031866599799398193, + "loss": 1.4069, + "theoretical_loss": 3.611014331662413, + "tokens_seen": 1218052096 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031865596790371116, + "loss": 1.5431, + "theoretical_loss": 3.610996651127293, + "tokens_seen": 1218117632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186459378134403, + "loss": 1.3603, + "theoretical_loss": 3.6109789718097076, + "tokens_seen": 1218183168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186359077231695, + "loss": 1.4312, + "theoretical_loss": 3.610961293709506, + "tokens_seen": 1218248704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186258776328987, + "loss": 1.3767, + "theoretical_loss": 3.6109436168265407, + "tokens_seen": 1218314240 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003186158475426279, + "loss": 1.4609, + "theoretical_loss": 3.6109259411606613, + "tokens_seen": 1218379776 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031860581745235707, + "loss": 1.4427, + "theoretical_loss": 3.6109082667117187, + "tokens_seen": 1218445312 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031859578736208625, + "loss": 1.4079, + "theoretical_loss": 3.6108905934795636, + "tokens_seen": 1218510848 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031858575727181543, + "loss": 1.3861, + "theoretical_loss": 3.610872921464047, + "tokens_seen": 1218576384 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031857572718154467, + "loss": 1.3814, + "theoretical_loss": 3.61085525066502, + "tokens_seen": 1218641920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003185656970912738, + "loss": 1.3986, + "theoretical_loss": 3.610837581082333, + "tokens_seen": 1218707456 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031855566700100303, + "loss": 1.4981, + "theoretical_loss": 3.610819912715837, + "tokens_seen": 1218772992 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003185456369107322, + "loss": 1.4228, + "theoretical_loss": 3.6108022455653828, + "tokens_seen": 1218838528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003185356068204614, + "loss": 1.2992, + "theoretical_loss": 3.6107845796308213, + "tokens_seen": 1218904064 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 445814, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.536476731300354, + "objective/train/theoretical_loss": 3.6107669149120034, + "objective/train/tokens_used": 1239429600, + "theoretical_loss": 3.6107669149120034, + "tokens_seen": 1218969600 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031852557673019057, + "loss": 1.4639, + "theoretical_loss": 3.6107669149120034, + "tokens_seen": 1218969600 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031851554663991975, + "loss": 1.359, + "theoretical_loss": 3.6107492514087802, + "tokens_seen": 1219035136 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031850551654964893, + "loss": 1.3333, + "theoretical_loss": 3.610731589121003, + "tokens_seen": 1219100672 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031849548645937817, + "loss": 1.3427, + "theoretical_loss": 3.610713928048523, + "tokens_seen": 1219166208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003184854563691073, + "loss": 1.5057, + "theoretical_loss": 3.61069626819119, + "tokens_seen": 1219231744 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031847542627883653, + "loss": 1.3283, + "theoretical_loss": 3.6106786095488563, + "tokens_seen": 1219297280 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031846539618856566, + "loss": 1.3551, + "theoretical_loss": 3.610660952121373, + "tokens_seen": 1219362816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003184553660982949, + "loss": 1.4644, + "theoretical_loss": 3.61064329590859, + "tokens_seen": 1219428352 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003184453360080241, + "loss": 1.4258, + "theoretical_loss": 3.61062564091036, + "tokens_seen": 1219493888 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031843530591775326, + "loss": 1.4393, + "theoretical_loss": 3.6106079871265333, + "tokens_seen": 1219559424 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031842527582748244, + "loss": 1.4107, + "theoretical_loss": 3.6105903345569614, + "tokens_seen": 1219624960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003184152457372116, + "loss": 1.3816, + "theoretical_loss": 3.6105726832014953, + "tokens_seen": 1219690496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003184052156469408, + "loss": 1.346, + "theoretical_loss": 3.6105550330599865, + "tokens_seen": 1219756032 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031839518555667004, + "loss": 1.4168, + "theoretical_loss": 3.6105373841322863, + "tokens_seen": 1219821568 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031838515546639916, + "loss": 1.3924, + "theoretical_loss": 3.610519736418246, + "tokens_seen": 1219887104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003183751253761284, + "loss": 1.4633, + "theoretical_loss": 3.6105020899177167, + "tokens_seen": 1219952640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003183650952858576, + "loss": 1.3074, + "theoretical_loss": 3.6104844446305497, + "tokens_seen": 1220018176 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031835506519558676, + "loss": 1.4259, + "theoretical_loss": 3.6104668005565976, + "tokens_seen": 1220083712 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031834503510531594, + "loss": 1.3824, + "theoretical_loss": 3.61044915769571, + "tokens_seen": 1220149248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003183350050150451, + "loss": 1.4188, + "theoretical_loss": 3.61043151604774, + "tokens_seen": 1220214784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003183249749247743, + "loss": 1.3478, + "theoretical_loss": 3.6104138756125375, + "tokens_seen": 1220280320 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031831494483450354, + "loss": 1.3921, + "theoretical_loss": 3.6103962363899553, + "tokens_seen": 1220345856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003183049147442327, + "loss": 1.4737, + "theoretical_loss": 3.610378598379844, + "tokens_seen": 1220411392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182948846539619, + "loss": 1.3294, + "theoretical_loss": 3.610360961582056, + "tokens_seen": 1220476928 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182848545636911, + "loss": 1.4476, + "theoretical_loss": 3.6103433259964426, + "tokens_seen": 1220542464 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031827482447342026, + "loss": 1.3328, + "theoretical_loss": 3.610325691622855, + "tokens_seen": 1220608000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182647943831495, + "loss": 1.4051, + "theoretical_loss": 3.610308058461145, + "tokens_seen": 1220673536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182547642928786, + "loss": 1.5538, + "theoretical_loss": 3.610290426511165, + "tokens_seen": 1220739072 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031824473420260786, + "loss": 1.4738, + "theoretical_loss": 3.6102727957727656, + "tokens_seen": 1220804608 + }, + { + "epoch": 0.37, + "learning_rate": 0.000318234704112337, + "loss": 1.3595, + "theoretical_loss": 3.610255166245799, + "tokens_seen": 1220870144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182246740220662, + "loss": 1.41, + "theoretical_loss": 3.610237537930117, + "tokens_seen": 1220935680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182146439317954, + "loss": 1.383, + "theoretical_loss": 3.6102199108255713, + "tokens_seen": 1221001216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003182046138415246, + "loss": 1.3477, + "theoretical_loss": 3.6102022849320132, + "tokens_seen": 1221066752 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031819458375125377, + "loss": 1.2898, + "theoretical_loss": 3.6101846602492955, + "tokens_seen": 1221132288 + }, + { + "epoch": 0.37, + "learning_rate": 0.000318184553660983, + "loss": 1.4267, + "theoretical_loss": 3.61016703677727, + "tokens_seen": 1221197824 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031817452357071213, + "loss": 1.4291, + "theoretical_loss": 3.6101494145157873, + "tokens_seen": 1221263360 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031816449348044136, + "loss": 1.4768, + "theoretical_loss": 3.6101317934647, + "tokens_seen": 1221328896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003181544633901705, + "loss": 1.5066, + "theoretical_loss": 3.610114173623861, + "tokens_seen": 1221394432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003181444332998997, + "loss": 1.5523, + "theoretical_loss": 3.6100965549931203, + "tokens_seen": 1221459968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003181344032096289, + "loss": 1.3409, + "theoretical_loss": 3.610078937572332, + "tokens_seen": 1221525504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003181243731193581, + "loss": 1.4623, + "theoretical_loss": 3.6100613213613464, + "tokens_seen": 1221591040 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031811434302908727, + "loss": 1.4677, + "theoretical_loss": 3.610043706360016, + "tokens_seen": 1221656576 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031810431293881645, + "loss": 1.3454, + "theoretical_loss": 3.6100260925681935, + "tokens_seen": 1221722112 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031809428284854563, + "loss": 1.3883, + "theoretical_loss": 3.61000847998573, + "tokens_seen": 1221787648 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031808425275827487, + "loss": 1.4725, + "theoretical_loss": 3.6099908686124786, + "tokens_seen": 1221853184 + }, + { + "epoch": 0.37, + "learning_rate": 0.000318074222668004, + "loss": 1.4451, + "theoretical_loss": 3.6099732584482904, + "tokens_seen": 1221918720 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031806419257773323, + "loss": 1.4038, + "theoretical_loss": 3.6099556494930187, + "tokens_seen": 1221984256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003180541624874624, + "loss": 1.383, + "theoretical_loss": 3.6099380417465143, + "tokens_seen": 1222049792 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003180441323971916, + "loss": 1.5994, + "theoretical_loss": 3.6099204352086307, + "tokens_seen": 1222115328 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031803410230692077, + "loss": 1.4165, + "theoretical_loss": 3.6099028298792195, + "tokens_seen": 1222180864 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 446976, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.392150640487671, + "objective/train/theoretical_loss": 3.6098852257581333, + "objective/train/tokens_used": 1242706400, + "theoretical_loss": 3.6098852257581333, + "tokens_seen": 1222246400 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031802407221664995, + "loss": 1.3782, + "theoretical_loss": 3.6098852257581333, + "tokens_seen": 1222246400 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031801404212637913, + "loss": 1.4168, + "theoretical_loss": 3.609867622845224, + "tokens_seen": 1222311936 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031800401203610837, + "loss": 1.3183, + "theoretical_loss": 3.6098500211403435, + "tokens_seen": 1222377472 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003179939819458375, + "loss": 1.359, + "theoretical_loss": 3.609832420643345, + "tokens_seen": 1222443008 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031798395185556673, + "loss": 1.4087, + "theoretical_loss": 3.609814821354081, + "tokens_seen": 1222508544 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031797392176529586, + "loss": 1.483, + "theoretical_loss": 3.6097972232724027, + "tokens_seen": 1222574080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003179638916750251, + "loss": 1.4091, + "theoretical_loss": 3.6097796263981636, + "tokens_seen": 1222639616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003179538615847543, + "loss": 1.3999, + "theoretical_loss": 3.6097620307312157, + "tokens_seen": 1222705152 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031794383149448346, + "loss": 1.3039, + "theoretical_loss": 3.609744436271412, + "tokens_seen": 1222770688 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031793380140421264, + "loss": 1.4834, + "theoretical_loss": 3.609726843018604, + "tokens_seen": 1222836224 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003179237713139418, + "loss": 1.4056, + "theoretical_loss": 3.609709250972645, + "tokens_seen": 1222901760 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317913741223671, + "loss": 1.337, + "theoretical_loss": 3.6096916601333873, + "tokens_seen": 1222967296 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031790371113340024, + "loss": 1.3769, + "theoretical_loss": 3.6096740705006836, + "tokens_seen": 1223032832 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031789368104312936, + "loss": 1.3783, + "theoretical_loss": 3.609656482074386, + "tokens_seen": 1223098368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003178836509528586, + "loss": 1.4552, + "theoretical_loss": 3.609638894854348, + "tokens_seen": 1223163904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003178736208625878, + "loss": 1.4315, + "theoretical_loss": 3.6096213088404214, + "tokens_seen": 1223229440 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031786359077231696, + "loss": 1.3547, + "theoretical_loss": 3.6096037240324597, + "tokens_seen": 1223294976 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031785356068204614, + "loss": 1.2759, + "theoretical_loss": 3.6095861404303147, + "tokens_seen": 1223360512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003178435305917753, + "loss": 1.3506, + "theoretical_loss": 3.609568558033839, + "tokens_seen": 1223426048 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003178335005015045, + "loss": 1.4752, + "theoretical_loss": 3.609550976842887, + "tokens_seen": 1223491584 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031782347041123374, + "loss": 1.3984, + "theoretical_loss": 3.6095333968573096, + "tokens_seen": 1223557120 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031781344032096287, + "loss": 1.2953, + "theoretical_loss": 3.609515818076961, + "tokens_seen": 1223622656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003178034102306921, + "loss": 1.358, + "theoretical_loss": 3.609498240501693, + "tokens_seen": 1223688192 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031779338014042123, + "loss": 1.4917, + "theoretical_loss": 3.6094806641313593, + "tokens_seen": 1223753728 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031778335005015046, + "loss": 1.4376, + "theoretical_loss": 3.609463088965812, + "tokens_seen": 1223819264 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031777331995987964, + "loss": 1.4069, + "theoretical_loss": 3.609445515004904, + "tokens_seen": 1223884800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003177632898696088, + "loss": 1.3674, + "theoretical_loss": 3.609427942248489, + "tokens_seen": 1223950336 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317753259779338, + "loss": 1.4811, + "theoretical_loss": 3.60941037069642, + "tokens_seen": 1224015872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003177432296890672, + "loss": 1.4051, + "theoretical_loss": 3.609392800348549, + "tokens_seen": 1224081408 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031773319959879637, + "loss": 1.4078, + "theoretical_loss": 3.609375231204729, + "tokens_seen": 1224146944 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003177231695085256, + "loss": 1.4529, + "theoretical_loss": 3.6093576632648148, + "tokens_seen": 1224212480 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031771313941825473, + "loss": 1.4264, + "theoretical_loss": 3.6093400965286575, + "tokens_seen": 1224278016 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031770310932798397, + "loss": 1.3401, + "theoretical_loss": 3.609322530996111, + "tokens_seen": 1224343552 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031769307923771315, + "loss": 1.3863, + "theoretical_loss": 3.609304966667028, + "tokens_seen": 1224409088 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031768304914744233, + "loss": 1.4373, + "theoretical_loss": 3.6092874035412628, + "tokens_seen": 1224474624 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003176730190571715, + "loss": 1.3796, + "theoretical_loss": 3.6092698416186675, + "tokens_seen": 1224540160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003176629889669007, + "loss": 1.4408, + "theoretical_loss": 3.609252280899095, + "tokens_seen": 1224605696 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031765295887662987, + "loss": 1.4963, + "theoretical_loss": 3.6092347213823994, + "tokens_seen": 1224671232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003176429287863591, + "loss": 1.4992, + "theoretical_loss": 3.6092171630684335, + "tokens_seen": 1224736768 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031763289869608823, + "loss": 1.4652, + "theoretical_loss": 3.609199605957051, + "tokens_seen": 1224802304 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031762286860581747, + "loss": 1.3796, + "theoretical_loss": 3.6091820500481044, + "tokens_seen": 1224867840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003176128385155466, + "loss": 1.3239, + "theoretical_loss": 3.6091644953414477, + "tokens_seen": 1224933376 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031760280842527583, + "loss": 1.5245, + "theoretical_loss": 3.6091469418369337, + "tokens_seen": 1224998912 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317592778335005, + "loss": 1.3386, + "theoretical_loss": 3.6091293895344165, + "tokens_seen": 1225064448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003175827482447342, + "loss": 1.4011, + "theoretical_loss": 3.609111838433749, + "tokens_seen": 1225129984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003175727181544634, + "loss": 1.4718, + "theoretical_loss": 3.6090942885347848, + "tokens_seen": 1225195520 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003175626880641926, + "loss": 1.381, + "theoretical_loss": 3.609076739837377, + "tokens_seen": 1225261056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003175526579739218, + "loss": 1.4799, + "theoretical_loss": 3.6090591923413795, + "tokens_seen": 1225326592 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031754262788365097, + "loss": 1.4677, + "theoretical_loss": 3.609041646046646, + "tokens_seen": 1225392128 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031753259779338015, + "loss": 1.5, + "theoretical_loss": 3.609024100953029, + "tokens_seen": 1225457664 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 448161, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.417766809463501, + "objective/train/theoretical_loss": 3.6090065570603826, + "objective/train/tokens_used": 1245983200, + "theoretical_loss": 3.6090065570603826, + "tokens_seen": 1225523200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031752256770310933, + "loss": 1.3795, + "theoretical_loss": 3.6090065570603826, + "tokens_seen": 1225523200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031751253761283857, + "loss": 1.3524, + "theoretical_loss": 3.608989014368561, + "tokens_seen": 1225588736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003175025075225677, + "loss": 1.3901, + "theoretical_loss": 3.608971472877417, + "tokens_seen": 1225654272 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031749247743229693, + "loss": 1.4119, + "theoretical_loss": 3.6089539325868047, + "tokens_seen": 1225719808 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031748244734202606, + "loss": 1.3927, + "theoretical_loss": 3.6089363934965775, + "tokens_seen": 1225785344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003174724172517553, + "loss": 1.4044, + "theoretical_loss": 3.608918855606589, + "tokens_seen": 1225850880 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003174623871614845, + "loss": 1.4839, + "theoretical_loss": 3.6089013189166934, + "tokens_seen": 1225916416 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031745235707121366, + "loss": 1.388, + "theoretical_loss": 3.608883783426744, + "tokens_seen": 1225981952 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031744232698094284, + "loss": 1.4752, + "theoretical_loss": 3.6088662491365944, + "tokens_seen": 1226047488 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317432296890672, + "loss": 1.4264, + "theoretical_loss": 3.608848716046099, + "tokens_seen": 1226113024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003174222668004012, + "loss": 1.4613, + "theoretical_loss": 3.6088311841551106, + "tokens_seen": 1226178560 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031741223671013044, + "loss": 1.3709, + "theoretical_loss": 3.608813653463484, + "tokens_seen": 1226244096 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031740220661985956, + "loss": 1.4611, + "theoretical_loss": 3.6087961239710724, + "tokens_seen": 1226309632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003173921765295888, + "loss": 1.4037, + "theoretical_loss": 3.6087785956777303, + "tokens_seen": 1226375168 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317382146439318, + "loss": 1.4345, + "theoretical_loss": 3.608761068583312, + "tokens_seen": 1226440704 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031737211634904716, + "loss": 1.3836, + "theoretical_loss": 3.6087435426876695, + "tokens_seen": 1226506240 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031736208625877634, + "loss": 1.3086, + "theoretical_loss": 3.6087260179906586, + "tokens_seen": 1226571776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003173520561685055, + "loss": 1.3665, + "theoretical_loss": 3.6087084944921326, + "tokens_seen": 1226637312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003173420260782347, + "loss": 1.343, + "theoretical_loss": 3.608690972191946, + "tokens_seen": 1226702848 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031733199598796394, + "loss": 1.4976, + "theoretical_loss": 3.608673451089952, + "tokens_seen": 1226768384 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031732196589769307, + "loss": 1.4233, + "theoretical_loss": 3.6086559311860054, + "tokens_seen": 1226833920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003173119358074223, + "loss": 1.3187, + "theoretical_loss": 3.6086384124799595, + "tokens_seen": 1226899456 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031730190571715143, + "loss": 1.3734, + "theoretical_loss": 3.6086208949716694, + "tokens_seen": 1226964992 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031729187562688066, + "loss": 1.3228, + "theoretical_loss": 3.608603378660989, + "tokens_seen": 1227030528 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031728184553660984, + "loss": 1.3362, + "theoretical_loss": 3.6085858635477717, + "tokens_seen": 1227096064 + }, + { + "epoch": 0.37, + "learning_rate": 0.000317271815446339, + "loss": 1.5182, + "theoretical_loss": 3.608568349631872, + "tokens_seen": 1227161600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003172617853560682, + "loss": 1.4528, + "theoretical_loss": 3.6085508369131447, + "tokens_seen": 1227227136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003172517552657974, + "loss": 1.3788, + "theoretical_loss": 3.608533325391444, + "tokens_seen": 1227292672 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031724172517552657, + "loss": 1.3315, + "theoretical_loss": 3.6085158150666232, + "tokens_seen": 1227358208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003172316950852558, + "loss": 1.3456, + "theoretical_loss": 3.6084983059385376, + "tokens_seen": 1227423744 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031722166499498493, + "loss": 1.4639, + "theoretical_loss": 3.608480798007041, + "tokens_seen": 1227489280 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031721163490471417, + "loss": 1.3562, + "theoretical_loss": 3.6084632912719874, + "tokens_seen": 1227554816 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031720160481444335, + "loss": 1.3864, + "theoretical_loss": 3.6084457857332324, + "tokens_seen": 1227620352 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031719157472417253, + "loss": 1.3374, + "theoretical_loss": 3.60842828139063, + "tokens_seen": 1227685888 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003171815446339017, + "loss": 1.4187, + "theoretical_loss": 3.608410778244033, + "tokens_seen": 1227751424 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003171715145436309, + "loss": 1.3301, + "theoretical_loss": 3.608393276293298, + "tokens_seen": 1227816960 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031716148445336007, + "loss": 1.3928, + "theoretical_loss": 3.6083757755382786, + "tokens_seen": 1227882496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003171514543630893, + "loss": 1.4255, + "theoretical_loss": 3.608358275978829, + "tokens_seen": 1227948032 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031714142427281843, + "loss": 1.5466, + "theoretical_loss": 3.608340777614804, + "tokens_seen": 1228013568 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031713139418254767, + "loss": 1.419, + "theoretical_loss": 3.6083232804460583, + "tokens_seen": 1228079104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003171213640922768, + "loss": 1.4515, + "theoretical_loss": 3.6083057844724467, + "tokens_seen": 1228144640 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031711133400200603, + "loss": 1.384, + "theoretical_loss": 3.6082882896938226, + "tokens_seen": 1228210176 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003171013039117352, + "loss": 1.3396, + "theoretical_loss": 3.6082707961100424, + "tokens_seen": 1228275712 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170912738214644, + "loss": 1.3321, + "theoretical_loss": 3.608253303720959, + "tokens_seen": 1228341248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170812437311936, + "loss": 1.4886, + "theoretical_loss": 3.6082358125264284, + "tokens_seen": 1228406784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170712136409228, + "loss": 1.5448, + "theoretical_loss": 3.6082183225263043, + "tokens_seen": 1228472320 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031706118355065194, + "loss": 1.4422, + "theoretical_loss": 3.608200833720443, + "tokens_seen": 1228537856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170511534603812, + "loss": 1.3346, + "theoretical_loss": 3.608183346108697, + "tokens_seen": 1228603392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170411233701103, + "loss": 1.3996, + "theoretical_loss": 3.6081658596909225, + "tokens_seen": 1228668928 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031703109327983954, + "loss": 1.4909, + "theoretical_loss": 3.6081483744669747, + "tokens_seen": 1228734464 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 449484, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5102331638336182, + "objective/train/theoretical_loss": 3.6081308904367067, + "objective/train/tokens_used": 1249260000, + "theoretical_loss": 3.6081308904367067, + "tokens_seen": 1228800000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170210631895687, + "loss": 1.5039, + "theoretical_loss": 3.6081308904367067, + "tokens_seen": 1228800000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170110330992979, + "loss": 1.4274, + "theoretical_loss": 3.6081134075999755, + "tokens_seen": 1228865536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003170010030090271, + "loss": 1.4319, + "theoretical_loss": 3.608095925956634, + "tokens_seen": 1228931072 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031699097291875626, + "loss": 1.3902, + "theoretical_loss": 3.6080784455065387, + "tokens_seen": 1228996608 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031698094282848544, + "loss": 1.4206, + "theoretical_loss": 3.6080609662495435, + "tokens_seen": 1229062144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003169709127382147, + "loss": 1.4066, + "theoretical_loss": 3.608043488185504, + "tokens_seen": 1229127680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003169608826479438, + "loss": 1.3855, + "theoretical_loss": 3.6080260113142746, + "tokens_seen": 1229193216 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031695085255767304, + "loss": 1.3704, + "theoretical_loss": 3.6080085356357103, + "tokens_seen": 1229258752 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031694082246740217, + "loss": 1.4079, + "theoretical_loss": 3.607991061149667, + "tokens_seen": 1229324288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003169307923771314, + "loss": 1.3624, + "theoretical_loss": 3.607973587855999, + "tokens_seen": 1229389824 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003169207622868606, + "loss": 1.3281, + "theoretical_loss": 3.607956115754561, + "tokens_seen": 1229455360 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031691073219658976, + "loss": 1.4048, + "theoretical_loss": 3.607938644845209, + "tokens_seen": 1229520896 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031690070210631894, + "loss": 1.4323, + "theoretical_loss": 3.6079211751277978, + "tokens_seen": 1229586432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003168906720160482, + "loss": 1.4699, + "theoretical_loss": 3.607903706602183, + "tokens_seen": 1229651968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003168806419257773, + "loss": 1.4755, + "theoretical_loss": 3.6078862392682187, + "tokens_seen": 1229717504 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031687061183550654, + "loss": 1.4445, + "theoretical_loss": 3.607868773125761, + "tokens_seen": 1229783040 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031686058174523567, + "loss": 1.5016, + "theoretical_loss": 3.607851308174665, + "tokens_seen": 1229848576 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003168505516549649, + "loss": 1.4814, + "theoretical_loss": 3.6078338444147855, + "tokens_seen": 1229914112 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003168405215646941, + "loss": 1.4757, + "theoretical_loss": 3.6078163818459785, + "tokens_seen": 1229979648 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031683049147442327, + "loss": 1.4087, + "theoretical_loss": 3.607798920468099, + "tokens_seen": 1230045184 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031682046138415245, + "loss": 1.4644, + "theoretical_loss": 3.607781460281002, + "tokens_seen": 1230110720 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031681043129388163, + "loss": 1.3783, + "theoretical_loss": 3.6077640012845427, + "tokens_seen": 1230176256 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031680040120361086, + "loss": 1.5007, + "theoretical_loss": 3.6077465434785774, + "tokens_seen": 1230241792 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031679037111334004, + "loss": 1.2929, + "theoretical_loss": 3.607729086862961, + "tokens_seen": 1230307328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003167803410230692, + "loss": 1.4509, + "theoretical_loss": 3.6077116314375486, + "tokens_seen": 1230372864 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003167703109327984, + "loss": 1.427, + "theoretical_loss": 3.6076941772021964, + "tokens_seen": 1230438400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003167602808425276, + "loss": 1.4652, + "theoretical_loss": 3.6076767241567596, + "tokens_seen": 1230503936 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031675025075225677, + "loss": 1.4143, + "theoretical_loss": 3.6076592723010936, + "tokens_seen": 1230569472 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316740220661986, + "loss": 1.493, + "theoretical_loss": 3.6076418216350534, + "tokens_seen": 1230635008 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031673019057171513, + "loss": 1.3455, + "theoretical_loss": 3.6076243721584955, + "tokens_seen": 1230700544 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031672016048144437, + "loss": 1.5358, + "theoretical_loss": 3.607606923871275, + "tokens_seen": 1230766080 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031671013039117355, + "loss": 1.3925, + "theoretical_loss": 3.6075894767732475, + "tokens_seen": 1230831616 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031670010030090273, + "loss": 1.4245, + "theoretical_loss": 3.607572030864269, + "tokens_seen": 1230897152 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003166900702106319, + "loss": 1.4289, + "theoretical_loss": 3.607554586144195, + "tokens_seen": 1230962688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003166800401203611, + "loss": 1.4442, + "theoretical_loss": 3.607537142612881, + "tokens_seen": 1231028224 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031667001003009027, + "loss": 1.4676, + "theoretical_loss": 3.6075197002701826, + "tokens_seen": 1231093760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003166599799398195, + "loss": 1.4135, + "theoretical_loss": 3.6075022591159556, + "tokens_seen": 1231159296 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031664994984954863, + "loss": 1.4319, + "theoretical_loss": 3.6074848191500566, + "tokens_seen": 1231224832 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031663991975927787, + "loss": 1.3112, + "theoretical_loss": 3.60746738037234, + "tokens_seen": 1231290368 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316629889669007, + "loss": 1.3342, + "theoretical_loss": 3.6074499427826625, + "tokens_seen": 1231355904 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031661985957873623, + "loss": 1.4626, + "theoretical_loss": 3.6074325063808796, + "tokens_seen": 1231421440 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003166098294884654, + "loss": 1.422, + "theoretical_loss": 3.607415071166848, + "tokens_seen": 1231486976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165997993981946, + "loss": 1.4992, + "theoretical_loss": 3.6073976371404224, + "tokens_seen": 1231552512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165897693079238, + "loss": 1.3198, + "theoretical_loss": 3.6073802043014593, + "tokens_seen": 1231618048 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316579739217653, + "loss": 1.4746, + "theoretical_loss": 3.6073627726498145, + "tokens_seen": 1231683584 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031656970912738214, + "loss": 1.4216, + "theoretical_loss": 3.6073453421853445, + "tokens_seen": 1231749120 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165596790371114, + "loss": 1.458, + "theoretical_loss": 3.607327912907904, + "tokens_seen": 1231814656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165496489468405, + "loss": 1.3661, + "theoretical_loss": 3.6073104848173503, + "tokens_seen": 1231880192 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031653961885656974, + "loss": 1.4016, + "theoretical_loss": 3.607293057913539, + "tokens_seen": 1231945728 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165295887662989, + "loss": 1.3641, + "theoretical_loss": 3.607275632196326, + "tokens_seen": 1232011264 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 450737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4954384565353394, + "objective/train/theoretical_loss": 3.6072582076655673, + "objective/train/tokens_used": 1252536800, + "theoretical_loss": 3.6072582076655673, + "tokens_seen": 1232076800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165195586760281, + "loss": 1.4593, + "theoretical_loss": 3.6072582076655673, + "tokens_seen": 1232076800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003165095285857573, + "loss": 1.3433, + "theoretical_loss": 3.60724078432112, + "tokens_seen": 1232142336 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031649949849548646, + "loss": 1.3409, + "theoretical_loss": 3.6072233621628387, + "tokens_seen": 1232207872 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031648946840521564, + "loss": 1.4601, + "theoretical_loss": 3.607205941190581, + "tokens_seen": 1232273408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003164794383149449, + "loss": 1.336, + "theoretical_loss": 3.607188521404202, + "tokens_seen": 1232338944 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316469408224674, + "loss": 1.4134, + "theoretical_loss": 3.6071711028035587, + "tokens_seen": 1232404480 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031645937813440324, + "loss": 1.4492, + "theoretical_loss": 3.6071536853885062, + "tokens_seen": 1232470016 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031644934804413237, + "loss": 1.3993, + "theoretical_loss": 3.6071362691589024, + "tokens_seen": 1232535552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003164393179538616, + "loss": 1.4207, + "theoretical_loss": 3.607118854114603, + "tokens_seen": 1232601088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003164292878635908, + "loss": 1.4032, + "theoretical_loss": 3.6071014402554633, + "tokens_seen": 1232666624 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031641925777331996, + "loss": 1.4491, + "theoretical_loss": 3.6070840275813403, + "tokens_seen": 1232732160 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031640922768304914, + "loss": 1.4534, + "theoretical_loss": 3.607066616092091, + "tokens_seen": 1232797696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003163991975927784, + "loss": 1.4903, + "theoretical_loss": 3.607049205787571, + "tokens_seen": 1232863232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003163891675025075, + "loss": 1.5437, + "theoretical_loss": 3.6070317966676377, + "tokens_seen": 1232928768 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031637913741223674, + "loss": 1.35, + "theoretical_loss": 3.6070143887321455, + "tokens_seen": 1232994304 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031636910732196587, + "loss": 1.3514, + "theoretical_loss": 3.6069969819809535, + "tokens_seen": 1233059840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003163590772316951, + "loss": 1.3766, + "theoretical_loss": 3.606979576413916, + "tokens_seen": 1233125376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003163490471414243, + "loss": 1.5054, + "theoretical_loss": 3.606962172030891, + "tokens_seen": 1233190912 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031633901705115347, + "loss": 1.4502, + "theoretical_loss": 3.606944768831734, + "tokens_seen": 1233256448 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031632898696088265, + "loss": 1.4126, + "theoretical_loss": 3.6069273668163024, + "tokens_seen": 1233321984 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031631895687061183, + "loss": 1.4454, + "theoretical_loss": 3.6069099659844523, + "tokens_seen": 1233387520 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316308926780341, + "loss": 1.4448, + "theoretical_loss": 3.6068925663360405, + "tokens_seen": 1233453056 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031629889669007024, + "loss": 1.3671, + "theoretical_loss": 3.606875167870923, + "tokens_seen": 1233518592 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031628886659979937, + "loss": 1.3836, + "theoretical_loss": 3.6068577705889577, + "tokens_seen": 1233584128 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003162788365095286, + "loss": 1.5713, + "theoretical_loss": 3.606840374490001, + "tokens_seen": 1233649664 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031626880641925773, + "loss": 1.3707, + "theoretical_loss": 3.606822979573908, + "tokens_seen": 1233715200 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031625877632898697, + "loss": 1.3042, + "theoretical_loss": 3.6068055858405375, + "tokens_seen": 1233780736 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031624874623871615, + "loss": 1.4601, + "theoretical_loss": 3.606788193289745, + "tokens_seen": 1233846272 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031623871614844533, + "loss": 1.392, + "theoretical_loss": 3.606770801921388, + "tokens_seen": 1233911808 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003162286860581745, + "loss": 1.431, + "theoretical_loss": 3.6067534117353235, + "tokens_seen": 1233977344 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031621865596790375, + "loss": 1.4173, + "theoretical_loss": 3.6067360227314067, + "tokens_seen": 1234042880 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003162086258776329, + "loss": 1.4593, + "theoretical_loss": 3.606718634909497, + "tokens_seen": 1234108416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003161985957873621, + "loss": 1.4402, + "theoretical_loss": 3.606701248269449, + "tokens_seen": 1234173952 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031618856569709124, + "loss": 1.3516, + "theoretical_loss": 3.6066838628111215, + "tokens_seen": 1234239488 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031617853560682047, + "loss": 1.3791, + "theoretical_loss": 3.6066664785343696, + "tokens_seen": 1234305024 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031616850551654965, + "loss": 1.397, + "theoretical_loss": 3.6066490954390513, + "tokens_seen": 1234370560 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031615847542627883, + "loss": 1.3393, + "theoretical_loss": 3.606631713525024, + "tokens_seen": 1234436096 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316148445336008, + "loss": 1.2784, + "theoretical_loss": 3.6066143327921436, + "tokens_seen": 1234501632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003161384152457372, + "loss": 1.4275, + "theoretical_loss": 3.606596953240268, + "tokens_seen": 1234567168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003161283851554664, + "loss": 1.409, + "theoretical_loss": 3.6065795748692544, + "tokens_seen": 1234632704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003161183550651956, + "loss": 1.4367, + "theoretical_loss": 3.606562197678959, + "tokens_seen": 1234698240 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031610832497492474, + "loss": 1.3715, + "theoretical_loss": 3.6065448216692393, + "tokens_seen": 1234763776 + }, + { + "epoch": 0.37, + "learning_rate": 0.000316098294884654, + "loss": 1.4477, + "theoretical_loss": 3.606527446839953, + "tokens_seen": 1234829312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160882647943831, + "loss": 1.4274, + "theoretical_loss": 3.6065100731909565, + "tokens_seen": 1234894848 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031607823470411234, + "loss": 1.3998, + "theoretical_loss": 3.6064927007221073, + "tokens_seen": 1234960384 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160682046138415, + "loss": 1.486, + "theoretical_loss": 3.6064753294332625, + "tokens_seen": 1235025920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160581745235707, + "loss": 1.4234, + "theoretical_loss": 3.6064579593242794, + "tokens_seen": 1235091456 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031604814443329994, + "loss": 1.369, + "theoretical_loss": 3.6064405903950156, + "tokens_seen": 1235156992 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160381143430291, + "loss": 1.4592, + "theoretical_loss": 3.6064232226453283, + "tokens_seen": 1235222528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160280842527583, + "loss": 1.4647, + "theoretical_loss": 3.606405856075074, + "tokens_seen": 1235288064 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 451987, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.448754072189331, + "objective/train/theoretical_loss": 3.606388490684111, + "objective/train/tokens_used": 1255813600, + "theoretical_loss": 3.606388490684111, + "tokens_seen": 1235353600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003160180541624875, + "loss": 1.4935, + "theoretical_loss": 3.606388490684111, + "tokens_seen": 1235353600 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031600802407221666, + "loss": 1.3969, + "theoretical_loss": 3.606371126472296, + "tokens_seen": 1235419136 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031599799398194584, + "loss": 1.4593, + "theoretical_loss": 3.606353763439487, + "tokens_seen": 1235484672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003159879638916751, + "loss": 1.3323, + "theoretical_loss": 3.6063364015855415, + "tokens_seen": 1235550208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003159779338014042, + "loss": 1.5284, + "theoretical_loss": 3.606319040910316, + "tokens_seen": 1235615744 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031596790371113344, + "loss": 1.412, + "theoretical_loss": 3.606301681413669, + "tokens_seen": 1235681280 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031595787362086257, + "loss": 1.4196, + "theoretical_loss": 3.6062843230954567, + "tokens_seen": 1235746816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003159478435305918, + "loss": 1.4465, + "theoretical_loss": 3.606266965955538, + "tokens_seen": 1235812352 + }, + { + "epoch": 0.37, + "learning_rate": 0.000315937813440321, + "loss": 1.5134, + "theoretical_loss": 3.6062496099937698, + "tokens_seen": 1235877888 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031592778335005016, + "loss": 1.3691, + "theoretical_loss": 3.60623225521001, + "tokens_seen": 1235943424 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031591775325977934, + "loss": 1.4249, + "theoretical_loss": 3.6062149016041154, + "tokens_seen": 1236008960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003159077231695086, + "loss": 1.4183, + "theoretical_loss": 3.606197549175944, + "tokens_seen": 1236074496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003158976930792377, + "loss": 1.3055, + "theoretical_loss": 3.606180197925354, + "tokens_seen": 1236140032 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031588766298896694, + "loss": 1.4175, + "theoretical_loss": 3.6061628478522025, + "tokens_seen": 1236205568 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031587763289869607, + "loss": 1.4368, + "theoretical_loss": 3.606145498956347, + "tokens_seen": 1236271104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003158676028084253, + "loss": 1.5197, + "theoretical_loss": 3.6061281512376464, + "tokens_seen": 1236336640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003158575727181545, + "loss": 1.4964, + "theoretical_loss": 3.6061108046959567, + "tokens_seen": 1236402176 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031584754262788367, + "loss": 1.4148, + "theoretical_loss": 3.6060934593311367, + "tokens_seen": 1236467712 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031583751253761285, + "loss": 1.4094, + "theoretical_loss": 3.6060761151430443, + "tokens_seen": 1236533248 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031582748244734203, + "loss": 1.3632, + "theoretical_loss": 3.6060587721315365, + "tokens_seen": 1236598784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003158174523570712, + "loss": 1.4055, + "theoretical_loss": 3.6060414302964716, + "tokens_seen": 1236664320 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031580742226680044, + "loss": 1.5018, + "theoretical_loss": 3.6060240896377085, + "tokens_seen": 1236729856 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031579739217652957, + "loss": 1.5174, + "theoretical_loss": 3.606006750155103, + "tokens_seen": 1236795392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003157873620862588, + "loss": 1.3954, + "theoretical_loss": 3.6059894118485145, + "tokens_seen": 1236860928 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031577733199598793, + "loss": 1.3836, + "theoretical_loss": 3.6059720747178003, + "tokens_seen": 1236926464 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031576730190571717, + "loss": 1.3474, + "theoretical_loss": 3.605954738762819, + "tokens_seen": 1236992000 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031575727181544635, + "loss": 1.4606, + "theoretical_loss": 3.605937403983428, + "tokens_seen": 1237057536 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031574724172517553, + "loss": 1.3134, + "theoretical_loss": 3.6059200703794856, + "tokens_seen": 1237123072 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003157372116349047, + "loss": 1.4068, + "theoretical_loss": 3.6059027379508493, + "tokens_seen": 1237188608 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031572718154463395, + "loss": 1.3159, + "theoretical_loss": 3.605885406697378, + "tokens_seen": 1237254144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003157171514543631, + "loss": 1.3871, + "theoretical_loss": 3.605868076618929, + "tokens_seen": 1237319680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0003157071213640923, + "loss": 1.4358, + "theoretical_loss": 3.605850747715361, + "tokens_seen": 1237385216 + }, + { + "epoch": 0.37, + "learning_rate": 0.00031569709127382144, + "loss": 1.3674, + "theoretical_loss": 3.6058334199865314, + "tokens_seen": 1237450752 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031568706118355067, + "loss": 1.4193, + "theoretical_loss": 3.605816093432299, + "tokens_seen": 1237516288 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031567703109327985, + "loss": 1.3252, + "theoretical_loss": 3.605798768052522, + "tokens_seen": 1237581824 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031566700100300903, + "loss": 1.3596, + "theoretical_loss": 3.6057814438470586, + "tokens_seen": 1237647360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003156569709127382, + "loss": 1.3645, + "theoretical_loss": 3.6057641208157665, + "tokens_seen": 1237712896 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003156469408224674, + "loss": 1.4549, + "theoretical_loss": 3.6057467989585046, + "tokens_seen": 1237778432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003156369107321966, + "loss": 1.3983, + "theoretical_loss": 3.605729478275131, + "tokens_seen": 1237843968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003156268806419258, + "loss": 1.4218, + "theoretical_loss": 3.6057121587655034, + "tokens_seen": 1237909504 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031561685055165494, + "loss": 1.4047, + "theoretical_loss": 3.6056948404294804, + "tokens_seen": 1237975040 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003156068204613842, + "loss": 1.3758, + "theoretical_loss": 3.6056775232669214, + "tokens_seen": 1238040576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155967903711133, + "loss": 1.3911, + "theoretical_loss": 3.6056602072776833, + "tokens_seen": 1238106112 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031558676028084254, + "loss": 1.4789, + "theoretical_loss": 3.6056428924616255, + "tokens_seen": 1238171648 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155767301905717, + "loss": 1.2943, + "theoretical_loss": 3.605625578818606, + "tokens_seen": 1238237184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155667001003009, + "loss": 1.4126, + "theoretical_loss": 3.605608266348483, + "tokens_seen": 1238302720 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155566700100301, + "loss": 1.3853, + "theoretical_loss": 3.6055909550511154, + "tokens_seen": 1238368256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155466399197593, + "loss": 1.3824, + "theoretical_loss": 3.605573644926362, + "tokens_seen": 1238433792 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031553660982948844, + "loss": 1.5198, + "theoretical_loss": 3.6055563359740805, + "tokens_seen": 1238499328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155265797392177, + "loss": 1.4335, + "theoretical_loss": 3.60553902819413, + "tokens_seen": 1238564864 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 453261, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.574123501777649, + "objective/train/theoretical_loss": 3.605521721586369, + "objective/train/tokens_used": 1259090400, + "theoretical_loss": 3.605521721586369, + "tokens_seen": 1238630400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003155165496489468, + "loss": 1.5134, + "theoretical_loss": 3.605521721586369, + "tokens_seen": 1238630400 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031550651955867604, + "loss": 1.4013, + "theoretical_loss": 3.605504416150656, + "tokens_seen": 1238695936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154964894684052, + "loss": 1.4736, + "theoretical_loss": 3.6054871118868492, + "tokens_seen": 1238761472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154864593781344, + "loss": 1.3952, + "theoretical_loss": 3.6054698087948083, + "tokens_seen": 1238827008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154764292878636, + "loss": 1.3863, + "theoretical_loss": 3.605452506874391, + "tokens_seen": 1238892544 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031546639919759277, + "loss": 1.2833, + "theoretical_loss": 3.6054352061254566, + "tokens_seen": 1238958080 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031545636910732195, + "loss": 1.48, + "theoretical_loss": 3.6054179065478635, + "tokens_seen": 1239023616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154463390170512, + "loss": 1.3576, + "theoretical_loss": 3.605400608141471, + "tokens_seen": 1239089152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154363089267803, + "loss": 1.4532, + "theoretical_loss": 3.6053833109061366, + "tokens_seen": 1239154688 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031542627883650954, + "loss": 1.3568, + "theoretical_loss": 3.605366014841721, + "tokens_seen": 1239220224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154162487462387, + "loss": 1.5705, + "theoretical_loss": 3.605348719948081, + "tokens_seen": 1239285760 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003154062186559679, + "loss": 1.3137, + "theoretical_loss": 3.605331426225076, + "tokens_seen": 1239351296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153961885656971, + "loss": 1.5307, + "theoretical_loss": 3.6053141336725663, + "tokens_seen": 1239416832 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031538615847542627, + "loss": 1.4698, + "theoretical_loss": 3.605296842290409, + "tokens_seen": 1239482368 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031537612838515545, + "loss": 1.3365, + "theoretical_loss": 3.605279552078464, + "tokens_seen": 1239547904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153660982948847, + "loss": 1.3749, + "theoretical_loss": 3.6052622630365896, + "tokens_seen": 1239613440 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153560682046138, + "loss": 1.3632, + "theoretical_loss": 3.605244975164646, + "tokens_seen": 1239678976 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031534603811434305, + "loss": 1.3993, + "theoretical_loss": 3.6052276884624908, + "tokens_seen": 1239744512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153360080240722, + "loss": 1.3416, + "theoretical_loss": 3.605210402929983, + "tokens_seen": 1239810048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153259779338014, + "loss": 1.3307, + "theoretical_loss": 3.6051931185669828, + "tokens_seen": 1239875584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003153159478435306, + "loss": 1.4716, + "theoretical_loss": 3.6051758353733487, + "tokens_seen": 1239941120 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031530591775325977, + "loss": 1.4149, + "theoretical_loss": 3.6051585533489394, + "tokens_seen": 1240006656 + }, + { + "epoch": 0.38, + "learning_rate": 0.000315295887662989, + "loss": 1.4475, + "theoretical_loss": 3.605141272493615, + "tokens_seen": 1240072192 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031528585757271813, + "loss": 1.3955, + "theoretical_loss": 3.6051239928072336, + "tokens_seen": 1240137728 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031527582748244737, + "loss": 1.42, + "theoretical_loss": 3.6051067142896542, + "tokens_seen": 1240203264 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031526579739217655, + "loss": 1.3214, + "theoretical_loss": 3.6050894369407374, + "tokens_seen": 1240268800 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031525576730190573, + "loss": 1.4281, + "theoretical_loss": 3.605072160760341, + "tokens_seen": 1240334336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003152457372116349, + "loss": 1.4132, + "theoretical_loss": 3.605054885748325, + "tokens_seen": 1240399872 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031523570712136415, + "loss": 1.516, + "theoretical_loss": 3.605037611904548, + "tokens_seen": 1240465408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003152256770310933, + "loss": 1.4551, + "theoretical_loss": 3.6050203392288704, + "tokens_seen": 1240530944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003152156469408225, + "loss": 1.5162, + "theoretical_loss": 3.6050030677211504, + "tokens_seen": 1240596480 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031520561685055164, + "loss": 1.4114, + "theoretical_loss": 3.604985797381248, + "tokens_seen": 1240662016 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031519558676028087, + "loss": 1.4428, + "theoretical_loss": 3.604968528209022, + "tokens_seen": 1240727552 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031518555667001005, + "loss": 1.3695, + "theoretical_loss": 3.6049512602043325, + "tokens_seen": 1240793088 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031517552657973924, + "loss": 1.3179, + "theoretical_loss": 3.604933993367038, + "tokens_seen": 1240858624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003151654964894684, + "loss": 1.4088, + "theoretical_loss": 3.6049167276969984, + "tokens_seen": 1240924160 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003151554663991976, + "loss": 1.3251, + "theoretical_loss": 3.604899463194074, + "tokens_seen": 1240989696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003151454363089268, + "loss": 1.4435, + "theoretical_loss": 3.604882199858123, + "tokens_seen": 1241055232 + }, + { + "epoch": 0.38, + "learning_rate": 0.000315135406218656, + "loss": 1.3917, + "theoretical_loss": 3.6048649376890047, + "tokens_seen": 1241120768 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031512537612838514, + "loss": 1.3837, + "theoretical_loss": 3.60484767668658, + "tokens_seen": 1241186304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003151153460381144, + "loss": 1.44, + "theoretical_loss": 3.6048304168507075, + "tokens_seen": 1241251840 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003151053159478435, + "loss": 1.4712, + "theoretical_loss": 3.604813158181247, + "tokens_seen": 1241317376 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031509528585757274, + "loss": 1.4021, + "theoretical_loss": 3.6047959006780586, + "tokens_seen": 1241382912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150852557673019, + "loss": 1.4392, + "theoretical_loss": 3.6047786443410006, + "tokens_seen": 1241448448 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150752256770311, + "loss": 1.526, + "theoretical_loss": 3.6047613891699335, + "tokens_seen": 1241513984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150651955867603, + "loss": 1.4744, + "theoretical_loss": 3.604744135164718, + "tokens_seen": 1241579520 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150551654964895, + "loss": 1.4328, + "theoretical_loss": 3.6047268823252114, + "tokens_seen": 1241645056 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031504513540621864, + "loss": 1.4786, + "theoretical_loss": 3.604709630651276, + "tokens_seen": 1241710592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150351053159479, + "loss": 1.4315, + "theoretical_loss": 3.6046923801427697, + "tokens_seen": 1241776128 + }, + { + "epoch": 0.38, + "learning_rate": 0.000315025075225677, + "loss": 1.4597, + "theoretical_loss": 3.604675130799553, + "tokens_seen": 1241841664 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 454384, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4350159168243408, + "objective/train/theoretical_loss": 3.6046578826214852, + "objective/train/tokens_used": 1262367200, + "theoretical_loss": 3.6046578826214852, + "tokens_seen": 1241907200 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031501504513540624, + "loss": 1.3846, + "theoretical_loss": 3.6046578826214852, + "tokens_seen": 1241907200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003150050150451354, + "loss": 1.4667, + "theoretical_loss": 3.604640635608427, + "tokens_seen": 1241972736 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149949849548646, + "loss": 1.4319, + "theoretical_loss": 3.6046233897602376, + "tokens_seen": 1242038272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149849548645938, + "loss": 1.4199, + "theoretical_loss": 3.604606145076777, + "tokens_seen": 1242103808 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031497492477432297, + "loss": 1.4324, + "theoretical_loss": 3.6045889015579045, + "tokens_seen": 1242169344 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031496489468405215, + "loss": 1.4126, + "theoretical_loss": 3.6045716592034815, + "tokens_seen": 1242234880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149548645937814, + "loss": 1.3686, + "theoretical_loss": 3.6045544180133664, + "tokens_seen": 1242300416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149448345035105, + "loss": 1.4263, + "theoretical_loss": 3.6045371779874205, + "tokens_seen": 1242365952 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031493480441323974, + "loss": 1.3391, + "theoretical_loss": 3.6045199391255025, + "tokens_seen": 1242431488 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149247743229689, + "loss": 1.4709, + "theoretical_loss": 3.6045027014274735, + "tokens_seen": 1242497024 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149147442326981, + "loss": 1.3417, + "theoretical_loss": 3.6044854648931928, + "tokens_seen": 1242562560 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003149047141424273, + "loss": 1.5652, + "theoretical_loss": 3.6044682295225203, + "tokens_seen": 1242628096 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031489468405215647, + "loss": 1.4316, + "theoretical_loss": 3.604450995315317, + "tokens_seen": 1242693632 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031488465396188565, + "loss": 1.4673, + "theoretical_loss": 3.604433762271442, + "tokens_seen": 1242759168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003148746238716149, + "loss": 1.3938, + "theoretical_loss": 3.6044165303907567, + "tokens_seen": 1242824704 + }, + { + "epoch": 0.38, + "learning_rate": 0.000314864593781344, + "loss": 1.4939, + "theoretical_loss": 3.60439929967312, + "tokens_seen": 1242890240 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031485456369107325, + "loss": 1.4982, + "theoretical_loss": 3.6043820701183926, + "tokens_seen": 1242955776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003148445336008024, + "loss": 1.3819, + "theoretical_loss": 3.6043648417264347, + "tokens_seen": 1243021312 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003148345035105316, + "loss": 1.4894, + "theoretical_loss": 3.6043476144971063, + "tokens_seen": 1243086848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003148244734202608, + "loss": 1.4659, + "theoretical_loss": 3.6043303884302684, + "tokens_seen": 1243152384 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031481444332998997, + "loss": 1.3962, + "theoretical_loss": 3.604313163525781, + "tokens_seen": 1243217920 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031480441323971915, + "loss": 1.3655, + "theoretical_loss": 3.604295939783503, + "tokens_seen": 1243283456 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031479438314944833, + "loss": 1.2314, + "theoretical_loss": 3.6042787172032966, + "tokens_seen": 1243348992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003147843530591775, + "loss": 1.406, + "theoretical_loss": 3.604261495785021, + "tokens_seen": 1243414528 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031477432296890675, + "loss": 1.3693, + "theoretical_loss": 3.6042442755285373, + "tokens_seen": 1243480064 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003147642928786359, + "loss": 1.4607, + "theoretical_loss": 3.6042270564337056, + "tokens_seen": 1243545600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003147542627883651, + "loss": 1.4147, + "theoretical_loss": 3.6042098385003865, + "tokens_seen": 1243611136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003147442326980943, + "loss": 1.4505, + "theoretical_loss": 3.6041926217284397, + "tokens_seen": 1243676672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003147342026078235, + "loss": 1.4845, + "theoretical_loss": 3.6041754061177267, + "tokens_seen": 1243742208 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031472417251755266, + "loss": 1.4875, + "theoretical_loss": 3.604158191668107, + "tokens_seen": 1243807744 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031471414242728184, + "loss": 1.4158, + "theoretical_loss": 3.6041409783794425, + "tokens_seen": 1243873280 + }, + { + "epoch": 0.38, + "learning_rate": 0.000314704112337011, + "loss": 1.4357, + "theoretical_loss": 3.6041237662515924, + "tokens_seen": 1243938816 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031469408224674025, + "loss": 1.4502, + "theoretical_loss": 3.6041065552844174, + "tokens_seen": 1244004352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003146840521564694, + "loss": 1.4074, + "theoretical_loss": 3.604089345477779, + "tokens_seen": 1244069888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003146740220661986, + "loss": 1.537, + "theoretical_loss": 3.6040721368315367, + "tokens_seen": 1244135424 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031466399197592774, + "loss": 1.3128, + "theoretical_loss": 3.604054929345552, + "tokens_seen": 1244200960 + }, + { + "epoch": 0.38, + "learning_rate": 0.000314653961885657, + "loss": 1.6144, + "theoretical_loss": 3.604037723019686, + "tokens_seen": 1244266496 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031464393179538616, + "loss": 1.4992, + "theoretical_loss": 3.6040205178537974, + "tokens_seen": 1244332032 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031463390170511534, + "loss": 1.384, + "theoretical_loss": 3.6040033138477487, + "tokens_seen": 1244397568 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003146238716148445, + "loss": 1.4913, + "theoretical_loss": 3.6039861110014004, + "tokens_seen": 1244463104 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003146138415245737, + "loss": 1.2978, + "theoretical_loss": 3.6039689093146126, + "tokens_seen": 1244528640 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003146038114343029, + "loss": 1.4401, + "theoretical_loss": 3.6039517087872466, + "tokens_seen": 1244594176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145937813440321, + "loss": 1.4818, + "theoretical_loss": 3.603934509419163, + "tokens_seen": 1244659712 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031458375125376125, + "loss": 1.3831, + "theoretical_loss": 3.603917311210223, + "tokens_seen": 1244725248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145737211634905, + "loss": 1.3807, + "theoretical_loss": 3.603900114160287, + "tokens_seen": 1244790784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145636910732197, + "loss": 1.3983, + "theoretical_loss": 3.6038829182692154, + "tokens_seen": 1244856320 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031455366098294884, + "loss": 1.4433, + "theoretical_loss": 3.6038657235368703, + "tokens_seen": 1244921856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145436308926781, + "loss": 1.4283, + "theoretical_loss": 3.603848529963112, + "tokens_seen": 1244987392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145336008024072, + "loss": 1.3734, + "theoretical_loss": 3.6038313375478013, + "tokens_seen": 1245052928 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031452357071213644, + "loss": 1.3363, + "theoretical_loss": 3.6038141462907998, + "tokens_seen": 1245118464 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 455531, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4752833843231201, + "objective/train/theoretical_loss": 3.6037969561919674, + "objective/train/tokens_used": 1265644000, + "theoretical_loss": 3.6037969561919674, + "tokens_seen": 1245184000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145135406218656, + "loss": 1.5299, + "theoretical_loss": 3.6037969561919674, + "tokens_seen": 1245184000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003145035105315948, + "loss": 1.39, + "theoretical_loss": 3.6037797672511664, + "tokens_seen": 1245249536 + }, + { + "epoch": 0.38, + "learning_rate": 0.000314493480441324, + "loss": 1.526, + "theoretical_loss": 3.603762579468257, + "tokens_seen": 1245315072 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031448345035105317, + "loss": 1.4233, + "theoretical_loss": 3.6037453928431007, + "tokens_seen": 1245380608 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031447342026078235, + "loss": 1.3487, + "theoretical_loss": 3.6037282073755583, + "tokens_seen": 1245446144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003144633901705116, + "loss": 1.5433, + "theoretical_loss": 3.6037110230654914, + "tokens_seen": 1245511680 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003144533600802407, + "loss": 1.349, + "theoretical_loss": 3.6036938399127605, + "tokens_seen": 1245577216 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031444332998996994, + "loss": 1.5682, + "theoretical_loss": 3.6036766579172266, + "tokens_seen": 1245642752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003144332998996991, + "loss": 1.3997, + "theoretical_loss": 3.603659477078752, + "tokens_seen": 1245708288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003144232698094283, + "loss": 1.3942, + "theoretical_loss": 3.6036422973971973, + "tokens_seen": 1245773824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003144132397191575, + "loss": 1.4915, + "theoretical_loss": 3.603625118872423, + "tokens_seen": 1245839360 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031440320962888667, + "loss": 1.3522, + "theoretical_loss": 3.6036079415042916, + "tokens_seen": 1245904896 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031439317953861585, + "loss": 1.4734, + "theoretical_loss": 3.603590765292664, + "tokens_seen": 1245970432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003143831494483451, + "loss": 1.3392, + "theoretical_loss": 3.603573590237401, + "tokens_seen": 1246035968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003143731193580742, + "loss": 1.4439, + "theoretical_loss": 3.603556416338365, + "tokens_seen": 1246101504 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031436308926780345, + "loss": 1.525, + "theoretical_loss": 3.6035392435954154, + "tokens_seen": 1246167040 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003143530591775326, + "loss": 1.5369, + "theoretical_loss": 3.603522072008416, + "tokens_seen": 1246232576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003143430290872618, + "loss": 1.4748, + "theoretical_loss": 3.6035049015772267, + "tokens_seen": 1246298112 + }, + { + "epoch": 0.38, + "learning_rate": 0.000314332998996991, + "loss": 1.3625, + "theoretical_loss": 3.6034877323017094, + "tokens_seen": 1246363648 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031432296890672017, + "loss": 1.329, + "theoretical_loss": 3.603470564181725, + "tokens_seen": 1246429184 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031431293881644935, + "loss": 1.4424, + "theoretical_loss": 3.6034533972171356, + "tokens_seen": 1246494720 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031430290872617853, + "loss": 1.4611, + "theoretical_loss": 3.6034362314078026, + "tokens_seen": 1246560256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142928786359077, + "loss": 1.454, + "theoretical_loss": 3.6034190667535877, + "tokens_seen": 1246625792 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031428284854563695, + "loss": 1.5631, + "theoretical_loss": 3.603401903254352, + "tokens_seen": 1246691328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142728184553661, + "loss": 1.4511, + "theoretical_loss": 3.6033847409099575, + "tokens_seen": 1246756864 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142627883650953, + "loss": 1.4439, + "theoretical_loss": 3.603367579720265, + "tokens_seen": 1246822400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142527582748245, + "loss": 1.3707, + "theoretical_loss": 3.603350419685137, + "tokens_seen": 1246887936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142427281845537, + "loss": 1.462, + "theoretical_loss": 3.603333260804435, + "tokens_seen": 1246953472 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031423269809428286, + "loss": 1.5791, + "theoretical_loss": 3.6033161030780203, + "tokens_seen": 1247019008 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031422266800401204, + "loss": 1.4496, + "theoretical_loss": 3.603298946505755, + "tokens_seen": 1247084544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003142126379137412, + "loss": 1.4597, + "theoretical_loss": 3.6032817910875004, + "tokens_seen": 1247150080 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031420260782347045, + "loss": 1.3909, + "theoretical_loss": 3.603264636823119, + "tokens_seen": 1247215616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141925777331996, + "loss": 1.4017, + "theoretical_loss": 3.6032474837124715, + "tokens_seen": 1247281152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141825476429288, + "loss": 1.3866, + "theoretical_loss": 3.6032303317554204, + "tokens_seen": 1247346688 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031417251755265794, + "loss": 1.3175, + "theoretical_loss": 3.603213180951827, + "tokens_seen": 1247412224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141624874623872, + "loss": 1.442, + "theoretical_loss": 3.603196031301554, + "tokens_seen": 1247477760 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031415245737211636, + "loss": 1.3116, + "theoretical_loss": 3.603178882804462, + "tokens_seen": 1247543296 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031414242728184554, + "loss": 1.4533, + "theoretical_loss": 3.6031617354604144, + "tokens_seen": 1247608832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141323971915747, + "loss": 1.4092, + "theoretical_loss": 3.6031445892692724, + "tokens_seen": 1247674368 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141223671013039, + "loss": 1.3328, + "theoretical_loss": 3.603127444230897, + "tokens_seen": 1247739904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141123370110331, + "loss": 1.4561, + "theoretical_loss": 3.603110300345152, + "tokens_seen": 1247805440 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003141023069207623, + "loss": 1.3672, + "theoretical_loss": 3.603093157611898, + "tokens_seen": 1247870976 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031409227683049145, + "loss": 1.5872, + "theoretical_loss": 3.603076016030997, + "tokens_seen": 1247936512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140822467402207, + "loss": 1.3722, + "theoretical_loss": 3.6030588756023123, + "tokens_seen": 1248002048 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031407221664994986, + "loss": 1.4683, + "theoretical_loss": 3.603041736325704, + "tokens_seen": 1248067584 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031406218655967904, + "loss": 1.3703, + "theoretical_loss": 3.603024598201036, + "tokens_seen": 1248133120 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140521564694082, + "loss": 1.3044, + "theoretical_loss": 3.6030074612281693, + "tokens_seen": 1248198656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140421263791374, + "loss": 1.4778, + "theoretical_loss": 3.6029903254069664, + "tokens_seen": 1248264192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140320962888666, + "loss": 1.4238, + "theoretical_loss": 3.6029731907372895, + "tokens_seen": 1248329728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140220661985958, + "loss": 1.5074, + "theoretical_loss": 3.602956057219001, + "tokens_seen": 1248395264 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 456712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3785254955291748, + "objective/train/theoretical_loss": 3.602938924851962, + "objective/train/tokens_used": 1268920800, + "theoretical_loss": 3.602938924851962, + "tokens_seen": 1248460800 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031401203610832495, + "loss": 1.3945, + "theoretical_loss": 3.602938924851962, + "tokens_seen": 1248460800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003140020060180542, + "loss": 1.4761, + "theoretical_loss": 3.6029217936360363, + "tokens_seen": 1248526336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139919759277833, + "loss": 1.4941, + "theoretical_loss": 3.602904663571085, + "tokens_seen": 1248591872 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031398194583751255, + "loss": 1.3944, + "theoretical_loss": 3.602887534656971, + "tokens_seen": 1248657408 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031397191574724173, + "loss": 1.459, + "theoretical_loss": 3.6028704068935555, + "tokens_seen": 1248722944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139618856569709, + "loss": 1.3382, + "theoretical_loss": 3.6028532802807023, + "tokens_seen": 1248788480 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139518555667001, + "loss": 1.4037, + "theoretical_loss": 3.602836154818273, + "tokens_seen": 1248854016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139418254764293, + "loss": 1.4552, + "theoretical_loss": 3.6028190305061303, + "tokens_seen": 1248919552 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031393179538615845, + "loss": 1.3768, + "theoretical_loss": 3.6028019073441357, + "tokens_seen": 1248985088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139217652958877, + "loss": 1.4984, + "theoretical_loss": 3.602784785332152, + "tokens_seen": 1249050624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003139117352056168, + "loss": 1.3251, + "theoretical_loss": 3.6027676644700426, + "tokens_seen": 1249116160 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031390170511534605, + "loss": 1.479, + "theoretical_loss": 3.6027505447576686, + "tokens_seen": 1249181696 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031389167502507523, + "loss": 1.5208, + "theoretical_loss": 3.6027334261948933, + "tokens_seen": 1249247232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138816449348044, + "loss": 1.4406, + "theoretical_loss": 3.6027163087815786, + "tokens_seen": 1249312768 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138716148445336, + "loss": 1.3644, + "theoretical_loss": 3.602699192517588, + "tokens_seen": 1249378304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138615847542628, + "loss": 1.4909, + "theoretical_loss": 3.6026820774027826, + "tokens_seen": 1249443840 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031385155466399196, + "loss": 1.4489, + "theoretical_loss": 3.6026649634370265, + "tokens_seen": 1249509376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138415245737212, + "loss": 1.4129, + "theoretical_loss": 3.6026478506201816, + "tokens_seen": 1249574912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138314944834503, + "loss": 1.4566, + "theoretical_loss": 3.6026307389521097, + "tokens_seen": 1249640448 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031382146439317955, + "loss": 1.5139, + "theoretical_loss": 3.602613628432675, + "tokens_seen": 1249705984 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031381143430290873, + "loss": 1.4231, + "theoretical_loss": 3.6025965190617395, + "tokens_seen": 1249771520 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003138014042126379, + "loss": 1.432, + "theoretical_loss": 3.6025794108391658, + "tokens_seen": 1249837056 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031379137412236715, + "loss": 1.408, + "theoretical_loss": 3.602562303764816, + "tokens_seen": 1249902592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137813440320963, + "loss": 1.4297, + "theoretical_loss": 3.6025451978385536, + "tokens_seen": 1249968128 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137713139418255, + "loss": 1.4472, + "theoretical_loss": 3.6025280930602417, + "tokens_seen": 1250033664 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137612838515547, + "loss": 1.4353, + "theoretical_loss": 3.6025109894297427, + "tokens_seen": 1250099200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137512537612839, + "loss": 1.3179, + "theoretical_loss": 3.602493886946919, + "tokens_seen": 1250164736 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031374122367101306, + "loss": 1.4821, + "theoretical_loss": 3.6024767856116338, + "tokens_seen": 1250230272 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031373119358074224, + "loss": 1.4869, + "theoretical_loss": 3.60245968542375, + "tokens_seen": 1250295808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137211634904714, + "loss": 1.3777, + "theoretical_loss": 3.6024425863831304, + "tokens_seen": 1250361344 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031371113340020065, + "loss": 1.3239, + "theoretical_loss": 3.6024254884896374, + "tokens_seen": 1250426880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003137011033099298, + "loss": 1.408, + "theoretical_loss": 3.6024083917431353, + "tokens_seen": 1250492416 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313691073219659, + "loss": 1.4207, + "theoretical_loss": 3.602391296143485, + "tokens_seen": 1250557952 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031368104312938814, + "loss": 1.4352, + "theoretical_loss": 3.6023742016905516, + "tokens_seen": 1250623488 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003136710130391174, + "loss": 1.4975, + "theoretical_loss": 3.602357108384197, + "tokens_seen": 1250689024 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031366098294884656, + "loss": 1.3539, + "theoretical_loss": 3.602340016224284, + "tokens_seen": 1250754560 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031365095285857574, + "loss": 1.412, + "theoretical_loss": 3.6023229252106765, + "tokens_seen": 1250820096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003136409227683049, + "loss": 1.4096, + "theoretical_loss": 3.6023058353432367, + "tokens_seen": 1250885632 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003136308926780341, + "loss": 1.4303, + "theoretical_loss": 3.602288746621828, + "tokens_seen": 1250951168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003136208625877633, + "loss": 1.5196, + "theoretical_loss": 3.602271659046314, + "tokens_seen": 1251016704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003136108324974925, + "loss": 1.3061, + "theoretical_loss": 3.602254572616557, + "tokens_seen": 1251082240 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031360080240722165, + "loss": 1.3553, + "theoretical_loss": 3.602237487332421, + "tokens_seen": 1251147776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135907723169509, + "loss": 1.4442, + "theoretical_loss": 3.6022204031937686, + "tokens_seen": 1251213312 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031358074222668006, + "loss": 1.4896, + "theoretical_loss": 3.602203320200463, + "tokens_seen": 1251278848 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031357071213640924, + "loss": 1.3556, + "theoretical_loss": 3.6021862383523677, + "tokens_seen": 1251344384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135606820461384, + "loss": 1.3244, + "theoretical_loss": 3.602169157649346, + "tokens_seen": 1251409920 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135506519558676, + "loss": 1.427, + "theoretical_loss": 3.602152078091261, + "tokens_seen": 1251475456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135406218655968, + "loss": 1.3305, + "theoretical_loss": 3.6021349996779763, + "tokens_seen": 1251540992 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313530591775326, + "loss": 1.4165, + "theoretical_loss": 3.6021179224093545, + "tokens_seen": 1251606528 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031352056168505515, + "loss": 1.3387, + "theoretical_loss": 3.6021008462852593, + "tokens_seen": 1251672064 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 458046, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.413086175918579, + "objective/train/theoretical_loss": 3.6020837713055545, + "objective/train/tokens_used": 1272197600, + "theoretical_loss": 3.6020837713055545, + "tokens_seen": 1251737600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135105315947844, + "loss": 1.4294, + "theoretical_loss": 3.6020837713055545, + "tokens_seen": 1251737600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003135005015045135, + "loss": 1.4362, + "theoretical_loss": 3.602066697470103, + "tokens_seen": 1251803136 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031349047141424275, + "loss": 1.4923, + "theoretical_loss": 3.6020496247787688, + "tokens_seen": 1251868672 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031348044132397193, + "loss": 1.3793, + "theoretical_loss": 3.602032553231415, + "tokens_seen": 1251934208 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003134704112337011, + "loss": 1.4908, + "theoretical_loss": 3.6020154828279045, + "tokens_seen": 1251999744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003134603811434303, + "loss": 1.3355, + "theoretical_loss": 3.6019984135681016, + "tokens_seen": 1252065280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003134503510531595, + "loss": 1.4165, + "theoretical_loss": 3.601981345451869, + "tokens_seen": 1252130816 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031344032096288865, + "loss": 1.467, + "theoretical_loss": 3.6019642784790715, + "tokens_seen": 1252196352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003134302908726179, + "loss": 1.4877, + "theoretical_loss": 3.6019472126495717, + "tokens_seen": 1252261888 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313420260782347, + "loss": 1.3164, + "theoretical_loss": 3.601930147963234, + "tokens_seen": 1252327424 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031341023069207625, + "loss": 1.4257, + "theoretical_loss": 3.6019130844199205, + "tokens_seen": 1252392960 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031340020060180543, + "loss": 1.3816, + "theoretical_loss": 3.601896022019496, + "tokens_seen": 1252458496 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133901705115346, + "loss": 1.4309, + "theoretical_loss": 3.601878960761824, + "tokens_seen": 1252524032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133801404212638, + "loss": 1.3934, + "theoretical_loss": 3.6018619006467683, + "tokens_seen": 1252589568 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313370110330993, + "loss": 1.4819, + "theoretical_loss": 3.601844841674192, + "tokens_seen": 1252655104 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031336008024072216, + "loss": 1.3448, + "theoretical_loss": 3.6018277838439596, + "tokens_seen": 1252720640 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133500501504514, + "loss": 1.4616, + "theoretical_loss": 3.601810727155934, + "tokens_seen": 1252786176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133400200601805, + "loss": 1.3306, + "theoretical_loss": 3.6017936716099794, + "tokens_seen": 1252851712 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031332998996990975, + "loss": 1.4621, + "theoretical_loss": 3.60177661720596, + "tokens_seen": 1252917248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133199598796389, + "loss": 1.4535, + "theoretical_loss": 3.601759563943739, + "tokens_seen": 1252982784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003133099297893681, + "loss": 1.3819, + "theoretical_loss": 3.601742511823181, + "tokens_seen": 1253048320 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132998996990973, + "loss": 1.4725, + "theoretical_loss": 3.6017254608441487, + "tokens_seen": 1253113856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132898696088265, + "loss": 1.5065, + "theoretical_loss": 3.6017084110065065, + "tokens_seen": 1253179392 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031327983951855566, + "loss": 1.5243, + "theoretical_loss": 3.601691362310119, + "tokens_seen": 1253244928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132698094282849, + "loss": 1.3856, + "theoretical_loss": 3.6016743147548493, + "tokens_seen": 1253310464 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313259779338014, + "loss": 1.3615, + "theoretical_loss": 3.6016572683405617, + "tokens_seen": 1253376000 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031324974924774326, + "loss": 1.4629, + "theoretical_loss": 3.60164022306712, + "tokens_seen": 1253441536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132397191574724, + "loss": 1.4215, + "theoretical_loss": 3.6016231789343887, + "tokens_seen": 1253507072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132296890672016, + "loss": 1.4217, + "theoretical_loss": 3.601606135942231, + "tokens_seen": 1253572608 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003132196589769308, + "loss": 1.4322, + "theoretical_loss": 3.601589094090512, + "tokens_seen": 1253638144 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031320962888666, + "loss": 1.4412, + "theoretical_loss": 3.6015720533790945, + "tokens_seen": 1253703680 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031319959879638916, + "loss": 1.3985, + "theoretical_loss": 3.601555013807844, + "tokens_seen": 1253769216 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031318956870611834, + "loss": 1.4258, + "theoretical_loss": 3.6015379753766235, + "tokens_seen": 1253834752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003131795386158475, + "loss": 1.4279, + "theoretical_loss": 3.6015209380852973, + "tokens_seen": 1253900288 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031316950852557676, + "loss": 1.4477, + "theoretical_loss": 3.60150390193373, + "tokens_seen": 1253965824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003131594784353059, + "loss": 1.5009, + "theoretical_loss": 3.6014868669217854, + "tokens_seen": 1254031360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003131494483450351, + "loss": 1.4116, + "theoretical_loss": 3.6014698330493284, + "tokens_seen": 1254096896 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031313941825476425, + "loss": 1.421, + "theoretical_loss": 3.6014528003162223, + "tokens_seen": 1254162432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003131293881644935, + "loss": 1.3708, + "theoretical_loss": 3.6014357687223324, + "tokens_seen": 1254227968 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031311935807422267, + "loss": 1.3909, + "theoretical_loss": 3.6014187382675216, + "tokens_seen": 1254293504 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031310932798395185, + "loss": 1.4746, + "theoretical_loss": 3.6014017089516557, + "tokens_seen": 1254359040 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031309929789368103, + "loss": 1.4278, + "theoretical_loss": 3.601384680774598, + "tokens_seen": 1254424576 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031308926780341026, + "loss": 1.397, + "theoretical_loss": 3.6013676537362134, + "tokens_seen": 1254490112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130792377131394, + "loss": 1.4189, + "theoretical_loss": 3.601350627836366, + "tokens_seen": 1254555648 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130692076228686, + "loss": 1.3963, + "theoretical_loss": 3.60133360307492, + "tokens_seen": 1254621184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130591775325978, + "loss": 1.4687, + "theoretical_loss": 3.6013165794517406, + "tokens_seen": 1254686720 + }, + { + "epoch": 0.38, + "learning_rate": 0.000313049147442327, + "loss": 1.3297, + "theoretical_loss": 3.6012995569666915, + "tokens_seen": 1254752256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130391173520562, + "loss": 1.5056, + "theoretical_loss": 3.601282535619638, + "tokens_seen": 1254817792 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031302908726178535, + "loss": 1.3802, + "theoretical_loss": 3.601265515410443, + "tokens_seen": 1254883328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130190571715146, + "loss": 1.3709, + "theoretical_loss": 3.6012484963389726, + "tokens_seen": 1254948864 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 459153, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4800946712493896, + "objective/train/theoretical_loss": 3.60123147840509, + "objective/train/tokens_used": 1275474400, + "theoretical_loss": 3.60123147840509, + "tokens_seen": 1255014400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003130090270812437, + "loss": 1.48, + "theoretical_loss": 3.60123147840509, + "tokens_seen": 1255014400 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031299899699097295, + "loss": 1.3625, + "theoretical_loss": 3.601214461608661, + "tokens_seen": 1255079936 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031298896690070213, + "loss": 1.4507, + "theoretical_loss": 3.60119744594955, + "tokens_seen": 1255145472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003129789368104313, + "loss": 1.3612, + "theoretical_loss": 3.601180431427621, + "tokens_seen": 1255211008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003129689067201605, + "loss": 1.418, + "theoretical_loss": 3.6011634180427397, + "tokens_seen": 1255276544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003129588766298897, + "loss": 1.3908, + "theoretical_loss": 3.6011464057947693, + "tokens_seen": 1255342080 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031294884653961885, + "loss": 1.4246, + "theoretical_loss": 3.601129394683575, + "tokens_seen": 1255407616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003129388164493481, + "loss": 1.4196, + "theoretical_loss": 3.6011123847090225, + "tokens_seen": 1255473152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003129287863590772, + "loss": 1.4278, + "theoretical_loss": 3.601095375870975, + "tokens_seen": 1255538688 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031291875626880645, + "loss": 1.4112, + "theoretical_loss": 3.601078368169298, + "tokens_seen": 1255604224 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031290872617853563, + "loss": 1.3988, + "theoretical_loss": 3.6010613616038567, + "tokens_seen": 1255669760 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128986960882648, + "loss": 1.4168, + "theoretical_loss": 3.6010443561745156, + "tokens_seen": 1255735296 + }, + { + "epoch": 0.38, + "learning_rate": 0.000312888665997994, + "loss": 1.3725, + "theoretical_loss": 3.6010273518811386, + "tokens_seen": 1255800832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128786359077232, + "loss": 1.4665, + "theoretical_loss": 3.601010348723592, + "tokens_seen": 1255866368 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031286860581745236, + "loss": 1.3802, + "theoretical_loss": 3.600993346701739, + "tokens_seen": 1255931904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128585757271816, + "loss": 1.3422, + "theoretical_loss": 3.6009763458154462, + "tokens_seen": 1255997440 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128485456369107, + "loss": 1.4102, + "theoretical_loss": 3.6009593460645775, + "tokens_seen": 1256062976 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031283851554663995, + "loss": 1.3426, + "theoretical_loss": 3.6009423474489983, + "tokens_seen": 1256128512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128284854563691, + "loss": 1.415, + "theoretical_loss": 3.6009253499685734, + "tokens_seen": 1256194048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128184553660983, + "loss": 1.4141, + "theoretical_loss": 3.6009083536231676, + "tokens_seen": 1256259584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003128084252758275, + "loss": 1.3612, + "theoretical_loss": 3.6008913584126456, + "tokens_seen": 1256325120 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127983951855567, + "loss": 1.3944, + "theoretical_loss": 3.6008743643368737, + "tokens_seen": 1256390656 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031278836509528586, + "loss": 1.4152, + "theoretical_loss": 3.6008573713957155, + "tokens_seen": 1256456192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127783350050151, + "loss": 1.4035, + "theoretical_loss": 3.6008403795890365, + "tokens_seen": 1256521728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127683049147442, + "loss": 1.5312, + "theoretical_loss": 3.6008233889167025, + "tokens_seen": 1256587264 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031275827482447346, + "loss": 1.4283, + "theoretical_loss": 3.6008063993785777, + "tokens_seen": 1256652800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127482447342026, + "loss": 1.3516, + "theoretical_loss": 3.600789410974528, + "tokens_seen": 1256718336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127382146439318, + "loss": 1.395, + "theoretical_loss": 3.600772423704418, + "tokens_seen": 1256783872 + }, + { + "epoch": 0.38, + "learning_rate": 0.000312728184553661, + "loss": 1.3964, + "theoretical_loss": 3.6007554375681132, + "tokens_seen": 1256849408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003127181544633902, + "loss": 1.4412, + "theoretical_loss": 3.600738452565478, + "tokens_seen": 1256914944 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031270812437311936, + "loss": 1.3339, + "theoretical_loss": 3.6007214686963795, + "tokens_seen": 1256980480 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031269809428284854, + "loss": 1.4072, + "theoretical_loss": 3.600704485960681, + "tokens_seen": 1257046016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003126880641925777, + "loss": 1.4663, + "theoretical_loss": 3.600687504358249, + "tokens_seen": 1257111552 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031267803410230696, + "loss": 1.492, + "theoretical_loss": 3.6006705238889483, + "tokens_seen": 1257177088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003126680040120361, + "loss": 1.3789, + "theoretical_loss": 3.600653544552644, + "tokens_seen": 1257242624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003126579739217653, + "loss": 1.4088, + "theoretical_loss": 3.600636566349202, + "tokens_seen": 1257308160 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031264794383149445, + "loss": 1.3657, + "theoretical_loss": 3.6006195892784874, + "tokens_seen": 1257373696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003126379137412237, + "loss": 1.4631, + "theoretical_loss": 3.6006026133403655, + "tokens_seen": 1257439232 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031262788365095287, + "loss": 1.3707, + "theoretical_loss": 3.6005856385347013, + "tokens_seen": 1257504768 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031261785356068205, + "loss": 1.4546, + "theoretical_loss": 3.6005686648613615, + "tokens_seen": 1257570304 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031260782347041123, + "loss": 1.3987, + "theoretical_loss": 3.6005516923202108, + "tokens_seen": 1257635840 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031259779338014046, + "loss": 1.3192, + "theoretical_loss": 3.6005347209111145, + "tokens_seen": 1257701376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003125877632898696, + "loss": 1.414, + "theoretical_loss": 3.600517750633938, + "tokens_seen": 1257766912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003125777331995988, + "loss": 1.3637, + "theoretical_loss": 3.600500781488548, + "tokens_seen": 1257832448 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031256770310932795, + "loss": 1.3835, + "theoretical_loss": 3.600483813474809, + "tokens_seen": 1257897984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003125576730190572, + "loss": 1.4632, + "theoretical_loss": 3.600466846592586, + "tokens_seen": 1257963520 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031254764292878637, + "loss": 1.4079, + "theoretical_loss": 3.6004498808417464, + "tokens_seen": 1258029056 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031253761283851555, + "loss": 1.3234, + "theoretical_loss": 3.6004329162221542, + "tokens_seen": 1258094592 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031252758274824473, + "loss": 1.3674, + "theoretical_loss": 3.600415952733676, + "tokens_seen": 1258160128 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003125175526579739, + "loss": 1.3814, + "theoretical_loss": 3.600398990376177, + "tokens_seen": 1258225664 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 460354, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.536881446838379, + "objective/train/theoretical_loss": 3.600382029149523, + "objective/train/tokens_used": 1278751200, + "theoretical_loss": 3.600382029149523, + "tokens_seen": 1258291200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003125075225677031, + "loss": 1.4546, + "theoretical_loss": 3.600382029149523, + "tokens_seen": 1258291200 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031249749247743233, + "loss": 1.3449, + "theoretical_loss": 3.60036506905358, + "tokens_seen": 1258356736 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031248746238716146, + "loss": 1.319, + "theoretical_loss": 3.600348110088213, + "tokens_seen": 1258422272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003124774322968907, + "loss": 1.3703, + "theoretical_loss": 3.600331152253289, + "tokens_seen": 1258487808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003124674022066198, + "loss": 1.3542, + "theoretical_loss": 3.6003141955486724, + "tokens_seen": 1258553344 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031245737211634905, + "loss": 1.4626, + "theoretical_loss": 3.6002972399742297, + "tokens_seen": 1258618880 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031244734202607823, + "loss": 1.4607, + "theoretical_loss": 3.600280285529827, + "tokens_seen": 1258684416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003124373119358074, + "loss": 1.4659, + "theoretical_loss": 3.6002633322153295, + "tokens_seen": 1258749952 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003124272818455366, + "loss": 1.4441, + "theoretical_loss": 3.6002463800306037, + "tokens_seen": 1258815488 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031241725175526583, + "loss": 1.447, + "theoretical_loss": 3.6002294289755152, + "tokens_seen": 1258881024 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031240722166499496, + "loss": 1.3161, + "theoretical_loss": 3.6002124790499295, + "tokens_seen": 1258946560 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123971915747242, + "loss": 1.4637, + "theoretical_loss": 3.600195530253713, + "tokens_seen": 1259012096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123871614844533, + "loss": 1.5488, + "theoretical_loss": 3.6001785825867323, + "tokens_seen": 1259077632 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031237713139418256, + "loss": 1.4295, + "theoretical_loss": 3.6001616360488526, + "tokens_seen": 1259143168 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031236710130391174, + "loss": 1.3686, + "theoretical_loss": 3.6001446906399397, + "tokens_seen": 1259208704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123570712136409, + "loss": 1.4689, + "theoretical_loss": 3.60012774635986, + "tokens_seen": 1259274240 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123470411233701, + "loss": 1.4561, + "theoretical_loss": 3.60011080320848, + "tokens_seen": 1259339776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123370110330993, + "loss": 1.4499, + "theoretical_loss": 3.600093861185665, + "tokens_seen": 1259405312 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031232698094282846, + "loss": 1.4854, + "theoretical_loss": 3.600076920291281, + "tokens_seen": 1259470848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123169508525577, + "loss": 1.4454, + "theoretical_loss": 3.600059980525195, + "tokens_seen": 1259536384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003123069207622869, + "loss": 1.4488, + "theoretical_loss": 3.600043041887273, + "tokens_seen": 1259601920 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031229689067201606, + "loss": 1.3884, + "theoretical_loss": 3.600026104377381, + "tokens_seen": 1259667456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003122868605817453, + "loss": 1.3907, + "theoretical_loss": 3.6000091679953847, + "tokens_seen": 1259732992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003122768304914744, + "loss": 1.4401, + "theoretical_loss": 3.5999922327411507, + "tokens_seen": 1259798528 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031226680040120366, + "loss": 1.3408, + "theoretical_loss": 3.5999752986145452, + "tokens_seen": 1259864064 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003122567703109328, + "loss": 1.5165, + "theoretical_loss": 3.599958365615435, + "tokens_seen": 1259929600 + }, + { + "epoch": 0.38, + "learning_rate": 0.000312246740220662, + "loss": 1.4598, + "theoretical_loss": 3.5999414337436852, + "tokens_seen": 1259995136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003122367101303912, + "loss": 1.4592, + "theoretical_loss": 3.5999245029991633, + "tokens_seen": 1260060672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003122266800401204, + "loss": 1.3392, + "theoretical_loss": 3.599907573381735, + "tokens_seen": 1260126208 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031221664994984956, + "loss": 1.445, + "theoretical_loss": 3.599890644891267, + "tokens_seen": 1260191744 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031220661985957874, + "loss": 1.4698, + "theoretical_loss": 3.5998737175276254, + "tokens_seen": 1260257280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003121965897693079, + "loss": 1.3474, + "theoretical_loss": 3.5998567912906765, + "tokens_seen": 1260322816 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031218655967903716, + "loss": 1.4796, + "theoretical_loss": 3.599839866180287, + "tokens_seen": 1260388352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003121765295887663, + "loss": 1.3928, + "theoretical_loss": 3.5998229421963233, + "tokens_seen": 1260453888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003121664994984955, + "loss": 1.5117, + "theoretical_loss": 3.5998060193386516, + "tokens_seen": 1260519424 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031215646940822465, + "loss": 1.43, + "theoretical_loss": 3.599789097607139, + "tokens_seen": 1260584960 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003121464393179539, + "loss": 1.3114, + "theoretical_loss": 3.5997721770016518, + "tokens_seen": 1260650496 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031213640922768307, + "loss": 1.3868, + "theoretical_loss": 3.599755257522056, + "tokens_seen": 1260716032 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031212637913741225, + "loss": 1.3977, + "theoretical_loss": 3.5997383391682187, + "tokens_seen": 1260781568 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031211634904714143, + "loss": 1.5182, + "theoretical_loss": 3.599721421940006, + "tokens_seen": 1260847104 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031210631895687066, + "loss": 1.4905, + "theoretical_loss": 3.599704505837286, + "tokens_seen": 1260912640 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003120962888665998, + "loss": 1.5003, + "theoretical_loss": 3.599687590859923, + "tokens_seen": 1260978176 + }, + { + "epoch": 0.38, + "learning_rate": 0.000312086258776329, + "loss": 1.3299, + "theoretical_loss": 3.599670677007785, + "tokens_seen": 1261043712 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031207622868605815, + "loss": 1.4685, + "theoretical_loss": 3.5996537642807387, + "tokens_seen": 1261109248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003120661985957874, + "loss": 1.3965, + "theoretical_loss": 3.5996368526786506, + "tokens_seen": 1261174784 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031205616850551657, + "loss": 1.4627, + "theoretical_loss": 3.5996199422013877, + "tokens_seen": 1261240320 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031204613841524575, + "loss": 1.4134, + "theoretical_loss": 3.599603032848816, + "tokens_seen": 1261305856 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031203610832497493, + "loss": 1.3195, + "theoretical_loss": 3.599586124620803, + "tokens_seen": 1261371392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003120260782347041, + "loss": 1.3685, + "theoretical_loss": 3.5995692175172147, + "tokens_seen": 1261436928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003120160481444333, + "loss": 1.3562, + "theoretical_loss": 3.599552311537919, + "tokens_seen": 1261502464 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 461530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3671680688858032, + "objective/train/theoretical_loss": 3.599535406682782, + "objective/train/tokens_used": 1282028000, + "theoretical_loss": 3.599535406682782, + "tokens_seen": 1261568000 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031200601805416253, + "loss": 1.3396, + "theoretical_loss": 3.599535406682782, + "tokens_seen": 1261568000 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031199598796389166, + "loss": 1.4294, + "theoretical_loss": 3.5995185029516708, + "tokens_seen": 1261633536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003119859578736209, + "loss": 1.4814, + "theoretical_loss": 3.5995016003444524, + "tokens_seen": 1261699072 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031197592778335, + "loss": 1.3987, + "theoretical_loss": 3.5994846988609934, + "tokens_seen": 1261764608 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031196589769307925, + "loss": 1.3854, + "theoretical_loss": 3.59946779850116, + "tokens_seen": 1261830144 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031195586760280843, + "loss": 1.4949, + "theoretical_loss": 3.599450899264821, + "tokens_seen": 1261895680 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003119458375125376, + "loss": 1.2849, + "theoretical_loss": 3.599434001151842, + "tokens_seen": 1261961216 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003119358074222668, + "loss": 1.36, + "theoretical_loss": 3.5994171041620904, + "tokens_seen": 1262026752 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031192577733199603, + "loss": 1.4741, + "theoretical_loss": 3.5994002082954335, + "tokens_seen": 1262092288 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031191574724172516, + "loss": 1.3745, + "theoretical_loss": 3.599383313551737, + "tokens_seen": 1262157824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003119057171514544, + "loss": 1.3655, + "theoretical_loss": 3.59936641993087, + "tokens_seen": 1262223360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003118956870611835, + "loss": 1.4798, + "theoretical_loss": 3.599349527432698, + "tokens_seen": 1262288896 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031188565697091276, + "loss": 1.32, + "theoretical_loss": 3.5993326360570888, + "tokens_seen": 1262354432 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031187562688064194, + "loss": 1.4029, + "theoretical_loss": 3.599315745803909, + "tokens_seen": 1262419968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003118655967903711, + "loss": 1.4571, + "theoretical_loss": 3.599298856673027, + "tokens_seen": 1262485504 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003118555667001003, + "loss": 1.3026, + "theoretical_loss": 3.5992819686643083, + "tokens_seen": 1262551040 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003118455366098295, + "loss": 1.4348, + "theoretical_loss": 3.5992650817776215, + "tokens_seen": 1262616576 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031183550651955866, + "loss": 1.3811, + "theoretical_loss": 3.599248196012833, + "tokens_seen": 1262682112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003118254764292879, + "loss": 1.4167, + "theoretical_loss": 3.59923131136981, + "tokens_seen": 1262747648 + }, + { + "epoch": 0.38, + "learning_rate": 0.000311815446339017, + "loss": 1.4081, + "theoretical_loss": 3.5992144278484206, + "tokens_seen": 1262813184 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031180541624874626, + "loss": 1.4151, + "theoretical_loss": 3.599197545448531, + "tokens_seen": 1262878720 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117953861584754, + "loss": 1.4367, + "theoretical_loss": 3.5991806641700093, + "tokens_seen": 1262944256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117853560682046, + "loss": 1.4016, + "theoretical_loss": 3.599163784012722, + "tokens_seen": 1263009792 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117753259779338, + "loss": 1.3629, + "theoretical_loss": 3.5991469049765374, + "tokens_seen": 1263075328 + }, + { + "epoch": 0.38, + "learning_rate": 0.000311765295887663, + "loss": 1.3695, + "theoretical_loss": 3.5991300270613227, + "tokens_seen": 1263140864 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031175526579739217, + "loss": 1.4175, + "theoretical_loss": 3.5991131502669447, + "tokens_seen": 1263206400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117452357071214, + "loss": 1.3593, + "theoretical_loss": 3.599096274593272, + "tokens_seen": 1263271936 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031173520561685053, + "loss": 1.3831, + "theoretical_loss": 3.5990794000401705, + "tokens_seen": 1263337472 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031172517552657976, + "loss": 1.4838, + "theoretical_loss": 3.5990625266075087, + "tokens_seen": 1263403008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117151454363089, + "loss": 1.4259, + "theoretical_loss": 3.5990456542951534, + "tokens_seen": 1263468544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003117051153460381, + "loss": 1.4726, + "theoretical_loss": 3.5990287831029733, + "tokens_seen": 1263534080 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003116950852557673, + "loss": 1.4794, + "theoretical_loss": 3.5990119130308345, + "tokens_seen": 1263599616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003116850551654965, + "loss": 1.472, + "theoretical_loss": 3.5989950440786056, + "tokens_seen": 1263665152 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031167502507522567, + "loss": 1.4495, + "theoretical_loss": 3.5989781762461535, + "tokens_seen": 1263730688 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031166499498495485, + "loss": 1.42, + "theoretical_loss": 3.5989613095333466, + "tokens_seen": 1263796224 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031165496489468403, + "loss": 1.4349, + "theoretical_loss": 3.5989444439400518, + "tokens_seen": 1263861760 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031164493480441327, + "loss": 1.4243, + "theoretical_loss": 3.598927579466137, + "tokens_seen": 1263927296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003116349047141424, + "loss": 1.4855, + "theoretical_loss": 3.5989107161114697, + "tokens_seen": 1263992832 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031162487462387163, + "loss": 1.44, + "theoretical_loss": 3.598893853875918, + "tokens_seen": 1264058368 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003116148445336008, + "loss": 1.2626, + "theoretical_loss": 3.598876992759349, + "tokens_seen": 1264123904 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031160481444333, + "loss": 1.4506, + "theoretical_loss": 3.598860132761631, + "tokens_seen": 1264189440 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031159478435305917, + "loss": 1.4236, + "theoretical_loss": 3.5988432738826317, + "tokens_seen": 1264254976 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031158475426278835, + "loss": 1.4448, + "theoretical_loss": 3.5988264161222183, + "tokens_seen": 1264320512 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031157472417251753, + "loss": 1.423, + "theoretical_loss": 3.5988095594802596, + "tokens_seen": 1264386048 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031156469408224677, + "loss": 1.4031, + "theoretical_loss": 3.5987927039566228, + "tokens_seen": 1264451584 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031155466399197595, + "loss": 1.5109, + "theoretical_loss": 3.5987758495511755, + "tokens_seen": 1264517120 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031154463390170513, + "loss": 1.3509, + "theoretical_loss": 3.5987589962637863, + "tokens_seen": 1264582656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003115346038114343, + "loss": 1.4043, + "theoretical_loss": 3.5987421440943224, + "tokens_seen": 1264648192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003115245737211635, + "loss": 1.3513, + "theoretical_loss": 3.5987252930426523, + "tokens_seen": 1264713728 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031151454363089273, + "loss": 1.4769, + "theoretical_loss": 3.5987084431086434, + "tokens_seen": 1264779264 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 462829, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4572224617004395, + "objective/train/theoretical_loss": 3.5986915942921636, + "objective/train/tokens_used": 1285304800, + "theoretical_loss": 3.5986915942921636, + "tokens_seen": 1264844800 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031150451354062186, + "loss": 1.4231, + "theoretical_loss": 3.5986915942921636, + "tokens_seen": 1264844800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003114944834503511, + "loss": 1.3841, + "theoretical_loss": 3.598674746593082, + "tokens_seen": 1264910336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003114844533600802, + "loss": 1.303, + "theoretical_loss": 3.598657900011265, + "tokens_seen": 1264975872 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031147442326980945, + "loss": 1.4117, + "theoretical_loss": 3.598641054546582, + "tokens_seen": 1265041408 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031146439317953863, + "loss": 1.4525, + "theoretical_loss": 3.5986242101989006, + "tokens_seen": 1265106944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003114543630892678, + "loss": 1.3634, + "theoretical_loss": 3.5986073669680883, + "tokens_seen": 1265172480 + }, + { + "epoch": 0.38, + "learning_rate": 0.000311444332998997, + "loss": 1.4388, + "theoretical_loss": 3.598590524854014, + "tokens_seen": 1265238016 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031143430290872623, + "loss": 1.3648, + "theoretical_loss": 3.5985736838565456, + "tokens_seen": 1265303552 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031142427281845536, + "loss": 1.3363, + "theoretical_loss": 3.598556843975551, + "tokens_seen": 1265369088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003114142427281846, + "loss": 1.3747, + "theoretical_loss": 3.5985400052108982, + "tokens_seen": 1265434624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003114042126379137, + "loss": 1.4299, + "theoretical_loss": 3.598523167562456, + "tokens_seen": 1265500160 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031139418254764296, + "loss": 1.4722, + "theoretical_loss": 3.5985063310300918, + "tokens_seen": 1265565696 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031138415245737214, + "loss": 1.4721, + "theoretical_loss": 3.5984894956136753, + "tokens_seen": 1265631232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113741223671013, + "loss": 1.334, + "theoretical_loss": 3.598472661313073, + "tokens_seen": 1265696768 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113640922768305, + "loss": 1.3365, + "theoretical_loss": 3.598455828128154, + "tokens_seen": 1265762304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113540621865597, + "loss": 1.4304, + "theoretical_loss": 3.598438996058787, + "tokens_seen": 1265827840 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031134403209628886, + "loss": 1.3558, + "theoretical_loss": 3.5984221651048394, + "tokens_seen": 1265893376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113340020060181, + "loss": 1.4362, + "theoretical_loss": 3.5984053352661807, + "tokens_seen": 1265958912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113239719157472, + "loss": 1.4157, + "theoretical_loss": 3.598388506542678, + "tokens_seen": 1266024448 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031131394182547646, + "loss": 1.5065, + "theoretical_loss": 3.5983716789342, + "tokens_seen": 1266089984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003113039117352056, + "loss": 1.5961, + "theoretical_loss": 3.598354852440616, + "tokens_seen": 1266155520 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112938816449348, + "loss": 1.5145, + "theoretical_loss": 3.5983380270617937, + "tokens_seen": 1266221056 + }, + { + "epoch": 0.38, + "learning_rate": 0.000311283851554664, + "loss": 1.3181, + "theoretical_loss": 3.5983212027976013, + "tokens_seen": 1266286592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112738214643932, + "loss": 1.4683, + "theoretical_loss": 3.5983043796479084, + "tokens_seen": 1266352128 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031126379137412237, + "loss": 1.3504, + "theoretical_loss": 3.598287557612582, + "tokens_seen": 1266417664 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112537612838516, + "loss": 1.4783, + "theoretical_loss": 3.5982707366914917, + "tokens_seen": 1266483200 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031124373119358073, + "loss": 1.4012, + "theoretical_loss": 3.598253916884506, + "tokens_seen": 1266548736 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031123370110330996, + "loss": 1.3152, + "theoretical_loss": 3.5982370981914924, + "tokens_seen": 1266614272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112236710130391, + "loss": 1.4297, + "theoretical_loss": 3.5982202806123205, + "tokens_seen": 1266679808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112136409227683, + "loss": 1.4053, + "theoretical_loss": 3.5982034641468594, + "tokens_seen": 1266745344 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003112036108324975, + "loss": 1.3593, + "theoretical_loss": 3.598186648794976, + "tokens_seen": 1266810880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003111935807422267, + "loss": 1.3833, + "theoretical_loss": 3.5981698345565407, + "tokens_seen": 1266876416 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031118355065195587, + "loss": 1.4412, + "theoretical_loss": 3.5981530214314215, + "tokens_seen": 1266941952 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031117352056168505, + "loss": 1.3664, + "theoretical_loss": 3.5981362094194864, + "tokens_seen": 1267007488 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031116349047141423, + "loss": 1.4587, + "theoretical_loss": 3.5981193985206055, + "tokens_seen": 1267073024 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031115346038114347, + "loss": 1.4314, + "theoretical_loss": 3.5981025887346467, + "tokens_seen": 1267138560 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003111434302908726, + "loss": 1.4087, + "theoretical_loss": 3.5980857800614787, + "tokens_seen": 1267204096 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031113340020060183, + "loss": 1.5034, + "theoretical_loss": 3.5980689725009705, + "tokens_seen": 1267269632 + }, + { + "epoch": 0.38, + "learning_rate": 0.000311123370110331, + "loss": 1.5401, + "theoretical_loss": 3.598052166052991, + "tokens_seen": 1267335168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003111133400200602, + "loss": 1.3589, + "theoretical_loss": 3.5980353607174087, + "tokens_seen": 1267400704 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031110330992978937, + "loss": 1.4406, + "theoretical_loss": 3.5980185564940927, + "tokens_seen": 1267466240 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031109327983951855, + "loss": 1.4458, + "theoretical_loss": 3.598001753382912, + "tokens_seen": 1267531776 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031108324974924773, + "loss": 1.5062, + "theoretical_loss": 3.5979849513837356, + "tokens_seen": 1267597312 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031107321965897697, + "loss": 1.3149, + "theoretical_loss": 3.597968150496432, + "tokens_seen": 1267662848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003110631895687061, + "loss": 1.4989, + "theoretical_loss": 3.5979513507208702, + "tokens_seen": 1267728384 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031105315947843533, + "loss": 1.4029, + "theoretical_loss": 3.5979345520569193, + "tokens_seen": 1267793920 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031104312938816446, + "loss": 1.4301, + "theoretical_loss": 3.597917754504448, + "tokens_seen": 1267859456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003110330992978937, + "loss": 1.4356, + "theoretical_loss": 3.5979009580633265, + "tokens_seen": 1267924992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003110230692076229, + "loss": 1.3978, + "theoretical_loss": 3.5978841627334224, + "tokens_seen": 1267990528 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031101303911735206, + "loss": 1.482, + "theoretical_loss": 3.5978673685146054, + "tokens_seen": 1268056064 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 464032, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4879343509674072, + "objective/train/theoretical_loss": 3.5978505754067442, + "objective/train/tokens_used": 1288581600, + "theoretical_loss": 3.5978505754067442, + "tokens_seen": 1268121600 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031100300902708124, + "loss": 1.4564, + "theoretical_loss": 3.5978505754067442, + "tokens_seen": 1268121600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109929789368104, + "loss": 1.396, + "theoretical_loss": 3.5978337834097083, + "tokens_seen": 1268187136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109829488465396, + "loss": 1.379, + "theoretical_loss": 3.5978169925233665, + "tokens_seen": 1268252672 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031097291875626883, + "loss": 1.3969, + "theoretical_loss": 3.597800202747589, + "tokens_seen": 1268318208 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031096288866599796, + "loss": 1.496, + "theoretical_loss": 3.5977834140822433, + "tokens_seen": 1268383744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109528585757272, + "loss": 1.3117, + "theoretical_loss": 3.5977666265272, + "tokens_seen": 1268449280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109428284854564, + "loss": 1.4523, + "theoretical_loss": 3.5977498400823267, + "tokens_seen": 1268514816 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031093279839518556, + "loss": 1.5134, + "theoretical_loss": 3.5977330547474944, + "tokens_seen": 1268580352 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031092276830491474, + "loss": 1.4309, + "theoretical_loss": 3.5977162705225716, + "tokens_seen": 1268645888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109127382146439, + "loss": 1.4145, + "theoretical_loss": 3.5976994874074277, + "tokens_seen": 1268711424 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003109027081243731, + "loss": 1.4165, + "theoretical_loss": 3.5976827054019314, + "tokens_seen": 1268776960 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031089267803410234, + "loss": 1.3883, + "theoretical_loss": 3.597665924505953, + "tokens_seen": 1268842496 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031088264794383147, + "loss": 1.3603, + "theoretical_loss": 3.5976491447193606, + "tokens_seen": 1268908032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003108726178535607, + "loss": 1.5444, + "theoretical_loss": 3.597632366042025, + "tokens_seen": 1268973568 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031086258776328983, + "loss": 1.4146, + "theoretical_loss": 3.597615588473815, + "tokens_seen": 1269039104 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031085255767301906, + "loss": 1.4038, + "theoretical_loss": 3.597598812014599, + "tokens_seen": 1269104640 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031084252758274824, + "loss": 1.4089, + "theoretical_loss": 3.5975820366642486, + "tokens_seen": 1269170176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003108324974924774, + "loss": 1.3633, + "theoretical_loss": 3.5975652624226315, + "tokens_seen": 1269235712 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003108224674022066, + "loss": 1.4301, + "theoretical_loss": 3.5975484892896175, + "tokens_seen": 1269301248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003108124373119358, + "loss": 1.4135, + "theoretical_loss": 3.597531717265076, + "tokens_seen": 1269366784 + }, + { + "epoch": 0.38, + "learning_rate": 0.000310802407221665, + "loss": 1.49, + "theoretical_loss": 3.5975149463488774, + "tokens_seen": 1269432320 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107923771313942, + "loss": 1.3977, + "theoretical_loss": 3.59749817654089, + "tokens_seen": 1269497856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107823470411234, + "loss": 1.4493, + "theoretical_loss": 3.5974814078409847, + "tokens_seen": 1269563392 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031077231695085257, + "loss": 1.3314, + "theoretical_loss": 3.5974646402490302, + "tokens_seen": 1269628928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107622868605818, + "loss": 1.4495, + "theoretical_loss": 3.597447873764896, + "tokens_seen": 1269694464 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031075225677031093, + "loss": 1.3601, + "theoretical_loss": 3.5974311083884523, + "tokens_seen": 1269760000 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031074222668004016, + "loss": 1.3855, + "theoretical_loss": 3.597414344119569, + "tokens_seen": 1269825536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107321965897693, + "loss": 1.4619, + "theoretical_loss": 3.5973975809581145, + "tokens_seen": 1269891072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107221664994985, + "loss": 1.3489, + "theoretical_loss": 3.597380818903959, + "tokens_seen": 1269956608 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107121364092277, + "loss": 1.3945, + "theoretical_loss": 3.5973640579569732, + "tokens_seen": 1270022144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003107021063189569, + "loss": 1.4615, + "theoretical_loss": 3.5973472981170262, + "tokens_seen": 1270087680 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031069207622868607, + "loss": 1.4175, + "theoretical_loss": 3.5973305393839876, + "tokens_seen": 1270153216 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031068204613841525, + "loss": 1.395, + "theoretical_loss": 3.5973137817577268, + "tokens_seen": 1270218752 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031067201604814443, + "loss": 1.3139, + "theoretical_loss": 3.5972970252381145, + "tokens_seen": 1270284288 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031066198595787367, + "loss": 1.3279, + "theoretical_loss": 3.59728026982502, + "tokens_seen": 1270349824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0003106519558676028, + "loss": 1.4045, + "theoretical_loss": 3.5972635155183132, + "tokens_seen": 1270415360 + }, + { + "epoch": 0.38, + "learning_rate": 0.00031064192577733203, + "loss": 1.3817, + "theoretical_loss": 3.597246762317864, + "tokens_seen": 1270480896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003106318956870612, + "loss": 1.3674, + "theoretical_loss": 3.5972300102235426, + "tokens_seen": 1270546432 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003106218655967904, + "loss": 1.4151, + "theoretical_loss": 3.597213259235219, + "tokens_seen": 1270611968 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031061183550651957, + "loss": 1.3735, + "theoretical_loss": 3.5971965093527625, + "tokens_seen": 1270677504 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031060180541624875, + "loss": 1.3977, + "theoretical_loss": 3.597179760576043, + "tokens_seen": 1270743040 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031059177532597793, + "loss": 1.3739, + "theoretical_loss": 3.5971630129049315, + "tokens_seen": 1270808576 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031058174523570717, + "loss": 1.3559, + "theoretical_loss": 3.597146266339297, + "tokens_seen": 1270874112 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003105717151454363, + "loss": 1.3368, + "theoretical_loss": 3.59712952087901, + "tokens_seen": 1270939648 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031056168505516553, + "loss": 1.4316, + "theoretical_loss": 3.5971127765239403, + "tokens_seen": 1271005184 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031055165496489466, + "loss": 1.4556, + "theoretical_loss": 3.597096033273958, + "tokens_seen": 1271070720 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003105416248746239, + "loss": 1.5142, + "theoretical_loss": 3.5970792911289333, + "tokens_seen": 1271136256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003105315947843531, + "loss": 1.3746, + "theoretical_loss": 3.5970625500887365, + "tokens_seen": 1271201792 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031052156469408226, + "loss": 1.3158, + "theoretical_loss": 3.5970458101532374, + "tokens_seen": 1271267328 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031051153460381144, + "loss": 1.3857, + "theoretical_loss": 3.5970290713223063, + "tokens_seen": 1271332864 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 465264, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4322786331176758, + "objective/train/theoretical_loss": 3.5970123335958135, + "objective/train/tokens_used": 1291858400, + "theoretical_loss": 3.5970123335958135, + "tokens_seen": 1271398400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003105015045135406, + "loss": 1.4114, + "theoretical_loss": 3.5970123335958135, + "tokens_seen": 1271398400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003104914744232698, + "loss": 1.4619, + "theoretical_loss": 3.5969955969736294, + "tokens_seen": 1271463936 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031048144433299904, + "loss": 1.4608, + "theoretical_loss": 3.5969788614556237, + "tokens_seen": 1271529472 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031047141424272816, + "loss": 1.3528, + "theoretical_loss": 3.596962127041667, + "tokens_seen": 1271595008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003104613841524574, + "loss": 1.4107, + "theoretical_loss": 3.5969453937316294, + "tokens_seen": 1271660544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003104513540621866, + "loss": 1.4113, + "theoretical_loss": 3.5969286615253813, + "tokens_seen": 1271726080 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031044132397191576, + "loss": 1.421, + "theoretical_loss": 3.5969119304227926, + "tokens_seen": 1271791616 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031043129388164494, + "loss": 1.3924, + "theoretical_loss": 3.5968952004237345, + "tokens_seen": 1271857152 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003104212637913741, + "loss": 1.3934, + "theoretical_loss": 3.5968784715280764, + "tokens_seen": 1271922688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003104112337011033, + "loss": 1.416, + "theoretical_loss": 3.5968617437356896, + "tokens_seen": 1271988224 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031040120361083254, + "loss": 1.4364, + "theoretical_loss": 3.5968450170464434, + "tokens_seen": 1272053760 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031039117352056167, + "loss": 1.3401, + "theoretical_loss": 3.596828291460209, + "tokens_seen": 1272119296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003103811434302909, + "loss": 1.3507, + "theoretical_loss": 3.596811566976857, + "tokens_seen": 1272184832 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031037111334002003, + "loss": 1.422, + "theoretical_loss": 3.596794843596258, + "tokens_seen": 1272250368 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031036108324974926, + "loss": 1.4495, + "theoretical_loss": 3.5967781213182817, + "tokens_seen": 1272315904 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031035105315947844, + "loss": 1.4842, + "theoretical_loss": 3.5967614001427988, + "tokens_seen": 1272381440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003103410230692076, + "loss": 1.336, + "theoretical_loss": 3.59674468006968, + "tokens_seen": 1272446976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003103309929789368, + "loss": 1.5121, + "theoretical_loss": 3.5967279610987957, + "tokens_seen": 1272512512 + }, + { + "epoch": 0.39, + "learning_rate": 0.000310320962888666, + "loss": 1.3095, + "theoretical_loss": 3.596711243230017, + "tokens_seen": 1272578048 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031031093279839517, + "loss": 1.3611, + "theoretical_loss": 3.5966945264632137, + "tokens_seen": 1272643584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003103009027081244, + "loss": 1.3129, + "theoretical_loss": 3.5966778107982575, + "tokens_seen": 1272709120 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031029087261785353, + "loss": 1.4077, + "theoretical_loss": 3.596661096235018, + "tokens_seen": 1272774656 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031028084252758277, + "loss": 1.4647, + "theoretical_loss": 3.5966443827733663, + "tokens_seen": 1272840192 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031027081243731195, + "loss": 1.4486, + "theoretical_loss": 3.596627670413173, + "tokens_seen": 1272905728 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031026078234704113, + "loss": 1.3441, + "theoretical_loss": 3.596610959154309, + "tokens_seen": 1272971264 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003102507522567703, + "loss": 1.4725, + "theoretical_loss": 3.5965942489966447, + "tokens_seen": 1273036800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003102407221664995, + "loss": 1.4738, + "theoretical_loss": 3.5965775399400512, + "tokens_seen": 1273102336 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031023069207622867, + "loss": 1.2957, + "theoretical_loss": 3.596560831984399, + "tokens_seen": 1273167872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003102206619859579, + "loss": 1.4152, + "theoretical_loss": 3.596544125129559, + "tokens_seen": 1273233408 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031021063189568703, + "loss": 1.4745, + "theoretical_loss": 3.596527419375402, + "tokens_seen": 1273298944 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031020060180541627, + "loss": 1.4206, + "theoretical_loss": 3.5965107147217994, + "tokens_seen": 1273364480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003101905717151454, + "loss": 1.3788, + "theoretical_loss": 3.5964940111686206, + "tokens_seen": 1273430016 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031018054162487463, + "loss": 1.4067, + "theoretical_loss": 3.5964773087157385, + "tokens_seen": 1273495552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003101705115346038, + "loss": 1.3185, + "theoretical_loss": 3.5964606073630216, + "tokens_seen": 1273561088 + }, + { + "epoch": 0.39, + "learning_rate": 0.000310160481444333, + "loss": 1.4269, + "theoretical_loss": 3.596443907110343, + "tokens_seen": 1273626624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003101504513540622, + "loss": 1.548, + "theoretical_loss": 3.5964272079575728, + "tokens_seen": 1273692160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003101404212637914, + "loss": 1.457, + "theoretical_loss": 3.5964105099045813, + "tokens_seen": 1273757696 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031013039117352054, + "loss": 1.3719, + "theoretical_loss": 3.596393812951241, + "tokens_seen": 1273823232 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031012036108324977, + "loss": 1.4618, + "theoretical_loss": 3.5963771170974215, + "tokens_seen": 1273888768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003101103309929789, + "loss": 1.542, + "theoretical_loss": 3.596360422342995, + "tokens_seen": 1273954304 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031010030090270813, + "loss": 1.4391, + "theoretical_loss": 3.596343728687831, + "tokens_seen": 1274019840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100902708124373, + "loss": 1.4601, + "theoretical_loss": 3.596327036131802, + "tokens_seen": 1274085376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100802407221665, + "loss": 1.4884, + "theoretical_loss": 3.5963103446747784, + "tokens_seen": 1274150912 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100702106318957, + "loss": 1.3446, + "theoretical_loss": 3.596293654316632, + "tokens_seen": 1274216448 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031006018054162486, + "loss": 1.5692, + "theoretical_loss": 3.5962769650572337, + "tokens_seen": 1274281984 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100501504513541, + "loss": 1.3497, + "theoretical_loss": 3.5962602768964533, + "tokens_seen": 1274347520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100401203610833, + "loss": 1.3476, + "theoretical_loss": 3.596243589834164, + "tokens_seen": 1274413056 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031003009027081246, + "loss": 1.3258, + "theoretical_loss": 3.596226903870236, + "tokens_seen": 1274478592 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031002006018054164, + "loss": 1.3971, + "theoretical_loss": 3.596210219004541, + "tokens_seen": 1274544128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003100100300902708, + "loss": 1.4759, + "theoretical_loss": 3.5961935352369494, + "tokens_seen": 1274609664 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 466434, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3834418058395386, + "objective/train/theoretical_loss": 3.5961768525673334, + "objective/train/tokens_used": 1295135200, + "theoretical_loss": 3.5961768525673334, + "tokens_seen": 1274675200 + }, + { + "epoch": 0.39, + "learning_rate": 0.00031, + "loss": 1.3254, + "theoretical_loss": 3.5961768525673334, + "tokens_seen": 1274675200 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030998996990972924, + "loss": 1.4055, + "theoretical_loss": 3.5961601709955637, + "tokens_seen": 1274740736 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030997993981945836, + "loss": 1.4646, + "theoretical_loss": 3.596143490521512, + "tokens_seen": 1274806272 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003099699097291876, + "loss": 1.5053, + "theoretical_loss": 3.5961268111450493, + "tokens_seen": 1274871808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003099598796389168, + "loss": 1.4537, + "theoretical_loss": 3.5961101328660474, + "tokens_seen": 1274937344 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030994984954864596, + "loss": 1.4891, + "theoretical_loss": 3.596093455684377, + "tokens_seen": 1275002880 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030993981945837514, + "loss": 1.3959, + "theoretical_loss": 3.59607677959991, + "tokens_seen": 1275068416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003099297893681043, + "loss": 1.3865, + "theoretical_loss": 3.596060104612518, + "tokens_seen": 1275133952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003099197592778335, + "loss": 1.4626, + "theoretical_loss": 3.5960434307220717, + "tokens_seen": 1275199488 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030990972918756274, + "loss": 1.3874, + "theoretical_loss": 3.5960267579284437, + "tokens_seen": 1275265024 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030989969909729187, + "loss": 1.5232, + "theoretical_loss": 3.5960100862315043, + "tokens_seen": 1275330560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003098896690070211, + "loss": 1.4557, + "theoretical_loss": 3.5959934156311255, + "tokens_seen": 1275396096 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030987963891675023, + "loss": 1.5292, + "theoretical_loss": 3.595976746127179, + "tokens_seen": 1275461632 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030986960882647946, + "loss": 1.4013, + "theoretical_loss": 3.5959600777195364, + "tokens_seen": 1275527168 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030985957873620864, + "loss": 1.4166, + "theoretical_loss": 3.5959434104080694, + "tokens_seen": 1275592704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003098495486459378, + "loss": 1.4705, + "theoretical_loss": 3.5959267441926483, + "tokens_seen": 1275658240 + }, + { + "epoch": 0.39, + "learning_rate": 0.000309839518555667, + "loss": 1.4736, + "theoretical_loss": 3.5959100790731466, + "tokens_seen": 1275723776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003098294884653962, + "loss": 1.4155, + "theoretical_loss": 3.595893415049435, + "tokens_seen": 1275789312 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030981945837512537, + "loss": 1.3672, + "theoretical_loss": 3.5958767521213852, + "tokens_seen": 1275854848 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003098094282848546, + "loss": 1.4489, + "theoretical_loss": 3.595860090288869, + "tokens_seen": 1275920384 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030979939819458373, + "loss": 1.4643, + "theoretical_loss": 3.595843429551758, + "tokens_seen": 1275985920 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030978936810431297, + "loss": 1.3936, + "theoretical_loss": 3.595826769909924, + "tokens_seen": 1276051456 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030977933801404215, + "loss": 1.4379, + "theoretical_loss": 3.595810111363238, + "tokens_seen": 1276116992 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030976930792377133, + "loss": 1.4387, + "theoretical_loss": 3.5957934539115732, + "tokens_seen": 1276182528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003097592778335005, + "loss": 1.5125, + "theoretical_loss": 3.5957767975548007, + "tokens_seen": 1276248064 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003097492477432297, + "loss": 1.3953, + "theoretical_loss": 3.595760142292792, + "tokens_seen": 1276313600 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030973921765295887, + "loss": 1.295, + "theoretical_loss": 3.5957434881254198, + "tokens_seen": 1276379136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003097291875626881, + "loss": 1.4542, + "theoretical_loss": 3.595726835052555, + "tokens_seen": 1276444672 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030971915747241723, + "loss": 1.4196, + "theoretical_loss": 3.5957101830740696, + "tokens_seen": 1276510208 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030970912738214647, + "loss": 1.428, + "theoretical_loss": 3.595693532189836, + "tokens_seen": 1276575744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003096990972918756, + "loss": 1.4377, + "theoretical_loss": 3.595676882399726, + "tokens_seen": 1276641280 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030968906720160483, + "loss": 1.5336, + "theoretical_loss": 3.5956602337036117, + "tokens_seen": 1276706816 + }, + { + "epoch": 0.39, + "learning_rate": 0.000309679037111334, + "loss": 1.5528, + "theoretical_loss": 3.5956435861013643, + "tokens_seen": 1276772352 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003096690070210632, + "loss": 1.4042, + "theoretical_loss": 3.595626939592856, + "tokens_seen": 1276837888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003096589769307924, + "loss": 1.4646, + "theoretical_loss": 3.59561029417796, + "tokens_seen": 1276903424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003096489468405216, + "loss": 1.3338, + "theoretical_loss": 3.595593649856547, + "tokens_seen": 1276968960 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030963891675025074, + "loss": 1.4128, + "theoretical_loss": 3.5955770066284893, + "tokens_seen": 1277034496 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030962888665997997, + "loss": 1.3806, + "theoretical_loss": 3.595560364493659, + "tokens_seen": 1277100032 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003096188565697091, + "loss": 1.5343, + "theoretical_loss": 3.595543723451929, + "tokens_seen": 1277165568 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030960882647943833, + "loss": 1.3885, + "theoretical_loss": 3.5955270835031703, + "tokens_seen": 1277231104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095987963891675, + "loss": 1.3045, + "theoretical_loss": 3.595510444647256, + "tokens_seen": 1277296640 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095887662988967, + "loss": 1.378, + "theoretical_loss": 3.5954938068840576, + "tokens_seen": 1277362176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095787362086259, + "loss": 1.4227, + "theoretical_loss": 3.5954771702134476, + "tokens_seen": 1277427712 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030956870611835506, + "loss": 1.43, + "theoretical_loss": 3.5954605346352975, + "tokens_seen": 1277493248 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030955867602808424, + "loss": 1.4379, + "theoretical_loss": 3.5954439001494802, + "tokens_seen": 1277558784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095486459378135, + "loss": 1.4307, + "theoretical_loss": 3.595427266755868, + "tokens_seen": 1277624320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095386158475426, + "loss": 1.4678, + "theoretical_loss": 3.595410634454333, + "tokens_seen": 1277689856 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030952858575727184, + "loss": 1.4092, + "theoretical_loss": 3.5953940032447473, + "tokens_seen": 1277755392 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030951855566700096, + "loss": 1.3773, + "theoretical_loss": 3.5953773731269836, + "tokens_seen": 1277820928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003095085255767302, + "loss": 1.5036, + "theoretical_loss": 3.595360744100914, + "tokens_seen": 1277886464 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 467639, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.375532627105713, + "objective/train/theoretical_loss": 3.595344116166411, + "objective/train/tokens_used": 1298412000, + "theoretical_loss": 3.595344116166411, + "tokens_seen": 1277952000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003094984954864594, + "loss": 1.3989, + "theoretical_loss": 3.595344116166411, + "tokens_seen": 1277952000 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030948846539618856, + "loss": 1.4385, + "theoretical_loss": 3.5953274893233464, + "tokens_seen": 1278017536 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030947843530591774, + "loss": 1.388, + "theoretical_loss": 3.595310863571594, + "tokens_seen": 1278083072 + }, + { + "epoch": 0.39, + "learning_rate": 0.000309468405215647, + "loss": 1.449, + "theoretical_loss": 3.5952942389110243, + "tokens_seen": 1278148608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003094583751253761, + "loss": 1.4236, + "theoretical_loss": 3.5952776153415114, + "tokens_seen": 1278214144 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030944834503510534, + "loss": 1.4537, + "theoretical_loss": 3.5952609928629267, + "tokens_seen": 1278279680 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030943831494483447, + "loss": 1.3039, + "theoretical_loss": 3.5952443714751428, + "tokens_seen": 1278345216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003094282848545637, + "loss": 1.3453, + "theoretical_loss": 3.5952277511780326, + "tokens_seen": 1278410752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003094182547642929, + "loss": 1.5362, + "theoretical_loss": 3.5952111319714684, + "tokens_seen": 1278476288 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030940822467402207, + "loss": 1.4024, + "theoretical_loss": 3.5951945138553234, + "tokens_seen": 1278541824 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030939819458375125, + "loss": 1.4913, + "theoretical_loss": 3.595177896829469, + "tokens_seen": 1278607360 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030938816449348043, + "loss": 1.4538, + "theoretical_loss": 3.5951612808937785, + "tokens_seen": 1278672896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003093781344032096, + "loss": 1.461, + "theoretical_loss": 3.5951446660481245, + "tokens_seen": 1278738432 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030936810431293884, + "loss": 1.3886, + "theoretical_loss": 3.5951280522923796, + "tokens_seen": 1278803968 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030935807422266797, + "loss": 1.3474, + "theoretical_loss": 3.5951114396264163, + "tokens_seen": 1278869504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003093480441323972, + "loss": 1.3567, + "theoretical_loss": 3.595094828050107, + "tokens_seen": 1278935040 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030933801404212633, + "loss": 1.4996, + "theoretical_loss": 3.5950782175633256, + "tokens_seen": 1279000576 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030932798395185557, + "loss": 1.3598, + "theoretical_loss": 3.5950616081659437, + "tokens_seen": 1279066112 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030931795386158475, + "loss": 1.4532, + "theoretical_loss": 3.5950449998578335, + "tokens_seen": 1279131648 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030930792377131393, + "loss": 1.4518, + "theoretical_loss": 3.5950283926388695, + "tokens_seen": 1279197184 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030929789368104317, + "loss": 1.4051, + "theoretical_loss": 3.5950117865089233, + "tokens_seen": 1279262720 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030928786359077235, + "loss": 1.3466, + "theoretical_loss": 3.594995181467868, + "tokens_seen": 1279328256 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030927783350050153, + "loss": 1.3382, + "theoretical_loss": 3.594978577515576, + "tokens_seen": 1279393792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003092678034102307, + "loss": 1.4202, + "theoretical_loss": 3.5949619746519206, + "tokens_seen": 1279459328 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003092577733199599, + "loss": 1.3888, + "theoretical_loss": 3.594945372876775, + "tokens_seen": 1279524864 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030924774322968907, + "loss": 1.5224, + "theoretical_loss": 3.594928772190011, + "tokens_seen": 1279590400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003092377131394183, + "loss": 1.4714, + "theoretical_loss": 3.5949121725915028, + "tokens_seen": 1279655936 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030922768304914743, + "loss": 1.3634, + "theoretical_loss": 3.5948955740811224, + "tokens_seen": 1279721472 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030921765295887667, + "loss": 1.4764, + "theoretical_loss": 3.5948789766587437, + "tokens_seen": 1279787008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003092076228686058, + "loss": 1.3966, + "theoretical_loss": 3.594862380324238, + "tokens_seen": 1279852544 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030919759277833503, + "loss": 1.3831, + "theoretical_loss": 3.5948457850774798, + "tokens_seen": 1279918080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091875626880642, + "loss": 1.4172, + "theoretical_loss": 3.5948291909183423, + "tokens_seen": 1279983616 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091775325977934, + "loss": 1.4799, + "theoretical_loss": 3.5948125978466967, + "tokens_seen": 1280049152 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091675025075226, + "loss": 1.4738, + "theoretical_loss": 3.594796005862418, + "tokens_seen": 1280114688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091574724172518, + "loss": 1.3944, + "theoretical_loss": 3.5947794149653785, + "tokens_seen": 1280180224 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030914744232698094, + "loss": 1.453, + "theoretical_loss": 3.5947628251554518, + "tokens_seen": 1280245760 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030913741223671017, + "loss": 1.4124, + "theoretical_loss": 3.59474623643251, + "tokens_seen": 1280311296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091273821464393, + "loss": 1.478, + "theoretical_loss": 3.5947296487964264, + "tokens_seen": 1280376832 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030911735205616853, + "loss": 1.3821, + "theoretical_loss": 3.594713062247075, + "tokens_seen": 1280442368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003091073219658977, + "loss": 1.3454, + "theoretical_loss": 3.594696476784329, + "tokens_seen": 1280507904 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090972918756269, + "loss": 1.3723, + "theoretical_loss": 3.5946798924080605, + "tokens_seen": 1280573440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090872617853561, + "loss": 1.4693, + "theoretical_loss": 3.5946633091181437, + "tokens_seen": 1280638976 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030907723169508526, + "loss": 1.5052, + "theoretical_loss": 3.5946467269144513, + "tokens_seen": 1280704512 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030906720160481444, + "loss": 1.3218, + "theoretical_loss": 3.5946301457968577, + "tokens_seen": 1280770048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090571715145437, + "loss": 1.5233, + "theoretical_loss": 3.5946135657652345, + "tokens_seen": 1280835584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090471414242728, + "loss": 1.54, + "theoretical_loss": 3.5945969868194556, + "tokens_seen": 1280901120 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030903711133400204, + "loss": 1.4335, + "theoretical_loss": 3.594580408959395, + "tokens_seen": 1280966656 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030902708124373116, + "loss": 1.5103, + "theoretical_loss": 3.5945638321849254, + "tokens_seen": 1281032192 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090170511534604, + "loss": 1.4355, + "theoretical_loss": 3.594547256495921, + "tokens_seen": 1281097728 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003090070210631896, + "loss": 1.3887, + "theoretical_loss": 3.594530681892254, + "tokens_seen": 1281163264 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 468763, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5929269790649414, + "objective/train/theoretical_loss": 3.594514108373799, + "objective/train/tokens_used": 1301688800, + "theoretical_loss": 3.594514108373799, + "tokens_seen": 1281228800 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030899699097291876, + "loss": 1.5103, + "theoretical_loss": 3.594514108373799, + "tokens_seen": 1281228800 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030898696088264794, + "loss": 1.4639, + "theoretical_loss": 3.5944975359404285, + "tokens_seen": 1281294336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003089769307923772, + "loss": 1.395, + "theoretical_loss": 3.594480964592016, + "tokens_seen": 1281359872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003089669007021063, + "loss": 1.5457, + "theoretical_loss": 3.594464394328436, + "tokens_seen": 1281425408 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030895687061183554, + "loss": 1.3873, + "theoretical_loss": 3.594447825149561, + "tokens_seen": 1281490944 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030894684052156467, + "loss": 1.4853, + "theoretical_loss": 3.5944312570552643, + "tokens_seen": 1281556480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003089368104312939, + "loss": 1.4074, + "theoretical_loss": 3.594414690045421, + "tokens_seen": 1281622016 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003089267803410231, + "loss": 1.4182, + "theoretical_loss": 3.594398124119903, + "tokens_seen": 1281687552 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030891675025075227, + "loss": 1.3768, + "theoretical_loss": 3.5943815592785846, + "tokens_seen": 1281753088 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030890672016048145, + "loss": 1.4815, + "theoretical_loss": 3.5943649955213397, + "tokens_seen": 1281818624 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030889669007021063, + "loss": 1.3932, + "theoretical_loss": 3.594348432848041, + "tokens_seen": 1281884160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003088866599799398, + "loss": 1.3888, + "theoretical_loss": 3.5943318712585635, + "tokens_seen": 1281949696 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030887662988966904, + "loss": 1.4115, + "theoretical_loss": 3.59431531075278, + "tokens_seen": 1282015232 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030886659979939817, + "loss": 1.4713, + "theoretical_loss": 3.594298751330564, + "tokens_seen": 1282080768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003088565697091274, + "loss": 1.4116, + "theoretical_loss": 3.5942821929917894, + "tokens_seen": 1282146304 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030884653961885653, + "loss": 1.4442, + "theoretical_loss": 3.5942656357363307, + "tokens_seen": 1282211840 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030883650952858577, + "loss": 1.4339, + "theoretical_loss": 3.5942490795640607, + "tokens_seen": 1282277376 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030882647943831495, + "loss": 1.3759, + "theoretical_loss": 3.5942325244748536, + "tokens_seen": 1282342912 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030881644934804413, + "loss": 1.388, + "theoretical_loss": 3.594215970468583, + "tokens_seen": 1282408448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003088064192577733, + "loss": 1.4045, + "theoretical_loss": 3.594199417545123, + "tokens_seen": 1282473984 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030879638916750255, + "loss": 1.2827, + "theoretical_loss": 3.5941828657043473, + "tokens_seen": 1282539520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003087863590772317, + "loss": 1.5216, + "theoretical_loss": 3.5941663149461296, + "tokens_seen": 1282605056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003087763289869609, + "loss": 1.2653, + "theoretical_loss": 3.594149765270344, + "tokens_seen": 1282670592 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030876629889669004, + "loss": 1.3181, + "theoretical_loss": 3.5941332166768647, + "tokens_seen": 1282736128 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030875626880641927, + "loss": 1.3116, + "theoretical_loss": 3.594116669165565, + "tokens_seen": 1282801664 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030874623871614845, + "loss": 1.2402, + "theoretical_loss": 3.594100122736319, + "tokens_seen": 1282867200 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030873620862587763, + "loss": 1.3772, + "theoretical_loss": 3.594083577389001, + "tokens_seen": 1282932736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003087261785356068, + "loss": 1.4007, + "theoretical_loss": 3.594067033123485, + "tokens_seen": 1282998272 + }, + { + "epoch": 0.39, + "learning_rate": 0.000308716148445336, + "loss": 1.3369, + "theoretical_loss": 3.5940504899396446, + "tokens_seen": 1283063808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003087061183550652, + "loss": 1.4474, + "theoretical_loss": 3.594033947837354, + "tokens_seen": 1283129344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086960882647944, + "loss": 1.3726, + "theoretical_loss": 3.5940174068164876, + "tokens_seen": 1283194880 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030868605817452354, + "loss": 1.4281, + "theoretical_loss": 3.594000866876919, + "tokens_seen": 1283260416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086760280842528, + "loss": 1.35, + "theoretical_loss": 3.5939843280185224, + "tokens_seen": 1283325952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086659979939819, + "loss": 1.3989, + "theoretical_loss": 3.593967790241172, + "tokens_seen": 1283391488 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030865596790371114, + "loss": 1.3295, + "theoretical_loss": 3.593951253544742, + "tokens_seen": 1283457024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086459378134403, + "loss": 1.3655, + "theoretical_loss": 3.593934717929107, + "tokens_seen": 1283522560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086359077231695, + "loss": 1.3551, + "theoretical_loss": 3.5939181833941403, + "tokens_seen": 1283588096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086258776328987, + "loss": 1.4508, + "theoretical_loss": 3.593901649939716, + "tokens_seen": 1283653632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003086158475426279, + "loss": 1.455, + "theoretical_loss": 3.5938851175657094, + "tokens_seen": 1283719168 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030860581745235704, + "loss": 1.36, + "theoretical_loss": 3.593868586271994, + "tokens_seen": 1283784704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003085957873620863, + "loss": 1.574, + "theoretical_loss": 3.593852056058444, + "tokens_seen": 1283850240 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003085857572718154, + "loss": 1.4189, + "theoretical_loss": 3.5938355269249347, + "tokens_seen": 1283915776 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030857572718154464, + "loss": 1.2831, + "theoretical_loss": 3.593818998871339, + "tokens_seen": 1283981312 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003085656970912738, + "loss": 1.382, + "theoretical_loss": 3.593802471897532, + "tokens_seen": 1284046848 + }, + { + "epoch": 0.39, + "learning_rate": 0.000308555667001003, + "loss": 1.4538, + "theoretical_loss": 3.593785946003388, + "tokens_seen": 1284112384 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030854563691073224, + "loss": 1.3863, + "theoretical_loss": 3.593769421188781, + "tokens_seen": 1284177920 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030853560682046137, + "loss": 1.3945, + "theoretical_loss": 3.5937528974535864, + "tokens_seen": 1284243456 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003085255767301906, + "loss": 1.448, + "theoretical_loss": 3.593736374797677, + "tokens_seen": 1284308992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003085155466399198, + "loss": 1.3106, + "theoretical_loss": 3.5937198532209287, + "tokens_seen": 1284374528 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030850551654964896, + "loss": 1.4984, + "theoretical_loss": 3.593703332723215, + "tokens_seen": 1284440064 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 470107, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4420971870422363, + "objective/train/theoretical_loss": 3.5936868133044113, + "objective/train/tokens_used": 1304965600, + "theoretical_loss": 3.5936868133044113, + "tokens_seen": 1284505600 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030849548645937814, + "loss": 1.3284, + "theoretical_loss": 3.5936868133044113, + "tokens_seen": 1284505600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003084854563691074, + "loss": 1.4344, + "theoretical_loss": 3.593670294964391, + "tokens_seen": 1284571136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003084754262788365, + "loss": 1.4721, + "theoretical_loss": 3.5936537777030293, + "tokens_seen": 1284636672 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030846539618856574, + "loss": 1.3492, + "theoretical_loss": 3.593637261520201, + "tokens_seen": 1284702208 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030845536609829487, + "loss": 1.4774, + "theoretical_loss": 3.59362074641578, + "tokens_seen": 1284767744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003084453360080241, + "loss": 1.3893, + "theoretical_loss": 3.593604232389641, + "tokens_seen": 1284833280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003084353059177533, + "loss": 1.3721, + "theoretical_loss": 3.593587719441659, + "tokens_seen": 1284898816 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030842527582748247, + "loss": 1.3432, + "theoretical_loss": 3.593571207571708, + "tokens_seen": 1284964352 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030841524573721165, + "loss": 1.3689, + "theoretical_loss": 3.593554696779664, + "tokens_seen": 1285029888 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030840521564694083, + "loss": 1.3561, + "theoretical_loss": 3.5935381870653997, + "tokens_seen": 1285095424 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030839518555667, + "loss": 1.4272, + "theoretical_loss": 3.5935216784287913, + "tokens_seen": 1285160960 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030838515546639924, + "loss": 1.3794, + "theoretical_loss": 3.5935051708697126, + "tokens_seen": 1285226496 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030837512537612837, + "loss": 1.4162, + "theoretical_loss": 3.5934886643880395, + "tokens_seen": 1285292032 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003083650952858576, + "loss": 1.4172, + "theoretical_loss": 3.593472158983645, + "tokens_seen": 1285357568 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030835506519558673, + "loss": 1.4866, + "theoretical_loss": 3.5934556546564056, + "tokens_seen": 1285423104 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030834503510531597, + "loss": 1.4353, + "theoretical_loss": 3.593439151406195, + "tokens_seen": 1285488640 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030833500501504515, + "loss": 1.3726, + "theoretical_loss": 3.5934226492328882, + "tokens_seen": 1285554176 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030832497492477433, + "loss": 1.4543, + "theoretical_loss": 3.593406148136361, + "tokens_seen": 1285619712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003083149448345035, + "loss": 1.4374, + "theoretical_loss": 3.593389648116487, + "tokens_seen": 1285685248 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030830491474423275, + "loss": 1.5442, + "theoretical_loss": 3.593373149173141, + "tokens_seen": 1285750784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003082948846539619, + "loss": 1.4145, + "theoretical_loss": 3.5933566513061987, + "tokens_seen": 1285816320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003082848545636911, + "loss": 1.321, + "theoretical_loss": 3.593340154515535, + "tokens_seen": 1285881856 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030827482447342024, + "loss": 1.3876, + "theoretical_loss": 3.5933236588010242, + "tokens_seen": 1285947392 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030826479438314947, + "loss": 1.4258, + "theoretical_loss": 3.5933071641625416, + "tokens_seen": 1286012928 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030825476429287865, + "loss": 1.4214, + "theoretical_loss": 3.5932906705999628, + "tokens_seen": 1286078464 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030824473420260783, + "loss": 1.544, + "theoretical_loss": 3.5932741781131616, + "tokens_seen": 1286144000 + }, + { + "epoch": 0.39, + "learning_rate": 0.000308234704112337, + "loss": 1.3557, + "theoretical_loss": 3.593257686702014, + "tokens_seen": 1286209536 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003082246740220662, + "loss": 1.3468, + "theoretical_loss": 3.5932411963663946, + "tokens_seen": 1286275072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003082146439317954, + "loss": 1.5142, + "theoretical_loss": 3.5932247071061783, + "tokens_seen": 1286340608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003082046138415246, + "loss": 1.3798, + "theoretical_loss": 3.593208218921241, + "tokens_seen": 1286406144 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030819458375125374, + "loss": 1.2842, + "theoretical_loss": 3.593191731811457, + "tokens_seen": 1286471680 + }, + { + "epoch": 0.39, + "learning_rate": 0.000308184553660983, + "loss": 1.4451, + "theoretical_loss": 3.5931752457767008, + "tokens_seen": 1286537216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081745235707121, + "loss": 1.4317, + "theoretical_loss": 3.5931587608168494, + "tokens_seen": 1286602752 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030816449348044134, + "loss": 1.3781, + "theoretical_loss": 3.5931422769317765, + "tokens_seen": 1286668288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081544633901705, + "loss": 1.3838, + "theoretical_loss": 3.593125794121358, + "tokens_seen": 1286733824 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081444332998997, + "loss": 1.4337, + "theoretical_loss": 3.5931093123854687, + "tokens_seen": 1286799360 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081344032096289, + "loss": 1.4322, + "theoretical_loss": 3.5930928317239843, + "tokens_seen": 1286864896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081243731193581, + "loss": 1.527, + "theoretical_loss": 3.5930763521367797, + "tokens_seen": 1286930432 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030811434302908724, + "loss": 1.4124, + "theoretical_loss": 3.59305987362373, + "tokens_seen": 1286995968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003081043129388165, + "loss": 1.3057, + "theoretical_loss": 3.5930433961847106, + "tokens_seen": 1287061504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003080942828485456, + "loss": 1.4561, + "theoretical_loss": 3.5930269198195974, + "tokens_seen": 1287127040 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030808425275827484, + "loss": 1.302, + "theoretical_loss": 3.593010444528265, + "tokens_seen": 1287192576 + }, + { + "epoch": 0.39, + "learning_rate": 0.000308074222668004, + "loss": 1.4027, + "theoretical_loss": 3.592993970310589, + "tokens_seen": 1287258112 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003080641925777332, + "loss": 1.4424, + "theoretical_loss": 3.5929774971664448, + "tokens_seen": 1287323648 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003080541624874624, + "loss": 1.481, + "theoretical_loss": 3.5929610250957076, + "tokens_seen": 1287389184 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030804413239719157, + "loss": 1.4728, + "theoretical_loss": 3.5929445540982536, + "tokens_seen": 1287454720 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030803410230692075, + "loss": 1.452, + "theoretical_loss": 3.5929280841739573, + "tokens_seen": 1287520256 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030802407221665, + "loss": 1.2967, + "theoretical_loss": 3.592911615322694, + "tokens_seen": 1287585792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003080140421263791, + "loss": 1.3805, + "theoretical_loss": 3.5928951475443403, + "tokens_seen": 1287651328 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030800401203610834, + "loss": 1.3129, + "theoretical_loss": 3.592878680838771, + "tokens_seen": 1287716864 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 471101, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3295934200286865, + "objective/train/theoretical_loss": 3.592862215205862, + "objective/train/tokens_used": 1308242400, + "theoretical_loss": 3.592862215205862, + "tokens_seen": 1287782400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079939819458375, + "loss": 1.4264, + "theoretical_loss": 3.592862215205862, + "tokens_seen": 1287782400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079839518555667, + "loss": 1.3445, + "theoretical_loss": 3.592845750645488, + "tokens_seen": 1287847936 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079739217652959, + "loss": 1.3913, + "theoretical_loss": 3.5928292871575254, + "tokens_seen": 1287913472 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030796389167502507, + "loss": 1.3753, + "theoretical_loss": 3.5928128247418494, + "tokens_seen": 1287979008 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030795386158475425, + "loss": 1.389, + "theoretical_loss": 3.5927963633983357, + "tokens_seen": 1288044544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079438314944835, + "loss": 1.4904, + "theoretical_loss": 3.59277990312686, + "tokens_seen": 1288110080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079338014042126, + "loss": 1.2977, + "theoretical_loss": 3.592763443927298, + "tokens_seen": 1288175616 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030792377131394185, + "loss": 1.3264, + "theoretical_loss": 3.592746985799525, + "tokens_seen": 1288241152 + }, + { + "epoch": 0.39, + "learning_rate": 0.000307913741223671, + "loss": 1.4901, + "theoretical_loss": 3.5927305287434175, + "tokens_seen": 1288306688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003079037111334002, + "loss": 1.4039, + "theoretical_loss": 3.59271407275885, + "tokens_seen": 1288372224 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078936810431294, + "loss": 1.3788, + "theoretical_loss": 3.5926976178456993, + "tokens_seen": 1288437760 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030788365095285857, + "loss": 1.3015, + "theoretical_loss": 3.5926811640038405, + "tokens_seen": 1288503296 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030787362086258775, + "loss": 1.3673, + "theoretical_loss": 3.5926647112331493, + "tokens_seen": 1288568832 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030786359077231693, + "loss": 1.351, + "theoretical_loss": 3.5926482595335023, + "tokens_seen": 1288634368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078535606820461, + "loss": 1.3839, + "theoretical_loss": 3.5926318089047746, + "tokens_seen": 1288699904 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030784353059177535, + "loss": 1.3611, + "theoretical_loss": 3.5926153593468424, + "tokens_seen": 1288765440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078335005015045, + "loss": 1.3964, + "theoretical_loss": 3.5925989108595813, + "tokens_seen": 1288830976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078234704112337, + "loss": 1.3587, + "theoretical_loss": 3.5925824634428674, + "tokens_seen": 1288896512 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078134403209629, + "loss": 1.419, + "theoretical_loss": 3.5925660170965763, + "tokens_seen": 1288962048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003078034102306921, + "loss": 1.3282, + "theoretical_loss": 3.5925495718205838, + "tokens_seen": 1289027584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003077933801404213, + "loss": 1.2527, + "theoretical_loss": 3.5925331276147663, + "tokens_seen": 1289093120 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030778335005015044, + "loss": 1.3458, + "theoretical_loss": 3.5925166844789995, + "tokens_seen": 1289158656 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030777331995987967, + "loss": 1.4937, + "theoretical_loss": 3.59250024241316, + "tokens_seen": 1289224192 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030776328986960885, + "loss": 1.4139, + "theoretical_loss": 3.5924838014171225, + "tokens_seen": 1289289728 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030775325977933803, + "loss": 1.3663, + "theoretical_loss": 3.592467361490764, + "tokens_seen": 1289355264 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003077432296890672, + "loss": 1.5283, + "theoretical_loss": 3.5924509226339607, + "tokens_seen": 1289420800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003077331995987964, + "loss": 1.4562, + "theoretical_loss": 3.5924344848465877, + "tokens_seen": 1289486336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003077231695085256, + "loss": 1.3294, + "theoretical_loss": 3.592418048128522, + "tokens_seen": 1289551872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003077131394182548, + "loss": 1.3883, + "theoretical_loss": 3.592401612479639, + "tokens_seen": 1289617408 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030770310932798394, + "loss": 1.3781, + "theoretical_loss": 3.5923851778998155, + "tokens_seen": 1289682944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076930792377132, + "loss": 1.3513, + "theoretical_loss": 3.5923687443889274, + "tokens_seen": 1289748480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076830491474423, + "loss": 1.3721, + "theoretical_loss": 3.5923523119468506, + "tokens_seen": 1289814016 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030767301905717154, + "loss": 1.3114, + "theoretical_loss": 3.5923358805734615, + "tokens_seen": 1289879552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076629889669007, + "loss": 1.3772, + "theoretical_loss": 3.5923194502686364, + "tokens_seen": 1289945088 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076529588766299, + "loss": 1.3949, + "theoretical_loss": 3.592303021032251, + "tokens_seen": 1290010624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076429287863591, + "loss": 1.4324, + "theoretical_loss": 3.5922865928641823, + "tokens_seen": 1290076160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076328986960883, + "loss": 1.4938, + "theoretical_loss": 3.592270165764306, + "tokens_seen": 1290141696 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030762286860581744, + "loss": 1.3389, + "theoretical_loss": 3.5922537397324987, + "tokens_seen": 1290207232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076128385155467, + "loss": 1.4498, + "theoretical_loss": 3.592237314768637, + "tokens_seen": 1290272768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003076028084252758, + "loss": 1.2659, + "theoretical_loss": 3.592220890872596, + "tokens_seen": 1290338304 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030759277833500504, + "loss": 1.429, + "theoretical_loss": 3.5922044680442533, + "tokens_seen": 1290403840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075827482447342, + "loss": 1.3788, + "theoretical_loss": 3.5921880462834848, + "tokens_seen": 1290469376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075727181544634, + "loss": 1.4261, + "theoretical_loss": 3.592171625590167, + "tokens_seen": 1290534912 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075626880641926, + "loss": 1.3653, + "theoretical_loss": 3.5921552059641764, + "tokens_seen": 1290600448 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030755265797392177, + "loss": 1.2655, + "theoretical_loss": 3.592138787405389, + "tokens_seen": 1290665984 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030754262788365095, + "loss": 1.3597, + "theoretical_loss": 3.5921223699136817, + "tokens_seen": 1290731520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075325977933802, + "loss": 1.334, + "theoretical_loss": 3.592105953488931, + "tokens_seen": 1290797056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075225677031093, + "loss": 1.2665, + "theoretical_loss": 3.592089538131013, + "tokens_seen": 1290862592 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030751253761283854, + "loss": 1.5482, + "theoretical_loss": 3.5920731238398043, + "tokens_seen": 1290928128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003075025075225677, + "loss": 1.3838, + "theoretical_loss": 3.5920567106151817, + "tokens_seen": 1290993664 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 472355, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4921298027038574, + "objective/train/theoretical_loss": 3.5920402984570217, + "objective/train/tokens_used": 1311519200, + "theoretical_loss": 3.5920402984570217, + "tokens_seen": 1291059200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074924774322969, + "loss": 1.4229, + "theoretical_loss": 3.5920402984570217, + "tokens_seen": 1291059200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074824473420261, + "loss": 1.431, + "theoretical_loss": 3.5920238873652006, + "tokens_seen": 1291124736 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030747241725175527, + "loss": 1.3105, + "theoretical_loss": 3.592007477339595, + "tokens_seen": 1291190272 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030746238716148445, + "loss": 1.3975, + "theoretical_loss": 3.591991068380082, + "tokens_seen": 1291255808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074523570712137, + "loss": 1.4301, + "theoretical_loss": 3.591974660486538, + "tokens_seen": 1291321344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074423269809428, + "loss": 1.4087, + "theoretical_loss": 3.5919582536588397, + "tokens_seen": 1291386880 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030743229689067205, + "loss": 1.3557, + "theoretical_loss": 3.591941847896863, + "tokens_seen": 1291452416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074222668004012, + "loss": 1.3586, + "theoretical_loss": 3.591925443200486, + "tokens_seen": 1291517952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074122367101304, + "loss": 1.4516, + "theoretical_loss": 3.5919090395695843, + "tokens_seen": 1291583488 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003074022066198596, + "loss": 1.4385, + "theoretical_loss": 3.5918926370040354, + "tokens_seen": 1291649024 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030739217652958877, + "loss": 1.4207, + "theoretical_loss": 3.5918762355037153, + "tokens_seen": 1291714560 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030738214643931795, + "loss": 1.4191, + "theoretical_loss": 3.5918598350685014, + "tokens_seen": 1291780096 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030737211634904713, + "loss": 1.2971, + "theoretical_loss": 3.5918434356982702, + "tokens_seen": 1291845632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003073620862587763, + "loss": 1.3921, + "theoretical_loss": 3.591827037392899, + "tokens_seen": 1291911168 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030735205616850555, + "loss": 1.3943, + "theoretical_loss": 3.5918106401522634, + "tokens_seen": 1291976704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003073420260782347, + "loss": 1.2751, + "theoretical_loss": 3.5917942439762416, + "tokens_seen": 1292042240 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003073319959879639, + "loss": 1.3703, + "theoretical_loss": 3.59177784886471, + "tokens_seen": 1292107776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003073219658976931, + "loss": 1.5064, + "theoretical_loss": 3.5917614548175454, + "tokens_seen": 1292173312 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003073119358074223, + "loss": 1.4084, + "theoretical_loss": 3.591745061834625, + "tokens_seen": 1292238848 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030730190571715146, + "loss": 1.4889, + "theoretical_loss": 3.5917286699158253, + "tokens_seen": 1292304384 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030729187562688064, + "loss": 1.4584, + "theoretical_loss": 3.5917122790610234, + "tokens_seen": 1292369920 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003072818455366098, + "loss": 1.4119, + "theoretical_loss": 3.5916958892700968, + "tokens_seen": 1292435456 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030727181544633905, + "loss": 1.4493, + "theoretical_loss": 3.591679500542922, + "tokens_seen": 1292500992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003072617853560682, + "loss": 1.418, + "theoretical_loss": 3.5916631128793757, + "tokens_seen": 1292566528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003072517552657974, + "loss": 1.3376, + "theoretical_loss": 3.591646726279336, + "tokens_seen": 1292632064 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030724172517552654, + "loss": 1.4055, + "theoretical_loss": 3.5916303407426797, + "tokens_seen": 1292697600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003072316950852558, + "loss": 1.4003, + "theoretical_loss": 3.5916139562692826, + "tokens_seen": 1292763136 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030722166499498496, + "loss": 1.2687, + "theoretical_loss": 3.5915975728590235, + "tokens_seen": 1292828672 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030721163490471414, + "loss": 1.4359, + "theoretical_loss": 3.5915811905117785, + "tokens_seen": 1292894208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003072016048144433, + "loss": 1.4806, + "theoretical_loss": 3.5915648092274246, + "tokens_seen": 1292959744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071915747241725, + "loss": 1.4174, + "theoretical_loss": 3.59154842900584, + "tokens_seen": 1293025280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071815446339017, + "loss": 1.3585, + "theoretical_loss": 3.5915320498469008, + "tokens_seen": 1293090816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071715145436309, + "loss": 1.3674, + "theoretical_loss": 3.5915156717504853, + "tokens_seen": 1293156352 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030716148445336005, + "loss": 1.3409, + "theoretical_loss": 3.5914992947164697, + "tokens_seen": 1293221888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071514543630893, + "loss": 1.3279, + "theoretical_loss": 3.591482918744732, + "tokens_seen": 1293287424 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030714142427281846, + "loss": 1.3569, + "theoretical_loss": 3.591466543835149, + "tokens_seen": 1293352960 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030713139418254764, + "loss": 1.4348, + "theoretical_loss": 3.5914501699875983, + "tokens_seen": 1293418496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071213640922768, + "loss": 1.3265, + "theoretical_loss": 3.5914337972019563, + "tokens_seen": 1293484032 + }, + { + "epoch": 0.39, + "learning_rate": 0.000307111334002006, + "loss": 1.4204, + "theoretical_loss": 3.591417425478102, + "tokens_seen": 1293549568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003071013039117352, + "loss": 1.4568, + "theoretical_loss": 3.591401054815911, + "tokens_seen": 1293615104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070912738214644, + "loss": 1.3909, + "theoretical_loss": 3.591384685215263, + "tokens_seen": 1293680640 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030708124373119355, + "loss": 1.4215, + "theoretical_loss": 3.5913683166760326, + "tokens_seen": 1293746176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070712136409228, + "loss": 1.4508, + "theoretical_loss": 3.591351949198099, + "tokens_seen": 1293811712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070611835506519, + "loss": 1.3758, + "theoretical_loss": 3.591335582781339, + "tokens_seen": 1293877248 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030705115346038115, + "loss": 1.5769, + "theoretical_loss": 3.5913192174256303, + "tokens_seen": 1293942784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070411233701104, + "loss": 1.3515, + "theoretical_loss": 3.5913028531308506, + "tokens_seen": 1294008320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070310932798395, + "loss": 1.4708, + "theoretical_loss": 3.591286489896877, + "tokens_seen": 1294073856 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030702106318956874, + "loss": 1.4523, + "theoretical_loss": 3.591270127723587, + "tokens_seen": 1294139392 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070110330992979, + "loss": 1.3873, + "theoretical_loss": 3.5912537666108584, + "tokens_seen": 1294204928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003070010030090271, + "loss": 1.4901, + "theoretical_loss": 3.591237406558568, + "tokens_seen": 1294270464 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 473604, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4121172428131104, + "objective/train/theoretical_loss": 3.591221047566594, + "objective/train/tokens_used": 1314796000, + "theoretical_loss": 3.591221047566594, + "tokens_seen": 1294336000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003069909729187563, + "loss": 1.3692, + "theoretical_loss": 3.591221047566594, + "tokens_seen": 1294336000 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030698094282848547, + "loss": 1.3759, + "theoretical_loss": 3.591204689634815, + "tokens_seen": 1294401536 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030697091273821465, + "loss": 1.4991, + "theoretical_loss": 3.5911883327631067, + "tokens_seen": 1294467072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003069608826479439, + "loss": 1.384, + "theoretical_loss": 3.5911719769513484, + "tokens_seen": 1294532608 + }, + { + "epoch": 0.39, + "learning_rate": 0.000306950852557673, + "loss": 1.378, + "theoretical_loss": 3.591155622199416, + "tokens_seen": 1294598144 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030694082246740225, + "loss": 1.3573, + "theoretical_loss": 3.591139268507189, + "tokens_seen": 1294663680 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003069307923771314, + "loss": 1.4551, + "theoretical_loss": 3.5911229158745437, + "tokens_seen": 1294729216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003069207622868606, + "loss": 1.3353, + "theoretical_loss": 3.5911065643013593, + "tokens_seen": 1294794752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003069107321965898, + "loss": 1.4002, + "theoretical_loss": 3.5910902137875116, + "tokens_seen": 1294860288 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030690070210631897, + "loss": 1.363, + "theoretical_loss": 3.59107386433288, + "tokens_seen": 1294925824 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030689067201604815, + "loss": 1.3672, + "theoretical_loss": 3.591057515937342, + "tokens_seen": 1294991360 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030688064192577733, + "loss": 1.5573, + "theoretical_loss": 3.591041168600774, + "tokens_seen": 1295056896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003068706118355065, + "loss": 1.4008, + "theoretical_loss": 3.591024822323056, + "tokens_seen": 1295122432 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030686058174523575, + "loss": 1.4396, + "theoretical_loss": 3.5910084771040642, + "tokens_seen": 1295187968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003068505516549649, + "loss": 1.3712, + "theoretical_loss": 3.590992132943677, + "tokens_seen": 1295253504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003068405215646941, + "loss": 1.5127, + "theoretical_loss": 3.5909757898417722, + "tokens_seen": 1295319040 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003068304914744233, + "loss": 1.3844, + "theoretical_loss": 3.590959447798228, + "tokens_seen": 1295384576 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003068204613841525, + "loss": 1.4624, + "theoretical_loss": 3.5909431068129223, + "tokens_seen": 1295450112 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030681043129388166, + "loss": 1.4163, + "theoretical_loss": 3.5909267668857323, + "tokens_seen": 1295515648 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030680040120361084, + "loss": 1.4503, + "theoretical_loss": 3.590910428016537, + "tokens_seen": 1295581184 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030679037111334, + "loss": 1.4414, + "theoretical_loss": 3.5908940902052136, + "tokens_seen": 1295646720 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030678034102306925, + "loss": 1.4978, + "theoretical_loss": 3.5908777534516405, + "tokens_seen": 1295712256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003067703109327984, + "loss": 1.3881, + "theoretical_loss": 3.5908614177556957, + "tokens_seen": 1295777792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003067602808425276, + "loss": 1.277, + "theoretical_loss": 3.590845083117257, + "tokens_seen": 1295843328 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030675025075225674, + "loss": 1.4483, + "theoretical_loss": 3.590828749536203, + "tokens_seen": 1295908864 + }, + { + "epoch": 0.39, + "learning_rate": 0.000306740220661986, + "loss": 1.3839, + "theoretical_loss": 3.590812417012411, + "tokens_seen": 1295974400 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030673019057171516, + "loss": 1.4635, + "theoretical_loss": 3.59079608554576, + "tokens_seen": 1296039936 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030672016048144434, + "loss": 1.3023, + "theoretical_loss": 3.5907797551361273, + "tokens_seen": 1296105472 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003067101303911735, + "loss": 1.3621, + "theoretical_loss": 3.5907634257833916, + "tokens_seen": 1296171008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003067001003009027, + "loss": 1.3786, + "theoretical_loss": 3.5907470974874305, + "tokens_seen": 1296236544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003066900702106319, + "loss": 1.3938, + "theoretical_loss": 3.590730770248123, + "tokens_seen": 1296302080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003066800401203611, + "loss": 1.4073, + "theoretical_loss": 3.5907144440653465, + "tokens_seen": 1296367616 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030667001003009025, + "loss": 1.4346, + "theoretical_loss": 3.5906981189389793, + "tokens_seen": 1296433152 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003066599799398195, + "loss": 1.429, + "theoretical_loss": 3.5906817948689005, + "tokens_seen": 1296498688 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030664994984954866, + "loss": 1.4939, + "theoretical_loss": 3.5906654718549875, + "tokens_seen": 1296564224 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030663991975927784, + "loss": 1.434, + "theoretical_loss": 3.590649149897119, + "tokens_seen": 1296629760 + }, + { + "epoch": 0.39, + "learning_rate": 0.000306629889669007, + "loss": 1.3767, + "theoretical_loss": 3.590632828995173, + "tokens_seen": 1296695296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003066198595787362, + "loss": 1.3865, + "theoretical_loss": 3.5906165091490285, + "tokens_seen": 1296760832 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003066098294884654, + "loss": 1.3375, + "theoretical_loss": 3.590600190358563, + "tokens_seen": 1296826368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003065997993981946, + "loss": 1.3971, + "theoretical_loss": 3.5905838726236556, + "tokens_seen": 1296891904 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030658976930792375, + "loss": 1.3835, + "theoretical_loss": 3.590567555944183, + "tokens_seen": 1296957440 + }, + { + "epoch": 0.39, + "learning_rate": 0.000306579739217653, + "loss": 1.3363, + "theoretical_loss": 3.5905512403200266, + "tokens_seen": 1297022976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003065697091273821, + "loss": 1.3805, + "theoretical_loss": 3.5905349257510624, + "tokens_seen": 1297088512 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030655967903711135, + "loss": 1.4096, + "theoretical_loss": 3.5905186122371697, + "tokens_seen": 1297154048 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030654964894684053, + "loss": 1.4061, + "theoretical_loss": 3.5905022997782265, + "tokens_seen": 1297219584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003065396188565697, + "loss": 1.3675, + "theoretical_loss": 3.5904859883741125, + "tokens_seen": 1297285120 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003065295887662989, + "loss": 1.3742, + "theoretical_loss": 3.5904696780247045, + "tokens_seen": 1297350656 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003065195586760281, + "loss": 1.3617, + "theoretical_loss": 3.5904533687298823, + "tokens_seen": 1297416192 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030650952858575725, + "loss": 1.3769, + "theoretical_loss": 3.5904370604895237, + "tokens_seen": 1297481728 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064994984954865, + "loss": 1.3493, + "theoretical_loss": 3.590420753303508, + "tokens_seen": 1297547264 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 474727, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.352555274963379, + "objective/train/theoretical_loss": 3.5904044471717134, + "objective/train/tokens_used": 1318072800, + "theoretical_loss": 3.5904044471717134, + "tokens_seen": 1297612800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064894684052156, + "loss": 1.3983, + "theoretical_loss": 3.5904044471717134, + "tokens_seen": 1297612800 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030647943831494485, + "loss": 1.3509, + "theoretical_loss": 3.5903881420940182, + "tokens_seen": 1297678336 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030646940822467403, + "loss": 1.3042, + "theoretical_loss": 3.5903718380703014, + "tokens_seen": 1297743872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064593781344032, + "loss": 1.3755, + "theoretical_loss": 3.5903555351004415, + "tokens_seen": 1297809408 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064493480441324, + "loss": 1.2939, + "theoretical_loss": 3.5903392331843174, + "tokens_seen": 1297874944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064393179538616, + "loss": 1.2842, + "theoretical_loss": 3.5903229323218078, + "tokens_seen": 1297940480 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030642928786359076, + "loss": 1.3168, + "theoretical_loss": 3.5903066325127906, + "tokens_seen": 1298006016 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030641925777332, + "loss": 1.3621, + "theoretical_loss": 3.590290333757146, + "tokens_seen": 1298071552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003064092276830491, + "loss": 1.412, + "theoretical_loss": 3.590274036054751, + "tokens_seen": 1298137088 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030639919759277835, + "loss": 1.4722, + "theoretical_loss": 3.5902577394054864, + "tokens_seen": 1298202624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063891675025075, + "loss": 1.3654, + "theoretical_loss": 3.590241443809229, + "tokens_seen": 1298268160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063791374122367, + "loss": 1.4049, + "theoretical_loss": 3.590225149265859, + "tokens_seen": 1298333696 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063691073219659, + "loss": 1.2821, + "theoretical_loss": 3.5902088557752547, + "tokens_seen": 1298399232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063590772316951, + "loss": 1.4857, + "theoretical_loss": 3.5901925633372946, + "tokens_seen": 1298464768 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030634904714142426, + "loss": 1.2882, + "theoretical_loss": 3.590176271951858, + "tokens_seen": 1298530304 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063390170511535, + "loss": 1.3827, + "theoretical_loss": 3.5901599816188243, + "tokens_seen": 1298595840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003063289869608826, + "loss": 1.2987, + "theoretical_loss": 3.5901436923380716, + "tokens_seen": 1298661376 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030631895687061186, + "loss": 1.2589, + "theoretical_loss": 3.5901274041094786, + "tokens_seen": 1298726912 + }, + { + "epoch": 0.39, + "learning_rate": 0.000306308926780341, + "loss": 1.5484, + "theoretical_loss": 3.590111116932925, + "tokens_seen": 1298792448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062988966900702, + "loss": 1.3646, + "theoretical_loss": 3.5900948308082894, + "tokens_seen": 1298857984 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030628886659979945, + "loss": 1.3239, + "theoretical_loss": 3.5900785457354516, + "tokens_seen": 1298923520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062788365095286, + "loss": 1.4018, + "theoretical_loss": 3.590062261714289, + "tokens_seen": 1298989056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062688064192578, + "loss": 1.3172, + "theoretical_loss": 3.5900459787446817, + "tokens_seen": 1299054592 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030625877632898694, + "loss": 1.389, + "theoretical_loss": 3.590029696826509, + "tokens_seen": 1299120128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062487462387162, + "loss": 1.3246, + "theoretical_loss": 3.5900134159596493, + "tokens_seen": 1299185664 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030623871614844536, + "loss": 1.3371, + "theoretical_loss": 3.5899971361439817, + "tokens_seen": 1299251200 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030622868605817454, + "loss": 1.4753, + "theoretical_loss": 3.589980857379386, + "tokens_seen": 1299316736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062186559679037, + "loss": 1.3395, + "theoretical_loss": 3.5899645796657405, + "tokens_seen": 1299382272 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003062086258776329, + "loss": 1.429, + "theoretical_loss": 3.5899483030029247, + "tokens_seen": 1299447808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061985957873621, + "loss": 1.4024, + "theoretical_loss": 3.5899320273908177, + "tokens_seen": 1299513344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061885656970913, + "loss": 1.4665, + "theoretical_loss": 3.5899157528292993, + "tokens_seen": 1299578880 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030617853560682045, + "loss": 1.4081, + "theoretical_loss": 3.5898994793182477, + "tokens_seen": 1299644416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061685055165497, + "loss": 1.3927, + "theoretical_loss": 3.589883206857543, + "tokens_seen": 1299709952 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030615847542627886, + "loss": 1.2844, + "theoretical_loss": 3.5898669354470636, + "tokens_seen": 1299775488 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030614844533600804, + "loss": 1.4277, + "theoretical_loss": 3.58985066508669, + "tokens_seen": 1299841024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061384152457372, + "loss": 1.4647, + "theoretical_loss": 3.5898343957763004, + "tokens_seen": 1299906560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061283851554664, + "loss": 1.472, + "theoretical_loss": 3.589818127515774, + "tokens_seen": 1299972096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061183550651956, + "loss": 1.3912, + "theoretical_loss": 3.5898018603049913, + "tokens_seen": 1300037632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003061083249749248, + "loss": 1.3506, + "theoretical_loss": 3.58978559414383, + "tokens_seen": 1300103168 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030609829488465395, + "loss": 1.3518, + "theoretical_loss": 3.589769329032171, + "tokens_seen": 1300168704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060882647943832, + "loss": 1.3942, + "theoretical_loss": 3.5897530649698934, + "tokens_seen": 1300234240 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060782347041123, + "loss": 1.3716, + "theoretical_loss": 3.5897368019568754, + "tokens_seen": 1300299776 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030606820461384155, + "loss": 1.4546, + "theoretical_loss": 3.589720539992998, + "tokens_seen": 1300365312 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030605817452357073, + "loss": 1.3456, + "theoretical_loss": 3.58970427907814, + "tokens_seen": 1300430848 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060481444332999, + "loss": 1.4543, + "theoretical_loss": 3.58968801921218, + "tokens_seen": 1300496384 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060381143430291, + "loss": 1.3588, + "theoretical_loss": 3.5896717603949995, + "tokens_seen": 1300561920 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060280842527583, + "loss": 1.3998, + "theoretical_loss": 3.589655502626476, + "tokens_seen": 1300627456 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030601805416248745, + "loss": 1.4861, + "theoretical_loss": 3.5896392459064903, + "tokens_seen": 1300692992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003060080240722167, + "loss": 1.3683, + "theoretical_loss": 3.5896229902349215, + "tokens_seen": 1300758528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059979939819458, + "loss": 1.3737, + "theoretical_loss": 3.5896067356116492, + "tokens_seen": 1300824064 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 475730, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.350705862045288, + "objective/train/theoretical_loss": 3.5895904820365527, + "objective/train/tokens_used": 1321349600, + "theoretical_loss": 3.5895904820365527, + "tokens_seen": 1300889600 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030598796389167505, + "loss": 1.4283, + "theoretical_loss": 3.5895904820365527, + "tokens_seen": 1300889600 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030597793380140423, + "loss": 1.3199, + "theoretical_loss": 3.589574229509512, + "tokens_seen": 1300955136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059679037111334, + "loss": 1.432, + "theoretical_loss": 3.589557978030407, + "tokens_seen": 1301020672 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059578736208626, + "loss": 1.3792, + "theoretical_loss": 3.5895417275991166, + "tokens_seen": 1301086208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059478435305918, + "loss": 1.4829, + "theoretical_loss": 3.5895254782155206, + "tokens_seen": 1301151744 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030593781344032096, + "loss": 1.3321, + "theoretical_loss": 3.589509229879499, + "tokens_seen": 1301217280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059277833500502, + "loss": 1.3899, + "theoretical_loss": 3.589492982590932, + "tokens_seen": 1301282816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003059177532597793, + "loss": 1.3861, + "theoretical_loss": 3.5894767363496984, + "tokens_seen": 1301348352 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030590772316950855, + "loss": 1.4635, + "theoretical_loss": 3.5894604911556782, + "tokens_seen": 1301413888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058976930792377, + "loss": 1.3917, + "theoretical_loss": 3.5894442470087515, + "tokens_seen": 1301479424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058876629889669, + "loss": 1.3487, + "theoretical_loss": 3.5894280039087976, + "tokens_seen": 1301544960 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058776328986961, + "loss": 1.3056, + "theoretical_loss": 3.589411761855697, + "tokens_seen": 1301610496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058676028084253, + "loss": 1.4497, + "theoretical_loss": 3.5893955208493287, + "tokens_seen": 1301676032 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030585757271815446, + "loss": 1.3963, + "theoretical_loss": 3.589379280889573, + "tokens_seen": 1301741568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058475426278837, + "loss": 1.2864, + "theoretical_loss": 3.589363041976309, + "tokens_seen": 1301807104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058375125376128, + "loss": 1.4961, + "theoretical_loss": 3.5893468041094185, + "tokens_seen": 1301872640 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030582748244734206, + "loss": 1.353, + "theoretical_loss": 3.58933056728878, + "tokens_seen": 1301938176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058174523570712, + "loss": 1.3595, + "theoretical_loss": 3.589314331514273, + "tokens_seen": 1302003712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003058074222668004, + "loss": 1.3446, + "theoretical_loss": 3.589298096785779, + "tokens_seen": 1302069248 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057973921765296, + "loss": 1.372, + "theoretical_loss": 3.589281863103176, + "tokens_seen": 1302134784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057873620862588, + "loss": 1.4399, + "theoretical_loss": 3.5892656304663455, + "tokens_seen": 1302200320 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030577733199598796, + "loss": 1.4903, + "theoretical_loss": 3.5892493988751673, + "tokens_seen": 1302265856 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030576730190571714, + "loss": 1.3421, + "theoretical_loss": 3.5892331683295207, + "tokens_seen": 1302331392 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057572718154463, + "loss": 1.3311, + "theoretical_loss": 3.5892169388292867, + "tokens_seen": 1302396928 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030574724172517556, + "loss": 1.3064, + "theoretical_loss": 3.589200710374344, + "tokens_seen": 1302462464 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057372116349047, + "loss": 1.4096, + "theoretical_loss": 3.5891844829645745, + "tokens_seen": 1302528000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057271815446339, + "loss": 1.5163, + "theoretical_loss": 3.589168256599857, + "tokens_seen": 1302593536 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030571715145436305, + "loss": 1.3776, + "theoretical_loss": 3.5891520312800718, + "tokens_seen": 1302659072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003057071213640923, + "loss": 1.3738, + "theoretical_loss": 3.5891358070051, + "tokens_seen": 1302724608 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030569709127382147, + "loss": 1.41, + "theoretical_loss": 3.5891195837748198, + "tokens_seen": 1302790144 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030568706118355065, + "loss": 1.4486, + "theoretical_loss": 3.589103361589113, + "tokens_seen": 1302855680 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030567703109327983, + "loss": 1.3587, + "theoretical_loss": 3.5890871404478597, + "tokens_seen": 1302921216 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030566700100300906, + "loss": 1.458, + "theoretical_loss": 3.58907092035094, + "tokens_seen": 1302986752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003056569709127382, + "loss": 1.4286, + "theoretical_loss": 3.5890547012982337, + "tokens_seen": 1303052288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003056469408224674, + "loss": 1.46, + "theoretical_loss": 3.589038483289621, + "tokens_seen": 1303117824 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030563691073219655, + "loss": 1.3423, + "theoretical_loss": 3.5890222663249824, + "tokens_seen": 1303183360 + }, + { + "epoch": 0.39, + "learning_rate": 0.0003056268806419258, + "loss": 1.4299, + "theoretical_loss": 3.589006050404199, + "tokens_seen": 1303248896 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030561685055165497, + "loss": 1.3782, + "theoretical_loss": 3.58898983552715, + "tokens_seen": 1303314432 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030560682046138415, + "loss": 1.4668, + "theoretical_loss": 3.5889736216937163, + "tokens_seen": 1303379968 + }, + { + "epoch": 0.39, + "learning_rate": 0.00030559679037111333, + "loss": 1.3946, + "theoretical_loss": 3.588957408903778, + "tokens_seen": 1303445504 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003055867602808425, + "loss": 1.393, + "theoretical_loss": 3.5889411971572156, + "tokens_seen": 1303511040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003055767301905717, + "loss": 1.4132, + "theoretical_loss": 3.5889249864539097, + "tokens_seen": 1303576576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030556670010030093, + "loss": 1.2884, + "theoretical_loss": 3.5889087767937404, + "tokens_seen": 1303642112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030555667001003006, + "loss": 1.4194, + "theoretical_loss": 3.588892568176588, + "tokens_seen": 1303707648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003055466399197593, + "loss": 1.4504, + "theoretical_loss": 3.5888763606023337, + "tokens_seen": 1303773184 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003055366098294885, + "loss": 1.3667, + "theoretical_loss": 3.588860154070857, + "tokens_seen": 1303838720 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030552657973921765, + "loss": 1.4751, + "theoretical_loss": 3.5888439485820394, + "tokens_seen": 1303904256 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003055165496489469, + "loss": 1.4915, + "theoretical_loss": 3.5888277441357603, + "tokens_seen": 1303969792 + }, + { + "epoch": 0.4, + "learning_rate": 0.000305506519558676, + "loss": 1.3596, + "theoretical_loss": 3.5888115407319017, + "tokens_seen": 1304035328 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030549648946840525, + "loss": 1.3442, + "theoretical_loss": 3.5887953383703426, + "tokens_seen": 1304100864 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 476908, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3124165534973145, + "objective/train/theoretical_loss": 3.588779137050965, + "objective/train/tokens_used": 1324626400, + "theoretical_loss": 3.588779137050965, + "tokens_seen": 1304166400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030548645937813443, + "loss": 1.2912, + "theoretical_loss": 3.588779137050965, + "tokens_seen": 1304166400 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003054764292878636, + "loss": 1.5225, + "theoretical_loss": 3.588762936773648, + "tokens_seen": 1304231936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003054663991975928, + "loss": 1.3739, + "theoretical_loss": 3.5887467375382736, + "tokens_seen": 1304297472 + }, + { + "epoch": 0.4, + "learning_rate": 0.000305456369107322, + "loss": 1.3239, + "theoretical_loss": 3.5887305393447217, + "tokens_seen": 1304363008 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030544633901705116, + "loss": 1.5515, + "theoretical_loss": 3.5887143421928727, + "tokens_seen": 1304428544 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003054363089267804, + "loss": 1.3962, + "theoretical_loss": 3.588698146082608, + "tokens_seen": 1304494080 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003054262788365095, + "loss": 1.3082, + "theoretical_loss": 3.5886819510138084, + "tokens_seen": 1304559616 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030541624874623875, + "loss": 1.4138, + "theoretical_loss": 3.5886657569863543, + "tokens_seen": 1304625152 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003054062186559679, + "loss": 1.3481, + "theoretical_loss": 3.588649564000126, + "tokens_seen": 1304690688 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053961885656971, + "loss": 1.4116, + "theoretical_loss": 3.5886333720550043, + "tokens_seen": 1304756224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053861584754263, + "loss": 1.3528, + "theoretical_loss": 3.5886171811508705, + "tokens_seen": 1304821760 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053761283851555, + "loss": 1.3031, + "theoretical_loss": 3.588600991287605, + "tokens_seen": 1304887296 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030536609829488466, + "loss": 1.5049, + "theoretical_loss": 3.588584802465089, + "tokens_seen": 1304952832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053560682046139, + "loss": 1.4682, + "theoretical_loss": 3.5885686146832025, + "tokens_seen": 1305018368 + }, + { + "epoch": 0.4, + "learning_rate": 0.000305346038114343, + "loss": 1.4601, + "theoretical_loss": 3.5885524279418277, + "tokens_seen": 1305083904 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030533600802407226, + "loss": 1.4646, + "theoretical_loss": 3.5885362422408447, + "tokens_seen": 1305149440 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053259779338014, + "loss": 1.3935, + "theoretical_loss": 3.5885200575801344, + "tokens_seen": 1305214976 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053159478435306, + "loss": 1.4099, + "theoretical_loss": 3.588503873959578, + "tokens_seen": 1305280512 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003053059177532598, + "loss": 1.4008, + "theoretical_loss": 3.588487691379055, + "tokens_seen": 1305346048 + }, + { + "epoch": 0.4, + "learning_rate": 0.000305295887662989, + "loss": 1.4282, + "theoretical_loss": 3.588471509838449, + "tokens_seen": 1305411584 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030528585757271816, + "loss": 1.3136, + "theoretical_loss": 3.5884553293376387, + "tokens_seen": 1305477120 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030527582748244734, + "loss": 1.3899, + "theoretical_loss": 3.588439149876506, + "tokens_seen": 1305542656 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003052657973921765, + "loss": 1.3317, + "theoretical_loss": 3.5884229714549316, + "tokens_seen": 1305608192 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030525576730190576, + "loss": 1.3077, + "theoretical_loss": 3.588406794072797, + "tokens_seen": 1305673728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003052457372116349, + "loss": 1.4584, + "theoretical_loss": 3.588390617729983, + "tokens_seen": 1305739264 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003052357071213641, + "loss": 1.4098, + "theoretical_loss": 3.5883744424263706, + "tokens_seen": 1305804800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030522567703109325, + "loss": 1.2969, + "theoretical_loss": 3.5883582681618407, + "tokens_seen": 1305870336 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003052156469408225, + "loss": 1.3198, + "theoretical_loss": 3.588342094936275, + "tokens_seen": 1305935872 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030520561685055167, + "loss": 1.2626, + "theoretical_loss": 3.5883259227495543, + "tokens_seen": 1306001408 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030519558676028085, + "loss": 1.3904, + "theoretical_loss": 3.5883097516015594, + "tokens_seen": 1306066944 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030518555667001003, + "loss": 1.4181, + "theoretical_loss": 3.588293581492172, + "tokens_seen": 1306132480 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030517552657973926, + "loss": 1.3004, + "theoretical_loss": 3.5882774124212733, + "tokens_seen": 1306198016 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003051654964894684, + "loss": 1.3469, + "theoretical_loss": 3.588261244388744, + "tokens_seen": 1306263552 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003051554663991976, + "loss": 1.4773, + "theoretical_loss": 3.5882450773944656, + "tokens_seen": 1306329088 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030514543630892675, + "loss": 1.4315, + "theoretical_loss": 3.5882289114383195, + "tokens_seen": 1306394624 + }, + { + "epoch": 0.4, + "learning_rate": 0.000305135406218656, + "loss": 1.2753, + "theoretical_loss": 3.5882127465201865, + "tokens_seen": 1306460160 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030512537612838517, + "loss": 1.3664, + "theoretical_loss": 3.5881965826399487, + "tokens_seen": 1306525696 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030511534603811435, + "loss": 1.3488, + "theoretical_loss": 3.588180419797487, + "tokens_seen": 1306591232 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030510531594784353, + "loss": 1.3916, + "theoretical_loss": 3.5881642579926822, + "tokens_seen": 1306656768 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050952858575727, + "loss": 1.4958, + "theoretical_loss": 3.588148097225416, + "tokens_seen": 1306722304 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050852557673019, + "loss": 1.3126, + "theoretical_loss": 3.58813193749557, + "tokens_seen": 1306787840 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030507522567703113, + "loss": 1.4076, + "theoretical_loss": 3.5881157788030253, + "tokens_seen": 1306853376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030506519558676026, + "loss": 1.471, + "theoretical_loss": 3.5880996211476637, + "tokens_seen": 1306918912 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050551654964895, + "loss": 1.3286, + "theoretical_loss": 3.588083464529366, + "tokens_seen": 1306984448 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050451354062186, + "loss": 1.3884, + "theoretical_loss": 3.5880673089480144, + "tokens_seen": 1307049984 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030503510531594785, + "loss": 1.3699, + "theoretical_loss": 3.5880511544034897, + "tokens_seen": 1307115520 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030502507522567703, + "loss": 1.4534, + "theoretical_loss": 3.5880350008956734, + "tokens_seen": 1307181056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050150451354062, + "loss": 1.5303, + "theoretical_loss": 3.588018848424448, + "tokens_seen": 1307246592 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003050050150451354, + "loss": 1.3762, + "theoretical_loss": 3.5880026969896934, + "tokens_seen": 1307312128 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030499498495486463, + "loss": 1.3032, + "theoretical_loss": 3.587986546591292, + "tokens_seen": 1307377664 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 478148, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4599223136901855, + "objective/train/theoretical_loss": 3.587970397229126, + "objective/train/tokens_used": 1327903200, + "theoretical_loss": 3.587970397229126, + "tokens_seen": 1307443200 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030498495486459376, + "loss": 1.3772, + "theoretical_loss": 3.587970397229126, + "tokens_seen": 1307443200 + }, + { + "epoch": 0.4, + "learning_rate": 0.000304974924774323, + "loss": 1.4604, + "theoretical_loss": 3.587954248903076, + "tokens_seen": 1307508736 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003049648946840521, + "loss": 1.2798, + "theoretical_loss": 3.5879381016130236, + "tokens_seen": 1307574272 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030495486459378136, + "loss": 1.4067, + "theoretical_loss": 3.5879219553588513, + "tokens_seen": 1307639808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030494483450351054, + "loss": 1.4226, + "theoretical_loss": 3.5879058101404397, + "tokens_seen": 1307705344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003049348044132397, + "loss": 1.4496, + "theoretical_loss": 3.587889665957671, + "tokens_seen": 1307770880 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003049247743229689, + "loss": 1.3529, + "theoretical_loss": 3.587873522810427, + "tokens_seen": 1307836416 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003049147442326981, + "loss": 1.3538, + "theoretical_loss": 3.5878573806985887, + "tokens_seen": 1307901952 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030490471414242726, + "loss": 1.4572, + "theoretical_loss": 3.5878412396220387, + "tokens_seen": 1307967488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048946840521565, + "loss": 1.4396, + "theoretical_loss": 3.587825099580658, + "tokens_seen": 1308033024 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048846539618856, + "loss": 1.4335, + "theoretical_loss": 3.5878089605743293, + "tokens_seen": 1308098560 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030487462387161486, + "loss": 1.4086, + "theoretical_loss": 3.5877928226029336, + "tokens_seen": 1308164096 + }, + { + "epoch": 0.4, + "learning_rate": 0.000304864593781344, + "loss": 1.392, + "theoretical_loss": 3.5877766856663524, + "tokens_seen": 1308229632 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048545636910732, + "loss": 1.303, + "theoretical_loss": 3.587760549764468, + "tokens_seen": 1308295168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048445336008024, + "loss": 1.3335, + "theoretical_loss": 3.5877444148971627, + "tokens_seen": 1308360704 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048345035105316, + "loss": 1.3644, + "theoretical_loss": 3.5877282810643174, + "tokens_seen": 1308426240 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030482447342026076, + "loss": 1.3623, + "theoretical_loss": 3.5877121482658145, + "tokens_seen": 1308491776 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030481444332999, + "loss": 1.2973, + "theoretical_loss": 3.587696016501536, + "tokens_seen": 1308557312 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003048044132397192, + "loss": 1.3681, + "theoretical_loss": 3.5876798857713634, + "tokens_seen": 1308622848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030479438314944836, + "loss": 1.4997, + "theoretical_loss": 3.5876637560751785, + "tokens_seen": 1308688384 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030478435305917754, + "loss": 1.2979, + "theoretical_loss": 3.5876476274128644, + "tokens_seen": 1308753920 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003047743229689067, + "loss": 1.3701, + "theoretical_loss": 3.587631499784301, + "tokens_seen": 1308819456 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030476429287863596, + "loss": 1.3848, + "theoretical_loss": 3.5876153731893727, + "tokens_seen": 1308884992 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003047542627883651, + "loss": 1.374, + "theoretical_loss": 3.58759924762796, + "tokens_seen": 1308950528 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003047442326980943, + "loss": 1.3395, + "theoretical_loss": 3.587583123099945, + "tokens_seen": 1309016064 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030473420260782345, + "loss": 1.4197, + "theoretical_loss": 3.58756699960521, + "tokens_seen": 1309081600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003047241725175527, + "loss": 1.4581, + "theoretical_loss": 3.587550877143637, + "tokens_seen": 1309147136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030471414242728187, + "loss": 1.4927, + "theoretical_loss": 3.5875347557151085, + "tokens_seen": 1309212672 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030470411233701105, + "loss": 1.4818, + "theoretical_loss": 3.5875186353195065, + "tokens_seen": 1309278208 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030469408224674023, + "loss": 1.434, + "theoretical_loss": 3.5875025159567118, + "tokens_seen": 1309343744 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030468405215646946, + "loss": 1.4428, + "theoretical_loss": 3.5874863976266083, + "tokens_seen": 1309409280 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003046740220661986, + "loss": 1.4784, + "theoretical_loss": 3.5874702803290774, + "tokens_seen": 1309474816 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003046639919759278, + "loss": 1.3917, + "theoretical_loss": 3.5874541640640016, + "tokens_seen": 1309540352 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030465396188565695, + "loss": 1.3554, + "theoretical_loss": 3.5874380488312623, + "tokens_seen": 1309605888 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003046439317953862, + "loss": 1.3796, + "theoretical_loss": 3.5874219346307425, + "tokens_seen": 1309671424 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030463390170511537, + "loss": 1.3752, + "theoretical_loss": 3.5874058214623243, + "tokens_seen": 1309736960 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030462387161484455, + "loss": 1.4086, + "theoretical_loss": 3.5873897093258895, + "tokens_seen": 1309802496 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030461384152457373, + "loss": 1.4265, + "theoretical_loss": 3.5873735982213213, + "tokens_seen": 1309868032 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003046038114343029, + "loss": 1.4563, + "theoretical_loss": 3.587357488148501, + "tokens_seen": 1309933568 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003045937813440321, + "loss": 1.438, + "theoretical_loss": 3.587341379107311, + "tokens_seen": 1309999104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030458375125376133, + "loss": 1.269, + "theoretical_loss": 3.587325271097635, + "tokens_seen": 1310064640 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030457372116349046, + "loss": 1.4539, + "theoretical_loss": 3.587309164119353, + "tokens_seen": 1310130176 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003045636910732197, + "loss": 1.4901, + "theoretical_loss": 3.5872930581723494, + "tokens_seen": 1310195712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003045536609829488, + "loss": 1.3181, + "theoretical_loss": 3.587276953256506, + "tokens_seen": 1310261248 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030454363089267805, + "loss": 1.2742, + "theoretical_loss": 3.587260849371705, + "tokens_seen": 1310326784 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030453360080240723, + "loss": 1.3893, + "theoretical_loss": 3.587244746517829, + "tokens_seen": 1310392320 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003045235707121364, + "loss": 1.4459, + "theoretical_loss": 3.5872286446947603, + "tokens_seen": 1310457856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003045135406218656, + "loss": 1.3133, + "theoretical_loss": 3.587212543902381, + "tokens_seen": 1310523392 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030450351053159483, + "loss": 1.3714, + "theoretical_loss": 3.5871964441405746, + "tokens_seen": 1310588928 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030449348044132396, + "loss": 1.3167, + "theoretical_loss": 3.587180345409222, + "tokens_seen": 1310654464 + }, + { + "debugging/distinct-1-grams": 0.7732995475821178, + "debugging/entropy-1-grams": 5.431960387820077, + "debugging/length": 501.4, + "debugging/num_segments": 10, + "debugging/score": 0.007191963134632398, + "debugging/score_std": 0.006508080924945895, + "epoch": 0.4, + "objective/train/docs_used": 479375, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4441468715667725, + "objective/train/theoretical_loss": 3.587164247708208, + "objective/train/tokens_used": 1331180000, + "theoretical_loss": 3.587164247708208, + "tokens_seen": 1310720000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044834503510532, + "loss": 1.4191, + "theoretical_loss": 3.587164247708208, + "tokens_seen": 1310720000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044734202607823, + "loss": 1.4383, + "theoretical_loss": 3.5871481510374137, + "tokens_seen": 1310785536 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030446339017051156, + "loss": 1.2686, + "theoretical_loss": 3.5871320553967214, + "tokens_seen": 1310851072 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030445336008024074, + "loss": 1.3596, + "theoretical_loss": 3.587115960786014, + "tokens_seen": 1310916608 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044433299899699, + "loss": 1.3787, + "theoretical_loss": 3.5870998672051746, + "tokens_seen": 1310982144 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044332998996991, + "loss": 1.3908, + "theoretical_loss": 3.5870837746540856, + "tokens_seen": 1311047680 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044232698094283, + "loss": 1.3944, + "theoretical_loss": 3.5870676831326294, + "tokens_seen": 1311113216 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030441323971915746, + "loss": 1.3777, + "theoretical_loss": 3.5870515926406887, + "tokens_seen": 1311178752 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003044032096288867, + "loss": 1.4491, + "theoretical_loss": 3.5870355031781465, + "tokens_seen": 1311244288 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043931795386158, + "loss": 1.4757, + "theoretical_loss": 3.587019414744885, + "tokens_seen": 1311309824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030438314944834506, + "loss": 1.4467, + "theoretical_loss": 3.587003327340787, + "tokens_seen": 1311375360 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043731193580742, + "loss": 1.3219, + "theoretical_loss": 3.586987240965736, + "tokens_seen": 1311440896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043630892678034, + "loss": 1.4154, + "theoretical_loss": 3.5869711556196133, + "tokens_seen": 1311506432 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043530591775326, + "loss": 1.2743, + "theoretical_loss": 3.5869550713023033, + "tokens_seen": 1311571968 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043430290872618, + "loss": 1.3945, + "theoretical_loss": 3.5869389880136877, + "tokens_seen": 1311637504 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030433299899699097, + "loss": 1.39, + "theoretical_loss": 3.5869229057536494, + "tokens_seen": 1311703040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003043229689067202, + "loss": 1.2351, + "theoretical_loss": 3.586906824522072, + "tokens_seen": 1311768576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030431293881644933, + "loss": 1.2513, + "theoretical_loss": 3.586890744318838, + "tokens_seen": 1311834112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030430290872617856, + "loss": 1.2541, + "theoretical_loss": 3.5868746651438297, + "tokens_seen": 1311899648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003042928786359077, + "loss": 1.3548, + "theoretical_loss": 3.5868585869969305, + "tokens_seen": 1311965184 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003042828485456369, + "loss": 1.3165, + "theoretical_loss": 3.5868425098780228, + "tokens_seen": 1312030720 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003042728184553661, + "loss": 1.4326, + "theoretical_loss": 3.5868264337869906, + "tokens_seen": 1312096256 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003042627883650953, + "loss": 1.3933, + "theoretical_loss": 3.5868103587237155, + "tokens_seen": 1312161792 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030425275827482447, + "loss": 1.4731, + "theoretical_loss": 3.586794284688082, + "tokens_seen": 1312227328 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030424272818455365, + "loss": 1.42, + "theoretical_loss": 3.586778211679972, + "tokens_seen": 1312292864 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030423269809428283, + "loss": 1.3948, + "theoretical_loss": 3.5867621396992684, + "tokens_seen": 1312358400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030422266800401207, + "loss": 1.4533, + "theoretical_loss": 3.5867460687458546, + "tokens_seen": 1312423936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003042126379137412, + "loss": 1.4085, + "theoretical_loss": 3.5867299988196137, + "tokens_seen": 1312489472 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030420260782347043, + "loss": 1.2579, + "theoretical_loss": 3.5867139299204287, + "tokens_seen": 1312555008 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003041925777331996, + "loss": 1.4262, + "theoretical_loss": 3.586697862048183, + "tokens_seen": 1312620544 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003041825476429288, + "loss": 1.3834, + "theoretical_loss": 3.586681795202759, + "tokens_seen": 1312686080 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030417251755265797, + "loss": 1.3947, + "theoretical_loss": 3.5866657293840403, + "tokens_seen": 1312751616 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030416248746238715, + "loss": 1.3426, + "theoretical_loss": 3.58664966459191, + "tokens_seen": 1312817152 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030415245737211633, + "loss": 1.3078, + "theoretical_loss": 3.586633600826251, + "tokens_seen": 1312882688 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030414242728184557, + "loss": 1.5592, + "theoretical_loss": 3.586617538086947, + "tokens_seen": 1312948224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003041323971915747, + "loss": 1.3773, + "theoretical_loss": 3.5866014763738807, + "tokens_seen": 1313013760 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030412236710130393, + "loss": 1.3, + "theoretical_loss": 3.5865854156869355, + "tokens_seen": 1313079296 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030411233701103306, + "loss": 1.2843, + "theoretical_loss": 3.5865693560259944, + "tokens_seen": 1313144832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003041023069207623, + "loss": 1.3442, + "theoretical_loss": 3.586553297390941, + "tokens_seen": 1313210368 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003040922768304915, + "loss": 1.5339, + "theoretical_loss": 3.5865372397816584, + "tokens_seen": 1313275904 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030408224674022066, + "loss": 1.3508, + "theoretical_loss": 3.5865211831980304, + "tokens_seen": 1313341440 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030407221664994984, + "loss": 1.4295, + "theoretical_loss": 3.5865051276399393, + "tokens_seen": 1313406976 + }, + { + "epoch": 0.4, + "learning_rate": 0.000304062186559679, + "loss": 1.4599, + "theoretical_loss": 3.5864890731072694, + "tokens_seen": 1313472512 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030405215646940825, + "loss": 1.3796, + "theoretical_loss": 3.5864730195999037, + "tokens_seen": 1313538048 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030404212637913743, + "loss": 1.4335, + "theoretical_loss": 3.586456967117725, + "tokens_seen": 1313603584 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003040320962888666, + "loss": 1.5189, + "theoretical_loss": 3.5864409156606176, + "tokens_seen": 1313669120 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003040220661985958, + "loss": 1.4177, + "theoretical_loss": 3.5864248652284645, + "tokens_seen": 1313734656 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030401203610832503, + "loss": 1.4471, + "theoretical_loss": 3.586408815821149, + "tokens_seen": 1313800192 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030400200601805416, + "loss": 1.3887, + "theoretical_loss": 3.5863927674385545, + "tokens_seen": 1313865728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039919759277834, + "loss": 1.482, + "theoretical_loss": 3.586376720080565, + "tokens_seen": 1313931264 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 480644, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4041553735733032, + "objective/train/theoretical_loss": 3.5863606737470635, + "objective/train/tokens_used": 1334456800, + "theoretical_loss": 3.5863606737470635, + "tokens_seen": 1313996800 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039819458375125, + "loss": 1.4058, + "theoretical_loss": 3.5863606737470635, + "tokens_seen": 1313996800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030397191574724176, + "loss": 1.3255, + "theoretical_loss": 3.5863446284379337, + "tokens_seen": 1314062336 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030396188565697094, + "loss": 1.3521, + "theoretical_loss": 3.586328584153059, + "tokens_seen": 1314127872 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039518555667001, + "loss": 1.3689, + "theoretical_loss": 3.586312540892323, + "tokens_seen": 1314193408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039418254764293, + "loss": 1.3273, + "theoretical_loss": 3.586296498655609, + "tokens_seen": 1314258944 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039317953861585, + "loss": 1.3923, + "theoretical_loss": 3.5862804574428013, + "tokens_seen": 1314324480 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030392176529588766, + "loss": 1.4899, + "theoretical_loss": 3.5862644172537825, + "tokens_seen": 1314390016 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003039117352056169, + "loss": 1.4307, + "theoretical_loss": 3.5862483780884373, + "tokens_seen": 1314455552 + }, + { + "epoch": 0.4, + "learning_rate": 0.000303901705115346, + "loss": 1.4152, + "theoretical_loss": 3.5862323399466485, + "tokens_seen": 1314521088 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030389167502507526, + "loss": 1.2525, + "theoretical_loss": 3.5862163028283005, + "tokens_seen": 1314586624 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003038816449348044, + "loss": 1.3796, + "theoretical_loss": 3.586200266733276, + "tokens_seen": 1314652160 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003038716148445336, + "loss": 1.47, + "theoretical_loss": 3.586184231661459, + "tokens_seen": 1314717696 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003038615847542628, + "loss": 1.3674, + "theoretical_loss": 3.586168197612734, + "tokens_seen": 1314783232 + }, + { + "epoch": 0.4, + "learning_rate": 0.000303851554663992, + "loss": 1.3189, + "theoretical_loss": 3.5861521645869843, + "tokens_seen": 1314848768 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030384152457372117, + "loss": 1.3616, + "theoretical_loss": 3.586136132584093, + "tokens_seen": 1314914304 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003038314944834504, + "loss": 1.4, + "theoretical_loss": 3.5861201016039446, + "tokens_seen": 1314979840 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030382146439317953, + "loss": 1.3592, + "theoretical_loss": 3.5861040716464228, + "tokens_seen": 1315045376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030381143430290876, + "loss": 1.4412, + "theoretical_loss": 3.5860880427114106, + "tokens_seen": 1315110912 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003038014042126379, + "loss": 1.4335, + "theoretical_loss": 3.586072014798793, + "tokens_seen": 1315176448 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003037913741223671, + "loss": 1.2986, + "theoretical_loss": 3.5860559879084537, + "tokens_seen": 1315241984 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003037813440320963, + "loss": 1.3534, + "theoretical_loss": 3.5860399620402754, + "tokens_seen": 1315307520 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003037713139418255, + "loss": 1.2479, + "theoretical_loss": 3.5860239371941436, + "tokens_seen": 1315373056 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030376128385155467, + "loss": 1.4894, + "theoretical_loss": 3.586007913369941, + "tokens_seen": 1315438592 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030375125376128385, + "loss": 1.4216, + "theoretical_loss": 3.585991890567552, + "tokens_seen": 1315504128 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030374122367101303, + "loss": 1.3966, + "theoretical_loss": 3.5859758687868606, + "tokens_seen": 1315569664 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030373119358074227, + "loss": 1.4871, + "theoretical_loss": 3.58595984802775, + "tokens_seen": 1315635200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003037211634904714, + "loss": 1.4634, + "theoretical_loss": 3.5859438282901057, + "tokens_seen": 1315700736 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030371113340020063, + "loss": 1.3224, + "theoretical_loss": 3.5859278095738105, + "tokens_seen": 1315766272 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003037011033099298, + "loss": 1.465, + "theoretical_loss": 3.585911791878748, + "tokens_seen": 1315831808 + }, + { + "epoch": 0.4, + "learning_rate": 0.000303691073219659, + "loss": 1.4127, + "theoretical_loss": 3.5858957752048037, + "tokens_seen": 1315897344 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030368104312938817, + "loss": 1.2726, + "theoretical_loss": 3.585879759551861, + "tokens_seen": 1315962880 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030367101303911735, + "loss": 1.5433, + "theoretical_loss": 3.5858637449198034, + "tokens_seen": 1316028416 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030366098294884653, + "loss": 1.463, + "theoretical_loss": 3.5858477313085158, + "tokens_seen": 1316093952 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030365095285857577, + "loss": 1.4266, + "theoretical_loss": 3.585831718717882, + "tokens_seen": 1316159488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003036409227683049, + "loss": 1.4186, + "theoretical_loss": 3.585815707147786, + "tokens_seen": 1316225024 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030363089267803413, + "loss": 1.3144, + "theoretical_loss": 3.5857996965981123, + "tokens_seen": 1316290560 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030362086258776326, + "loss": 1.3453, + "theoretical_loss": 3.5857836870687443, + "tokens_seen": 1316356096 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003036108324974925, + "loss": 1.4946, + "theoretical_loss": 3.585767678559567, + "tokens_seen": 1316421632 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003036008024072217, + "loss": 1.3083, + "theoretical_loss": 3.585751671070464, + "tokens_seen": 1316487168 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030359077231695086, + "loss": 1.4828, + "theoretical_loss": 3.5857356646013203, + "tokens_seen": 1316552704 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030358074222668004, + "loss": 1.5263, + "theoretical_loss": 3.5857196591520197, + "tokens_seen": 1316618240 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003035707121364092, + "loss": 1.4828, + "theoretical_loss": 3.585703654722446, + "tokens_seen": 1316683776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003035606820461384, + "loss": 1.3731, + "theoretical_loss": 3.5856876513124845, + "tokens_seen": 1316749312 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030355065195586763, + "loss": 1.4647, + "theoretical_loss": 3.585671648922019, + "tokens_seen": 1316814848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030354062186559676, + "loss": 1.3319, + "theoretical_loss": 3.585655647550933, + "tokens_seen": 1316880384 + }, + { + "epoch": 0.4, + "learning_rate": 0.000303530591775326, + "loss": 1.499, + "theoretical_loss": 3.5856396471991117, + "tokens_seen": 1316945920 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003035205616850552, + "loss": 1.3704, + "theoretical_loss": 3.5856236478664396, + "tokens_seen": 1317011456 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030351053159478436, + "loss": 1.3508, + "theoretical_loss": 3.5856076495528013, + "tokens_seen": 1317076992 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030350050150451354, + "loss": 1.4525, + "theoretical_loss": 3.58559165225808, + "tokens_seen": 1317142528 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034904714142427, + "loss": 1.5277, + "theoretical_loss": 3.5855756559821614, + "tokens_seen": 1317208064 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 481887, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4952340126037598, + "objective/train/theoretical_loss": 3.585559660724929, + "objective/train/tokens_used": 1337733600, + "theoretical_loss": 3.585559660724929, + "tokens_seen": 1317273600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034804413239719, + "loss": 1.4446, + "theoretical_loss": 3.585559660724929, + "tokens_seen": 1317273600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030347041123370114, + "loss": 1.5123, + "theoretical_loss": 3.5855436664862674, + "tokens_seen": 1317339136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030346038114343026, + "loss": 1.6431, + "theoretical_loss": 3.5855276732660615, + "tokens_seen": 1317404672 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034503510531595, + "loss": 1.5149, + "theoretical_loss": 3.5855116810641956, + "tokens_seen": 1317470208 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034403209628886, + "loss": 1.3399, + "theoretical_loss": 3.5854956898805543, + "tokens_seen": 1317535744 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030343029087261786, + "loss": 1.3893, + "theoretical_loss": 3.5854796997150222, + "tokens_seen": 1317601280 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030342026078234704, + "loss": 1.4779, + "theoretical_loss": 3.5854637105674834, + "tokens_seen": 1317666816 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034102306920762, + "loss": 1.2435, + "theoretical_loss": 3.5854477224378227, + "tokens_seen": 1317732352 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003034002006018054, + "loss": 1.5063, + "theoretical_loss": 3.5854317353259244, + "tokens_seen": 1317797888 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003033901705115346, + "loss": 1.3444, + "theoretical_loss": 3.585415749231674, + "tokens_seen": 1317863424 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030338014042126377, + "loss": 1.4685, + "theoretical_loss": 3.585399764154955, + "tokens_seen": 1317928960 + }, + { + "epoch": 0.4, + "learning_rate": 0.000303370110330993, + "loss": 1.4464, + "theoretical_loss": 3.585383780095653, + "tokens_seen": 1317994496 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030336008024072213, + "loss": 1.4199, + "theoretical_loss": 3.585367797053652, + "tokens_seen": 1318060032 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030335005015045137, + "loss": 1.343, + "theoretical_loss": 3.5853518150288366, + "tokens_seen": 1318125568 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030334002006018055, + "loss": 1.3931, + "theoretical_loss": 3.5853358340210923, + "tokens_seen": 1318191104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030332998996990973, + "loss": 1.471, + "theoretical_loss": 3.5853198540303035, + "tokens_seen": 1318256640 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003033199598796389, + "loss": 1.392, + "theoretical_loss": 3.585303875056354, + "tokens_seen": 1318322176 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003033099297893681, + "loss": 1.382, + "theoretical_loss": 3.5852878970991298, + "tokens_seen": 1318387712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003032998996990973, + "loss": 1.4131, + "theoretical_loss": 3.5852719201585153, + "tokens_seen": 1318453248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003032898696088265, + "loss": 1.3679, + "theoretical_loss": 3.5852559442343948, + "tokens_seen": 1318518784 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003032798395185557, + "loss": 1.4746, + "theoretical_loss": 3.5852399693266532, + "tokens_seen": 1318584320 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030326980942828487, + "loss": 1.2565, + "theoretical_loss": 3.585223995435176, + "tokens_seen": 1318649856 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030325977933801405, + "loss": 1.5029, + "theoretical_loss": 3.5852080225598475, + "tokens_seen": 1318715392 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030324974924774323, + "loss": 1.3832, + "theoretical_loss": 3.585192050700553, + "tokens_seen": 1318780928 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030323971915747247, + "loss": 1.3518, + "theoretical_loss": 3.5851760798571766, + "tokens_seen": 1318846464 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003032296890672016, + "loss": 1.4096, + "theoretical_loss": 3.585160110029604, + "tokens_seen": 1318912000 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030321965897693083, + "loss": 1.3467, + "theoretical_loss": 3.58514414121772, + "tokens_seen": 1318977536 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030320962888666, + "loss": 1.3083, + "theoretical_loss": 3.585128173421409, + "tokens_seen": 1319043072 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003031995987963892, + "loss": 1.5003, + "theoretical_loss": 3.585112206640556, + "tokens_seen": 1319108608 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030318956870611837, + "loss": 1.4056, + "theoretical_loss": 3.585096240875047, + "tokens_seen": 1319174144 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030317953861584755, + "loss": 1.3797, + "theoretical_loss": 3.5850802761247658, + "tokens_seen": 1319239680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030316950852557673, + "loss": 1.3238, + "theoretical_loss": 3.585064312389598, + "tokens_seen": 1319305216 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030315947843530597, + "loss": 1.3395, + "theoretical_loss": 3.585048349669429, + "tokens_seen": 1319370752 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003031494483450351, + "loss": 1.4135, + "theoretical_loss": 3.5850323879641426, + "tokens_seen": 1319436288 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030313941825476433, + "loss": 1.3029, + "theoretical_loss": 3.585016427273625, + "tokens_seen": 1319501824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030312938816449346, + "loss": 1.4678, + "theoretical_loss": 3.5850004675977614, + "tokens_seen": 1319567360 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003031193580742227, + "loss": 1.4274, + "theoretical_loss": 3.5849845089364356, + "tokens_seen": 1319632896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003031093279839519, + "loss": 1.3841, + "theoretical_loss": 3.5849685512895344, + "tokens_seen": 1319698432 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030309929789368106, + "loss": 1.3805, + "theoretical_loss": 3.584952594656942, + "tokens_seen": 1319763968 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030308926780341024, + "loss": 1.4408, + "theoretical_loss": 3.584936639038543, + "tokens_seen": 1319829504 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003030792377131394, + "loss": 1.3963, + "theoretical_loss": 3.584920684434224, + "tokens_seen": 1319895040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003030692076228686, + "loss": 1.4658, + "theoretical_loss": 3.584904730843869, + "tokens_seen": 1319960576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030305917753259783, + "loss": 1.3268, + "theoretical_loss": 3.5848887782673637, + "tokens_seen": 1320026112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030304914744232696, + "loss": 1.2807, + "theoretical_loss": 3.5848728267045935, + "tokens_seen": 1320091648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003030391173520562, + "loss": 1.4384, + "theoretical_loss": 3.5848568761554436, + "tokens_seen": 1320157184 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003030290872617854, + "loss": 1.2982, + "theoretical_loss": 3.5848409266197994, + "tokens_seen": 1320222720 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030301905717151456, + "loss": 1.365, + "theoretical_loss": 3.5848249780975454, + "tokens_seen": 1320288256 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030300902708124374, + "loss": 1.3141, + "theoretical_loss": 3.584809030588568, + "tokens_seen": 1320353792 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029989969909729, + "loss": 1.3041, + "theoretical_loss": 3.5847930840927518, + "tokens_seen": 1320419328 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029889669007021, + "loss": 1.4792, + "theoretical_loss": 3.584777138609982, + "tokens_seen": 1320484864 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 483102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3544766902923584, + "objective/train/theoretical_loss": 3.584761194140145, + "objective/train/tokens_used": 1341010400, + "theoretical_loss": 3.584761194140145, + "tokens_seen": 1320550400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030297893681043134, + "loss": 1.4125, + "theoretical_loss": 3.584761194140145, + "tokens_seen": 1320550400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030296890672016046, + "loss": 1.4311, + "theoretical_loss": 3.584745250683125, + "tokens_seen": 1320615936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029588766298897, + "loss": 1.2878, + "theoretical_loss": 3.584729308238808, + "tokens_seen": 1320681472 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029488465396188, + "loss": 1.4046, + "theoretical_loss": 3.5847133668070796, + "tokens_seen": 1320747008 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030293881644934806, + "loss": 1.4737, + "theoretical_loss": 3.584697426387825, + "tokens_seen": 1320812544 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030292878635907724, + "loss": 1.3683, + "theoretical_loss": 3.58468148698093, + "tokens_seen": 1320878080 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029187562688064, + "loss": 1.476, + "theoretical_loss": 3.5846655485862793, + "tokens_seen": 1320943616 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003029087261785356, + "loss": 1.4314, + "theoretical_loss": 3.5846496112037585, + "tokens_seen": 1321009152 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003028986960882648, + "loss": 1.2849, + "theoretical_loss": 3.5846336748332543, + "tokens_seen": 1321074688 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030288866599799397, + "loss": 1.419, + "theoretical_loss": 3.584617739474651, + "tokens_seen": 1321140224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003028786359077232, + "loss": 1.3994, + "theoretical_loss": 3.5846018051278348, + "tokens_seen": 1321205760 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030286860581745233, + "loss": 1.475, + "theoretical_loss": 3.58458587179269, + "tokens_seen": 1321271296 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030285857572718157, + "loss": 1.46, + "theoretical_loss": 3.5845699394691044, + "tokens_seen": 1321336832 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030284854563691075, + "loss": 1.5907, + "theoretical_loss": 3.584554008156962, + "tokens_seen": 1321402368 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030283851554663993, + "loss": 1.4471, + "theoretical_loss": 3.5845380778561493, + "tokens_seen": 1321467904 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003028284854563691, + "loss": 1.45, + "theoretical_loss": 3.584522148566551, + "tokens_seen": 1321533440 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003028184553660983, + "loss": 1.4593, + "theoretical_loss": 3.5845062202880538, + "tokens_seen": 1321598976 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030280842527582747, + "loss": 1.4071, + "theoretical_loss": 3.5844902930205422, + "tokens_seen": 1321664512 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003027983951855567, + "loss": 1.3544, + "theoretical_loss": 3.584474366763903, + "tokens_seen": 1321730048 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030278836509528583, + "loss": 1.4185, + "theoretical_loss": 3.584458441518022, + "tokens_seen": 1321795584 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030277833500501507, + "loss": 1.4896, + "theoretical_loss": 3.584442517282784, + "tokens_seen": 1321861120 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003027683049147442, + "loss": 1.4569, + "theoretical_loss": 3.584426594058075, + "tokens_seen": 1321926656 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030275827482447343, + "loss": 1.354, + "theoretical_loss": 3.5844106718437807, + "tokens_seen": 1321992192 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003027482447342026, + "loss": 1.3603, + "theoretical_loss": 3.584394750639788, + "tokens_seen": 1322057728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003027382146439318, + "loss": 1.3652, + "theoretical_loss": 3.5843788304459814, + "tokens_seen": 1322123264 + }, + { + "epoch": 0.4, + "learning_rate": 0.000302728184553661, + "loss": 1.4732, + "theoretical_loss": 3.5843629112622475, + "tokens_seen": 1322188800 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003027181544633902, + "loss": 1.2148, + "theoretical_loss": 3.584346993088472, + "tokens_seen": 1322254336 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030270812437311934, + "loss": 1.4106, + "theoretical_loss": 3.58433107592454, + "tokens_seen": 1322319872 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030269809428284857, + "loss": 1.2846, + "theoretical_loss": 3.5843151597703393, + "tokens_seen": 1322385408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026880641925777, + "loss": 1.4567, + "theoretical_loss": 3.5842992446257536, + "tokens_seen": 1322450944 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030267803410230693, + "loss": 1.4007, + "theoretical_loss": 3.58428333049067, + "tokens_seen": 1322516480 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026680040120361, + "loss": 1.4045, + "theoretical_loss": 3.5842674173649742, + "tokens_seen": 1322582016 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026579739217653, + "loss": 1.4129, + "theoretical_loss": 3.5842515052485524, + "tokens_seen": 1322647552 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026479438314945, + "loss": 1.4495, + "theoretical_loss": 3.5842355941412904, + "tokens_seen": 1322713088 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030263791374122366, + "loss": 1.4172, + "theoretical_loss": 3.5842196840430742, + "tokens_seen": 1322778624 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030262788365095284, + "loss": 1.4151, + "theoretical_loss": 3.5842037749537896, + "tokens_seen": 1322844160 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026178535606821, + "loss": 1.4229, + "theoretical_loss": 3.5841878668733234, + "tokens_seen": 1322909696 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003026078234704112, + "loss": 1.4502, + "theoretical_loss": 3.5841719598015604, + "tokens_seen": 1322975232 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030259779338014044, + "loss": 1.4404, + "theoretical_loss": 3.584156053738388, + "tokens_seen": 1323040768 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030258776328986956, + "loss": 1.3643, + "theoretical_loss": 3.5841401486836917, + "tokens_seen": 1323106304 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003025777331995988, + "loss": 1.3837, + "theoretical_loss": 3.5841242446373576, + "tokens_seen": 1323171840 + }, + { + "epoch": 0.4, + "learning_rate": 0.000302567703109328, + "loss": 1.3925, + "theoretical_loss": 3.584108341599271, + "tokens_seen": 1323237376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030255767301905716, + "loss": 1.4714, + "theoretical_loss": 3.5840924395693197, + "tokens_seen": 1323302912 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003025476429287864, + "loss": 1.3206, + "theoretical_loss": 3.5840765385473894, + "tokens_seen": 1323368448 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003025376128385156, + "loss": 1.4278, + "theoretical_loss": 3.584060638533365, + "tokens_seen": 1323433984 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030252758274824476, + "loss": 1.3995, + "theoretical_loss": 3.584044739527134, + "tokens_seen": 1323499520 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030251755265797394, + "loss": 1.5273, + "theoretical_loss": 3.584028841528583, + "tokens_seen": 1323565056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003025075225677031, + "loss": 1.3855, + "theoretical_loss": 3.5840129445375966, + "tokens_seen": 1323630592 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003024974924774323, + "loss": 1.4354, + "theoretical_loss": 3.583997048554062, + "tokens_seen": 1323696128 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030248746238716154, + "loss": 1.4105, + "theoretical_loss": 3.5839811535778656, + "tokens_seen": 1323761664 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 484332, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4385075569152832, + "objective/train/theoretical_loss": 3.583965259608894, + "objective/train/tokens_used": 1344287200, + "theoretical_loss": 3.583965259608894, + "tokens_seen": 1323827200 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030247743229689066, + "loss": 1.4113, + "theoretical_loss": 3.583965259608894, + "tokens_seen": 1323827200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003024674022066199, + "loss": 1.3397, + "theoretical_loss": 3.583949366647033, + "tokens_seen": 1323892736 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030245737211634903, + "loss": 1.3868, + "theoretical_loss": 3.5839334746921683, + "tokens_seen": 1323958272 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030244734202607826, + "loss": 1.3468, + "theoretical_loss": 3.5839175837441877, + "tokens_seen": 1324023808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030243731193580744, + "loss": 1.5082, + "theoretical_loss": 3.583901693802977, + "tokens_seen": 1324089344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003024272818455366, + "loss": 1.3419, + "theoretical_loss": 3.583885804868422, + "tokens_seen": 1324154880 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003024172517552658, + "loss": 1.2941, + "theoretical_loss": 3.5838699169404094, + "tokens_seen": 1324220416 + }, + { + "epoch": 0.4, + "learning_rate": 0.000302407221664995, + "loss": 1.5615, + "theoretical_loss": 3.583854030018826, + "tokens_seen": 1324285952 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030239719157472417, + "loss": 1.3827, + "theoretical_loss": 3.583838144103558, + "tokens_seen": 1324351488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003023871614844534, + "loss": 1.3755, + "theoretical_loss": 3.583822259194492, + "tokens_seen": 1324417024 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030237713139418253, + "loss": 1.3493, + "theoretical_loss": 3.5838063752915144, + "tokens_seen": 1324482560 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030236710130391177, + "loss": 1.5209, + "theoretical_loss": 3.5837904923945114, + "tokens_seen": 1324548096 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030235707121364095, + "loss": 1.3844, + "theoretical_loss": 3.5837746105033697, + "tokens_seen": 1324613632 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030234704112337013, + "loss": 1.4778, + "theoretical_loss": 3.5837587296179763, + "tokens_seen": 1324679168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003023370110330993, + "loss": 1.517, + "theoretical_loss": 3.5837428497382167, + "tokens_seen": 1324744704 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003023269809428285, + "loss": 1.4613, + "theoretical_loss": 3.583726970863979, + "tokens_seen": 1324810240 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030231695085255767, + "loss": 1.5094, + "theoretical_loss": 3.5837110929951486, + "tokens_seen": 1324875776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003023069207622869, + "loss": 1.4015, + "theoretical_loss": 3.5836952161316122, + "tokens_seen": 1324941312 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030229689067201603, + "loss": 1.3941, + "theoretical_loss": 3.5836793402732567, + "tokens_seen": 1325006848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030228686058174527, + "loss": 1.3789, + "theoretical_loss": 3.5836634654199693, + "tokens_seen": 1325072384 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003022768304914744, + "loss": 1.3721, + "theoretical_loss": 3.5836475915716357, + "tokens_seen": 1325137920 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030226680040120363, + "loss": 1.4166, + "theoretical_loss": 3.5836317187281432, + "tokens_seen": 1325203456 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003022567703109328, + "loss": 1.4305, + "theoretical_loss": 3.583615846889378, + "tokens_seen": 1325268992 + }, + { + "epoch": 0.4, + "learning_rate": 0.000302246740220662, + "loss": 1.3924, + "theoretical_loss": 3.5835999760552273, + "tokens_seen": 1325334528 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003022367101303912, + "loss": 1.4529, + "theoretical_loss": 3.5835841062255773, + "tokens_seen": 1325400064 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003022266800401204, + "loss": 1.345, + "theoretical_loss": 3.583568237400316, + "tokens_seen": 1325465600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030221664994984954, + "loss": 1.4085, + "theoretical_loss": 3.583552369579328, + "tokens_seen": 1325531136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030220661985957877, + "loss": 1.2958, + "theoretical_loss": 3.5835365027625024, + "tokens_seen": 1325596672 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021965897693079, + "loss": 1.5059, + "theoretical_loss": 3.583520636949725, + "tokens_seen": 1325662208 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030218655967903713, + "loss": 1.3527, + "theoretical_loss": 3.5835047721408824, + "tokens_seen": 1325727744 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021765295887663, + "loss": 1.479, + "theoretical_loss": 3.583488908335861, + "tokens_seen": 1325793280 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021664994984955, + "loss": 1.4099, + "theoretical_loss": 3.583473045534549, + "tokens_seen": 1325858816 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021564694082247, + "loss": 1.2993, + "theoretical_loss": 3.583457183736833, + "tokens_seen": 1325924352 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030214643931795386, + "loss": 1.3726, + "theoretical_loss": 3.583441322942599, + "tokens_seen": 1325989888 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030213640922768304, + "loss": 1.4664, + "theoretical_loss": 3.583425463151735, + "tokens_seen": 1326055424 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021263791374123, + "loss": 1.3791, + "theoretical_loss": 3.5834096043641264, + "tokens_seen": 1326120960 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003021163490471414, + "loss": 1.4466, + "theoretical_loss": 3.583393746579662, + "tokens_seen": 1326186496 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030210631895687064, + "loss": 1.4498, + "theoretical_loss": 3.5833778897982276, + "tokens_seen": 1326252032 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030209628886659976, + "loss": 1.4914, + "theoretical_loss": 3.5833620340197108, + "tokens_seen": 1326317568 + }, + { + "epoch": 0.4, + "learning_rate": 0.000302086258776329, + "loss": 1.4777, + "theoretical_loss": 3.5833461792439985, + "tokens_seen": 1326383104 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003020762286860582, + "loss": 1.4246, + "theoretical_loss": 3.5833303254709774, + "tokens_seen": 1326448640 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030206619859578736, + "loss": 1.4569, + "theoretical_loss": 3.5833144727005344, + "tokens_seen": 1326514176 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030205616850551654, + "loss": 1.4418, + "theoretical_loss": 3.583298620932557, + "tokens_seen": 1326579712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003020461384152458, + "loss": 1.4166, + "theoretical_loss": 3.583282770166933, + "tokens_seen": 1326645248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003020361083249749, + "loss": 1.347, + "theoretical_loss": 3.583266920403548, + "tokens_seen": 1326710784 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030202607823470414, + "loss": 1.3504, + "theoretical_loss": 3.58325107164229, + "tokens_seen": 1326776320 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030201604814443327, + "loss": 1.4303, + "theoretical_loss": 3.583235223883046, + "tokens_seen": 1326841856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003020060180541625, + "loss": 1.4248, + "theoretical_loss": 3.5832193771257033, + "tokens_seen": 1326907392 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003019959879638917, + "loss": 1.4386, + "theoretical_loss": 3.583203531370149, + "tokens_seen": 1326972928 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030198595787362087, + "loss": 1.3984, + "theoretical_loss": 3.5831876866162693, + "tokens_seen": 1327038464 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 485470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.484735369682312, + "objective/train/theoretical_loss": 3.5831718428639534, + "objective/train/tokens_used": 1347564000, + "theoretical_loss": 3.5831718428639534, + "tokens_seen": 1327104000 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030197592778335005, + "loss": 1.5138, + "theoretical_loss": 3.5831718428639534, + "tokens_seen": 1327104000 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030196589769307923, + "loss": 1.3163, + "theoretical_loss": 3.583156000113087, + "tokens_seen": 1327169536 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003019558676028084, + "loss": 1.4776, + "theoretical_loss": 3.583140158363558, + "tokens_seen": 1327235072 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030194583751253764, + "loss": 1.3715, + "theoretical_loss": 3.5831243176152534, + "tokens_seen": 1327300608 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030193580742226677, + "loss": 1.531, + "theoretical_loss": 3.583108477868061, + "tokens_seen": 1327366144 + }, + { + "epoch": 0.4, + "learning_rate": 0.000301925777331996, + "loss": 1.3158, + "theoretical_loss": 3.583092639121867, + "tokens_seen": 1327431680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030191574724172513, + "loss": 1.3382, + "theoretical_loss": 3.58307680137656, + "tokens_seen": 1327497216 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030190571715145437, + "loss": 1.4171, + "theoretical_loss": 3.5830609646320264, + "tokens_seen": 1327562752 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030189568706118355, + "loss": 1.4703, + "theoretical_loss": 3.5830451288881537, + "tokens_seen": 1327628288 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030188565697091273, + "loss": 1.4825, + "theoretical_loss": 3.58302929414483, + "tokens_seen": 1327693824 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003018756268806419, + "loss": 1.3308, + "theoretical_loss": 3.583013460401942, + "tokens_seen": 1327759360 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030186559679037115, + "loss": 1.3854, + "theoretical_loss": 3.5829976276593776, + "tokens_seen": 1327824896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003018555667001003, + "loss": 1.4411, + "theoretical_loss": 3.582981795917024, + "tokens_seen": 1327890432 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003018455366098295, + "loss": 1.4532, + "theoretical_loss": 3.5829659651747683, + "tokens_seen": 1327955968 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030183550651955864, + "loss": 1.4622, + "theoretical_loss": 3.5829501354324984, + "tokens_seen": 1328021504 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030182547642928787, + "loss": 1.4131, + "theoretical_loss": 3.5829343066901016, + "tokens_seen": 1328087040 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030181544633901705, + "loss": 1.3633, + "theoretical_loss": 3.582918478947466, + "tokens_seen": 1328152576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030180541624874623, + "loss": 1.4493, + "theoretical_loss": 3.582902652204478, + "tokens_seen": 1328218112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030179538615847547, + "loss": 1.3362, + "theoretical_loss": 3.582886826461026, + "tokens_seen": 1328283648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003017853560682046, + "loss": 1.3754, + "theoretical_loss": 3.582871001716997, + "tokens_seen": 1328349184 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030177532597793383, + "loss": 1.4855, + "theoretical_loss": 3.582855177972279, + "tokens_seen": 1328414720 + }, + { + "epoch": 0.4, + "learning_rate": 0.000301765295887663, + "loss": 1.3959, + "theoretical_loss": 3.58283935522676, + "tokens_seen": 1328480256 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003017552657973922, + "loss": 1.3929, + "theoretical_loss": 3.582823533480327, + "tokens_seen": 1328545792 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003017452357071214, + "loss": 1.3393, + "theoretical_loss": 3.5828077127328672, + "tokens_seen": 1328611328 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003017352056168506, + "loss": 1.3582, + "theoretical_loss": 3.5827918929842695, + "tokens_seen": 1328676864 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030172517552657974, + "loss": 1.3713, + "theoretical_loss": 3.5827760742344203, + "tokens_seen": 1328742400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030171514543630897, + "loss": 1.4705, + "theoretical_loss": 3.5827602564832084, + "tokens_seen": 1328807936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003017051153460381, + "loss": 1.4144, + "theoretical_loss": 3.582744439730521, + "tokens_seen": 1328873472 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030169508525576733, + "loss": 1.4151, + "theoretical_loss": 3.5827286239762453, + "tokens_seen": 1328939008 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003016850551654965, + "loss": 1.3932, + "theoretical_loss": 3.58271280922027, + "tokens_seen": 1329004544 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003016750250752257, + "loss": 1.5006, + "theoretical_loss": 3.582696995462482, + "tokens_seen": 1329070080 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003016649949849549, + "loss": 1.3758, + "theoretical_loss": 3.5826811827027703, + "tokens_seen": 1329135616 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030165496489468406, + "loss": 1.3622, + "theoretical_loss": 3.582665370941021, + "tokens_seen": 1329201152 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030164493480441324, + "loss": 1.5078, + "theoretical_loss": 3.5826495601771233, + "tokens_seen": 1329266688 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003016349047141425, + "loss": 1.3061, + "theoretical_loss": 3.582633750410964, + "tokens_seen": 1329332224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003016248746238716, + "loss": 1.3909, + "theoretical_loss": 3.5826179416424324, + "tokens_seen": 1329397760 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030161484453360084, + "loss": 1.2959, + "theoretical_loss": 3.582602133871415, + "tokens_seen": 1329463296 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030160481444332996, + "loss": 1.4638, + "theoretical_loss": 3.5825863270978, + "tokens_seen": 1329528832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003015947843530592, + "loss": 1.3812, + "theoretical_loss": 3.582570521321476, + "tokens_seen": 1329594368 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003015847542627884, + "loss": 1.3174, + "theoretical_loss": 3.58255471654233, + "tokens_seen": 1329659904 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030157472417251756, + "loss": 1.3092, + "theoretical_loss": 3.58253891276025, + "tokens_seen": 1329725440 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030156469408224674, + "loss": 1.2868, + "theoretical_loss": 3.582523109975125, + "tokens_seen": 1329790976 + }, + { + "epoch": 0.4, + "learning_rate": 0.000301554663991976, + "loss": 1.4141, + "theoretical_loss": 3.582507308186842, + "tokens_seen": 1329856512 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003015446339017051, + "loss": 1.3649, + "theoretical_loss": 3.5824915073952894, + "tokens_seen": 1329922048 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030153460381143434, + "loss": 1.4829, + "theoretical_loss": 3.582475707600355, + "tokens_seen": 1329987584 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030152457372116347, + "loss": 1.3213, + "theoretical_loss": 3.582459908801927, + "tokens_seen": 1330053120 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003015145436308927, + "loss": 1.3482, + "theoretical_loss": 3.5824441109998935, + "tokens_seen": 1330118656 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003015045135406219, + "loss": 1.4378, + "theoretical_loss": 3.582428314194142, + "tokens_seen": 1330184192 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030149448345035107, + "loss": 1.3412, + "theoretical_loss": 3.5824125183845617, + "tokens_seen": 1330249728 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030148445336008025, + "loss": 1.4569, + "theoretical_loss": 3.5823967235710397, + "tokens_seen": 1330315264 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 486514, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4111610651016235, + "objective/train/theoretical_loss": 3.582380929753465, + "objective/train/tokens_used": 1350840800, + "theoretical_loss": 3.582380929753465, + "tokens_seen": 1330380800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030147442326980943, + "loss": 1.3567, + "theoretical_loss": 3.582380929753465, + "tokens_seen": 1330380800 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003014643931795386, + "loss": 1.4954, + "theoretical_loss": 3.5823651369317244, + "tokens_seen": 1330446336 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030145436308926784, + "loss": 1.432, + "theoretical_loss": 3.5823493451057074, + "tokens_seen": 1330511872 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030144433299899697, + "loss": 1.3623, + "theoretical_loss": 3.5823335542753014, + "tokens_seen": 1330577408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003014343029087262, + "loss": 1.36, + "theoretical_loss": 3.582317764440395, + "tokens_seen": 1330642944 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030142427281845533, + "loss": 1.444, + "theoretical_loss": 3.5823019756008767, + "tokens_seen": 1330708480 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030141424272818457, + "loss": 1.4404, + "theoretical_loss": 3.582286187756634, + "tokens_seen": 1330774016 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030140421263791375, + "loss": 1.3721, + "theoretical_loss": 3.5822704009075554, + "tokens_seen": 1330839552 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030139418254764293, + "loss": 1.3665, + "theoretical_loss": 3.5822546150535293, + "tokens_seen": 1330905088 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003013841524573721, + "loss": 1.4776, + "theoretical_loss": 3.582238830194444, + "tokens_seen": 1330970624 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030137412236710135, + "loss": 1.4566, + "theoretical_loss": 3.5822230463301876, + "tokens_seen": 1331036160 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003013640922768305, + "loss": 1.3538, + "theoretical_loss": 3.5822072634606483, + "tokens_seen": 1331101696 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003013540621865597, + "loss": 1.4143, + "theoretical_loss": 3.582191481585715, + "tokens_seen": 1331167232 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030134403209628884, + "loss": 1.4227, + "theoretical_loss": 3.5821757007052764, + "tokens_seen": 1331232768 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030133400200601807, + "loss": 1.3895, + "theoretical_loss": 3.5821599208192194, + "tokens_seen": 1331298304 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030132397191574725, + "loss": 1.3595, + "theoretical_loss": 3.5821441419274334, + "tokens_seen": 1331363840 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030131394182547643, + "loss": 1.4476, + "theoretical_loss": 3.582128364029807, + "tokens_seen": 1331429376 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003013039117352056, + "loss": 1.477, + "theoretical_loss": 3.582112587126228, + "tokens_seen": 1331494912 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012938816449348, + "loss": 1.4082, + "theoretical_loss": 3.582096811216585, + "tokens_seen": 1331560448 + }, + { + "epoch": 0.4, + "learning_rate": 0.000301283851554664, + "loss": 1.3626, + "theoretical_loss": 3.5820810363007665, + "tokens_seen": 1331625984 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012738214643932, + "loss": 1.3563, + "theoretical_loss": 3.5820652623786615, + "tokens_seen": 1331691520 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030126379137412234, + "loss": 1.4377, + "theoretical_loss": 3.582049489450158, + "tokens_seen": 1331757056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012537612838516, + "loss": 1.4728, + "theoretical_loss": 3.582033717515144, + "tokens_seen": 1331822592 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012437311935807, + "loss": 1.4448, + "theoretical_loss": 3.58201794657351, + "tokens_seen": 1331888128 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030123370110330994, + "loss": 1.3415, + "theoretical_loss": 3.582002176625142, + "tokens_seen": 1331953664 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012236710130391, + "loss": 1.3385, + "theoretical_loss": 3.5819864076699304, + "tokens_seen": 1332019200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012136409227683, + "loss": 1.2442, + "theoretical_loss": 3.581970639707763, + "tokens_seen": 1332084736 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003012036108324975, + "loss": 1.3399, + "theoretical_loss": 3.5819548727385286, + "tokens_seen": 1332150272 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003011935807422267, + "loss": 1.4284, + "theoretical_loss": 3.581939106762116, + "tokens_seen": 1332215808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030118355065195584, + "loss": 1.2861, + "theoretical_loss": 3.5819233417784133, + "tokens_seen": 1332281344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003011735205616851, + "loss": 1.558, + "theoretical_loss": 3.5819075777873097, + "tokens_seen": 1332346880 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003011634904714142, + "loss": 1.3892, + "theoretical_loss": 3.581891814788694, + "tokens_seen": 1332412416 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030115346038114344, + "loss": 1.4107, + "theoretical_loss": 3.5818760527824542, + "tokens_seen": 1332477952 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003011434302908726, + "loss": 1.4755, + "theoretical_loss": 3.5818602917684794, + "tokens_seen": 1332543488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003011334002006018, + "loss": 1.3791, + "theoretical_loss": 3.581844531746659, + "tokens_seen": 1332609024 + }, + { + "epoch": 0.4, + "learning_rate": 0.000301123370110331, + "loss": 1.3777, + "theoretical_loss": 3.5818287727168805, + "tokens_seen": 1332674560 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030111334002006016, + "loss": 1.4388, + "theoretical_loss": 3.5818130146790335, + "tokens_seen": 1332740096 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030110330992978935, + "loss": 1.3594, + "theoretical_loss": 3.581797257633007, + "tokens_seen": 1332805632 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003010932798395186, + "loss": 1.2805, + "theoretical_loss": 3.581781501578689, + "tokens_seen": 1332871168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003010832497492477, + "loss": 1.4209, + "theoretical_loss": 3.581765746515969, + "tokens_seen": 1332936704 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030107321965897694, + "loss": 1.4671, + "theoretical_loss": 3.5817499924447356, + "tokens_seen": 1333002240 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030106318956870607, + "loss": 1.4728, + "theoretical_loss": 3.5817342393648772, + "tokens_seen": 1333067776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003010531594784353, + "loss": 1.3859, + "theoretical_loss": 3.5817184872762833, + "tokens_seen": 1333133312 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030104312938816454, + "loss": 1.3458, + "theoretical_loss": 3.581702736178843, + "tokens_seen": 1333198848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030103309929789367, + "loss": 1.4805, + "theoretical_loss": 3.5816869860724445, + "tokens_seen": 1333264384 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003010230692076229, + "loss": 1.2637, + "theoretical_loss": 3.581671236956977, + "tokens_seen": 1333329920 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003010130391173521, + "loss": 1.4525, + "theoretical_loss": 3.5816554888323298, + "tokens_seen": 1333395456 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030100300902708127, + "loss": 1.3893, + "theoretical_loss": 3.5816397416983916, + "tokens_seen": 1333460992 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030099297893681045, + "loss": 1.3959, + "theoretical_loss": 3.5816239955550513, + "tokens_seen": 1333526528 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030098294884653963, + "loss": 1.373, + "theoretical_loss": 3.581608250402198, + "tokens_seen": 1333592064 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 487665, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3523489236831665, + "objective/train/theoretical_loss": 3.581592506239721, + "objective/train/tokens_used": 1354117600, + "theoretical_loss": 3.581592506239721, + "tokens_seen": 1333657600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003009729187562688, + "loss": 1.3593, + "theoretical_loss": 3.581592506239721, + "tokens_seen": 1333657600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030096288866599804, + "loss": 1.458, + "theoretical_loss": 3.581576763067509, + "tokens_seen": 1333723136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030095285857572717, + "loss": 1.4316, + "theoretical_loss": 3.581561020885451, + "tokens_seen": 1333788672 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003009428284854564, + "loss": 1.4219, + "theoretical_loss": 3.5815452796934357, + "tokens_seen": 1333854208 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030093279839518553, + "loss": 1.3444, + "theoretical_loss": 3.5815295394913536, + "tokens_seen": 1333919744 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030092276830491477, + "loss": 1.2993, + "theoretical_loss": 3.5815138002790925, + "tokens_seen": 1333985280 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030091273821464395, + "loss": 1.3006, + "theoretical_loss": 3.581498062056542, + "tokens_seen": 1334050816 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030090270812437313, + "loss": 1.3445, + "theoretical_loss": 3.581482324823591, + "tokens_seen": 1334116352 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003008926780341023, + "loss": 1.4211, + "theoretical_loss": 3.5814665885801293, + "tokens_seen": 1334181888 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030088264794383155, + "loss": 1.4998, + "theoretical_loss": 3.581450853326045, + "tokens_seen": 1334247424 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003008726178535607, + "loss": 1.4737, + "theoretical_loss": 3.5814351190612284, + "tokens_seen": 1334312960 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003008625877632899, + "loss": 1.4636, + "theoretical_loss": 3.5814193857855683, + "tokens_seen": 1334378496 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030085255767301904, + "loss": 1.351, + "theoretical_loss": 3.581403653498954, + "tokens_seen": 1334444032 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030084252758274827, + "loss": 1.3552, + "theoretical_loss": 3.5813879222012748, + "tokens_seen": 1334509568 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030083249749247745, + "loss": 1.4397, + "theoretical_loss": 3.5813721918924193, + "tokens_seen": 1334575104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030082246740220663, + "loss": 1.3992, + "theoretical_loss": 3.581356462572278, + "tokens_seen": 1334640640 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003008124373119358, + "loss": 1.4572, + "theoretical_loss": 3.5813407342407384, + "tokens_seen": 1334706176 + }, + { + "epoch": 0.4, + "learning_rate": 0.000300802407221665, + "loss": 1.4269, + "theoretical_loss": 3.581325006897692, + "tokens_seen": 1334771712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007923771313942, + "loss": 1.314, + "theoretical_loss": 3.581309280543027, + "tokens_seen": 1334837248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007823470411234, + "loss": 1.3095, + "theoretical_loss": 3.5812935551766327, + "tokens_seen": 1334902784 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030077231695085254, + "loss": 1.3284, + "theoretical_loss": 3.581277830798399, + "tokens_seen": 1334968320 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007622868605818, + "loss": 1.3382, + "theoretical_loss": 3.5812621074082145, + "tokens_seen": 1335033856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007522567703109, + "loss": 1.4331, + "theoretical_loss": 3.5812463850059695, + "tokens_seen": 1335099392 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030074222668004014, + "loss": 1.4166, + "theoretical_loss": 3.5812306635915525, + "tokens_seen": 1335164928 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007321965897693, + "loss": 1.4871, + "theoretical_loss": 3.581214943164854, + "tokens_seen": 1335230464 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007221664994985, + "loss": 1.2961, + "theoretical_loss": 3.581199223725762, + "tokens_seen": 1335296000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007121364092277, + "loss": 1.3331, + "theoretical_loss": 3.5811835052741685, + "tokens_seen": 1335361536 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003007021063189569, + "loss": 1.3608, + "theoretical_loss": 3.58116778780996, + "tokens_seen": 1335427072 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030069207622868604, + "loss": 1.4132, + "theoretical_loss": 3.5811520713330283, + "tokens_seen": 1335492608 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003006820461384153, + "loss": 1.2784, + "theoretical_loss": 3.5811363558432623, + "tokens_seen": 1335558144 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003006720160481444, + "loss": 1.4417, + "theoretical_loss": 3.5811206413405507, + "tokens_seen": 1335623680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030066198595787364, + "loss": 1.3343, + "theoretical_loss": 3.581104927824784, + "tokens_seen": 1335689216 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003006519558676028, + "loss": 1.3503, + "theoretical_loss": 3.581089215295852, + "tokens_seen": 1335754752 + }, + { + "epoch": 0.4, + "learning_rate": 0.000300641925777332, + "loss": 1.2874, + "theoretical_loss": 3.581073503753643, + "tokens_seen": 1335820288 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003006318956870612, + "loss": 1.3512, + "theoretical_loss": 3.5810577931980476, + "tokens_seen": 1335885824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030062186559679036, + "loss": 1.327, + "theoretical_loss": 3.5810420836289554, + "tokens_seen": 1335951360 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030061183550651955, + "loss": 1.3075, + "theoretical_loss": 3.5810263750462568, + "tokens_seen": 1336016896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003006018054162488, + "loss": 1.3907, + "theoretical_loss": 3.58101066744984, + "tokens_seen": 1336082432 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003005917753259779, + "loss": 1.3309, + "theoretical_loss": 3.580994960839595, + "tokens_seen": 1336147968 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030058174523570714, + "loss": 1.2837, + "theoretical_loss": 3.5809792552154125, + "tokens_seen": 1336213504 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030057171514543627, + "loss": 1.303, + "theoretical_loss": 3.5809635505771817, + "tokens_seen": 1336279040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003005616850551655, + "loss": 1.423, + "theoretical_loss": 3.5809478469247917, + "tokens_seen": 1336344576 + }, + { + "epoch": 0.4, + "learning_rate": 0.0003005516549648947, + "loss": 1.2671, + "theoretical_loss": 3.580932144258133, + "tokens_seen": 1336410112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00030054162487462387, + "loss": 1.3545, + "theoretical_loss": 3.580916442577096, + "tokens_seen": 1336475648 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030053159478435305, + "loss": 1.4537, + "theoretical_loss": 3.5809007418815693, + "tokens_seen": 1336541184 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003005215646940823, + "loss": 1.3785, + "theoretical_loss": 3.580885042171443, + "tokens_seen": 1336606720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003005115346038114, + "loss": 1.3948, + "theoretical_loss": 3.5808693434466075, + "tokens_seen": 1336672256 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030050150451354065, + "loss": 1.3145, + "theoretical_loss": 3.580853645706952, + "tokens_seen": 1336737792 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003004914744232698, + "loss": 1.3829, + "theoretical_loss": 3.580837948952367, + "tokens_seen": 1336803328 + }, + { + "epoch": 0.41, + "learning_rate": 0.000300481444332999, + "loss": 1.565, + "theoretical_loss": 3.580822253182742, + "tokens_seen": 1336868864 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 488954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2186024188995361, + "objective/train/theoretical_loss": 3.5808065583979665, + "objective/train/tokens_used": 1357394400, + "theoretical_loss": 3.5808065583979665, + "tokens_seen": 1336934400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003004714142427282, + "loss": 1.3343, + "theoretical_loss": 3.5808065583979665, + "tokens_seen": 1336934400 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030046138415245737, + "loss": 1.3774, + "theoretical_loss": 3.5807908645979314, + "tokens_seen": 1336999936 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030045135406218655, + "loss": 1.4031, + "theoretical_loss": 3.580775171782526, + "tokens_seen": 1337065472 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030044132397191573, + "loss": 1.3839, + "theoretical_loss": 3.5807594799516402, + "tokens_seen": 1337131008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003004312938816449, + "loss": 1.3771, + "theoretical_loss": 3.5807437891051643, + "tokens_seen": 1337196544 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030042126379137415, + "loss": 1.3133, + "theoretical_loss": 3.5807280992429886, + "tokens_seen": 1337262080 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003004112337011033, + "loss": 1.3101, + "theoretical_loss": 3.5807124103650025, + "tokens_seen": 1337327616 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003004012036108325, + "loss": 1.4693, + "theoretical_loss": 3.580696722471096, + "tokens_seen": 1337393152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003003911735205617, + "loss": 1.4483, + "theoretical_loss": 3.58068103556116, + "tokens_seen": 1337458688 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003003811434302909, + "loss": 1.3109, + "theoretical_loss": 3.5806653496350838, + "tokens_seen": 1337524224 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030037111334002006, + "loss": 1.4612, + "theoretical_loss": 3.5806496646927575, + "tokens_seen": 1337589760 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030036108324974924, + "loss": 1.4151, + "theoretical_loss": 3.5806339807340715, + "tokens_seen": 1337655296 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003003510531594784, + "loss": 1.3924, + "theoretical_loss": 3.5806182977589156, + "tokens_seen": 1337720832 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030034102306920765, + "loss": 1.4416, + "theoretical_loss": 3.580602615767181, + "tokens_seen": 1337786368 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003003309929789368, + "loss": 1.3684, + "theoretical_loss": 3.5805869347587564, + "tokens_seen": 1337851904 + }, + { + "epoch": 0.41, + "learning_rate": 0.000300320962888666, + "loss": 1.3952, + "theoretical_loss": 3.580571254733533, + "tokens_seen": 1337917440 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030031093279839514, + "loss": 1.4065, + "theoretical_loss": 3.5805555756914, + "tokens_seen": 1337982976 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003003009027081244, + "loss": 1.3119, + "theoretical_loss": 3.5805398976322484, + "tokens_seen": 1338048512 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002908726178536, + "loss": 1.4242, + "theoretical_loss": 3.5805242205559686, + "tokens_seen": 1338114048 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030028084252758274, + "loss": 1.3252, + "theoretical_loss": 3.5805085444624503, + "tokens_seen": 1338179584 + }, + { + "epoch": 0.41, + "learning_rate": 0.000300270812437312, + "loss": 1.3908, + "theoretical_loss": 3.580492869351584, + "tokens_seen": 1338245120 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002607823470411, + "loss": 1.2952, + "theoretical_loss": 3.58047719522326, + "tokens_seen": 1338310656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030025075225677034, + "loss": 1.3753, + "theoretical_loss": 3.5804615220773686, + "tokens_seen": 1338376192 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002407221664995, + "loss": 1.3544, + "theoretical_loss": 3.5804458499138, + "tokens_seen": 1338441728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002306920762287, + "loss": 1.3921, + "theoretical_loss": 3.5804301787324446, + "tokens_seen": 1338507264 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002206619859579, + "loss": 1.3909, + "theoretical_loss": 3.5804145085331927, + "tokens_seen": 1338572800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003002106318956871, + "loss": 1.2724, + "theoretical_loss": 3.580398839315935, + "tokens_seen": 1338638336 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030020060180541624, + "loss": 1.2966, + "theoretical_loss": 3.5803831710805616, + "tokens_seen": 1338703872 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003001905717151455, + "loss": 1.367, + "theoretical_loss": 3.580367503826963, + "tokens_seen": 1338769408 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003001805416248746, + "loss": 1.4275, + "theoretical_loss": 3.5803518375550296, + "tokens_seen": 1338834944 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030017051153460384, + "loss": 1.3832, + "theoretical_loss": 3.580336172264652, + "tokens_seen": 1338900480 + }, + { + "epoch": 0.41, + "learning_rate": 0.000300160481444333, + "loss": 1.4251, + "theoretical_loss": 3.5803205079557197, + "tokens_seen": 1338966016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003001504513540622, + "loss": 1.3821, + "theoretical_loss": 3.5803048446281247, + "tokens_seen": 1339031552 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003001404212637914, + "loss": 1.3425, + "theoretical_loss": 3.5802891822817564, + "tokens_seen": 1339097088 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030013039117352056, + "loss": 1.3667, + "theoretical_loss": 3.580273520916506, + "tokens_seen": 1339162624 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030012036108324975, + "loss": 1.4295, + "theoretical_loss": 3.5802578605322637, + "tokens_seen": 1339228160 + }, + { + "epoch": 0.41, + "learning_rate": 0.000300110330992979, + "loss": 1.3386, + "theoretical_loss": 3.58024220112892, + "tokens_seen": 1339293696 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003001003009027081, + "loss": 1.4143, + "theoretical_loss": 3.580226542706365, + "tokens_seen": 1339359232 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030009027081243734, + "loss": 1.408, + "theoretical_loss": 3.5802108852644903, + "tokens_seen": 1339424768 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030008024072216647, + "loss": 1.4166, + "theoretical_loss": 3.580195228803186, + "tokens_seen": 1339490304 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003000702106318957, + "loss": 1.3504, + "theoretical_loss": 3.5801795733223427, + "tokens_seen": 1339555840 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003000601805416249, + "loss": 1.4448, + "theoretical_loss": 3.5801639188218513, + "tokens_seen": 1339621376 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030005015045135407, + "loss": 1.351, + "theoretical_loss": 3.5801482653016015, + "tokens_seen": 1339686912 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030004012036108325, + "loss": 1.4343, + "theoretical_loss": 3.5801326127614854, + "tokens_seen": 1339752448 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003000300902708125, + "loss": 1.4068, + "theoretical_loss": 3.5801169612013926, + "tokens_seen": 1339817984 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003000200601805416, + "loss": 1.2409, + "theoretical_loss": 3.580101310621214, + "tokens_seen": 1339883520 + }, + { + "epoch": 0.41, + "learning_rate": 0.00030001003009027085, + "loss": 1.3399, + "theoretical_loss": 3.580085661020841, + "tokens_seen": 1339949056 + }, + { + "epoch": 0.41, + "learning_rate": 0.0003, + "loss": 1.4657, + "theoretical_loss": 3.5800700124001636, + "tokens_seen": 1340014592 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002999899699097292, + "loss": 1.4037, + "theoretical_loss": 3.5800543647590732, + "tokens_seen": 1340080128 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002999799398194584, + "loss": 1.319, + "theoretical_loss": 3.5800387180974598, + "tokens_seen": 1340145664 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 490192, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3784217834472656, + "objective/train/theoretical_loss": 3.580023072415215, + "objective/train/tokens_used": 1360671200, + "theoretical_loss": 3.580023072415215, + "tokens_seen": 1340211200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029996990972918757, + "loss": 1.4038, + "theoretical_loss": 3.580023072415215, + "tokens_seen": 1340211200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029995987963891675, + "loss": 1.4503, + "theoretical_loss": 3.5800074277122285, + "tokens_seen": 1340276736 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029994984954864593, + "loss": 1.2704, + "theoretical_loss": 3.579991783988392, + "tokens_seen": 1340342272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002999398194583751, + "loss": 1.3725, + "theoretical_loss": 3.5799761412435966, + "tokens_seen": 1340407808 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029992978936810435, + "loss": 1.3945, + "theoretical_loss": 3.5799604994777328, + "tokens_seen": 1340473344 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002999197592778335, + "loss": 1.4597, + "theoretical_loss": 3.5799448586906912, + "tokens_seen": 1340538880 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002999097291875627, + "loss": 1.393, + "theoretical_loss": 3.579929218882363, + "tokens_seen": 1340604416 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002998996990972919, + "loss": 1.44, + "theoretical_loss": 3.5799135800526383, + "tokens_seen": 1340669952 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002998896690070211, + "loss": 1.5026, + "theoretical_loss": 3.57989794220141, + "tokens_seen": 1340735488 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029987963891675026, + "loss": 1.4503, + "theoretical_loss": 3.579882305328567, + "tokens_seen": 1340801024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029986960882647944, + "loss": 1.3903, + "theoretical_loss": 3.5798666694340016, + "tokens_seen": 1340866560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002998595787362086, + "loss": 1.1536, + "theoretical_loss": 3.5798510345176036, + "tokens_seen": 1340932096 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029984954864593785, + "loss": 1.4637, + "theoretical_loss": 3.579835400579265, + "tokens_seen": 1340997632 + }, + { + "epoch": 0.41, + "learning_rate": 0.000299839518555667, + "loss": 1.505, + "theoretical_loss": 3.5798197676188765, + "tokens_seen": 1341063168 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002998294884653962, + "loss": 1.3663, + "theoretical_loss": 3.57980413563633, + "tokens_seen": 1341128704 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029981945837512534, + "loss": 1.3909, + "theoretical_loss": 3.5797885046315145, + "tokens_seen": 1341194240 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002998094282848546, + "loss": 1.3826, + "theoretical_loss": 3.579772874604323, + "tokens_seen": 1341259776 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029979939819458376, + "loss": 1.416, + "theoretical_loss": 3.5797572455546454, + "tokens_seen": 1341325312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029978936810431294, + "loss": 1.48, + "theoretical_loss": 3.579741617482374, + "tokens_seen": 1341390848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002997793380140421, + "loss": 1.3819, + "theoretical_loss": 3.5797259903873986, + "tokens_seen": 1341456384 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002997693079237713, + "loss": 1.3071, + "theoretical_loss": 3.579710364269611, + "tokens_seen": 1341521920 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002997592778335005, + "loss": 1.4703, + "theoretical_loss": 3.579694739128902, + "tokens_seen": 1341587456 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002997492477432297, + "loss": 1.4725, + "theoretical_loss": 3.5796791149651637, + "tokens_seen": 1341652992 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029973921765295885, + "loss": 1.3667, + "theoretical_loss": 3.5796634917782866, + "tokens_seen": 1341718528 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002997291875626881, + "loss": 1.3591, + "theoretical_loss": 3.579647869568162, + "tokens_seen": 1341784064 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029971915747241726, + "loss": 1.4038, + "theoretical_loss": 3.579632248334681, + "tokens_seen": 1341849600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029970912738214644, + "loss": 1.4465, + "theoretical_loss": 3.579616628077735, + "tokens_seen": 1341915136 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996990972918756, + "loss": 1.4856, + "theoretical_loss": 3.579601008797215, + "tokens_seen": 1341980672 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996890672016048, + "loss": 1.3964, + "theoretical_loss": 3.579585390493013, + "tokens_seen": 1342046208 + }, + { + "epoch": 0.41, + "learning_rate": 0.000299679037111334, + "loss": 1.4991, + "theoretical_loss": 3.5795697731650193, + "tokens_seen": 1342111744 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996690070210632, + "loss": 1.5026, + "theoretical_loss": 3.5795541568131264, + "tokens_seen": 1342177280 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029965897693079235, + "loss": 1.3596, + "theoretical_loss": 3.5795385414372247, + "tokens_seen": 1342242816 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996489468405216, + "loss": 1.3736, + "theoretical_loss": 3.579522927037206, + "tokens_seen": 1342308352 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996389167502507, + "loss": 1.4225, + "theoretical_loss": 3.5795073136129614, + "tokens_seen": 1342373888 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029962888665997995, + "loss": 1.4771, + "theoretical_loss": 3.5794917011643825, + "tokens_seen": 1342439424 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029961885656970913, + "loss": 1.4136, + "theoretical_loss": 3.57947608969136, + "tokens_seen": 1342504960 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002996088264794383, + "loss": 1.5392, + "theoretical_loss": 3.579460479193787, + "tokens_seen": 1342570496 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995987963891675, + "loss": 1.3477, + "theoretical_loss": 3.579444869671553, + "tokens_seen": 1342636032 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029958876629889667, + "loss": 1.4235, + "theoretical_loss": 3.5794292611245506, + "tokens_seen": 1342701568 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029957873620862585, + "loss": 1.3877, + "theoretical_loss": 3.5794136535526713, + "tokens_seen": 1342767104 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995687061183551, + "loss": 1.3736, + "theoretical_loss": 3.5793980469558058, + "tokens_seen": 1342832640 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995586760280842, + "loss": 1.3983, + "theoretical_loss": 3.5793824413338466, + "tokens_seen": 1342898176 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029954864593781345, + "loss": 1.2562, + "theoretical_loss": 3.579366836686684, + "tokens_seen": 1342963712 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995386158475427, + "loss": 1.4877, + "theoretical_loss": 3.579351233014211, + "tokens_seen": 1343029248 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995285857572718, + "loss": 1.4193, + "theoretical_loss": 3.579335630316318, + "tokens_seen": 1343094784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029951855566700105, + "loss": 1.364, + "theoretical_loss": 3.5793200285928974, + "tokens_seen": 1343160320 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002995085255767302, + "loss": 1.3463, + "theoretical_loss": 3.57930442784384, + "tokens_seen": 1343225856 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994984954864594, + "loss": 1.4222, + "theoretical_loss": 3.5792888280690383, + "tokens_seen": 1343291392 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994884653961886, + "loss": 1.3026, + "theoretical_loss": 3.5792732292683835, + "tokens_seen": 1343356928 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029947843530591777, + "loss": 1.3491, + "theoretical_loss": 3.5792576314417666, + "tokens_seen": 1343422464 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 490931, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5292123556137085, + "objective/train/theoretical_loss": 3.5792420345890807, + "objective/train/tokens_used": 1363948000, + "theoretical_loss": 3.5792420345890807, + "tokens_seen": 1343488000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029946840521564695, + "loss": 1.4192, + "theoretical_loss": 3.5792420345890807, + "tokens_seen": 1343488000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029945837512537613, + "loss": 1.337, + "theoretical_loss": 3.579226438710216, + "tokens_seen": 1343553536 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994483450351053, + "loss": 1.3697, + "theoretical_loss": 3.579210843805065, + "tokens_seen": 1343619072 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029943831494483455, + "loss": 1.3229, + "theoretical_loss": 3.5791952498735196, + "tokens_seen": 1343684608 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994282848545637, + "loss": 1.3772, + "theoretical_loss": 3.579179656915471, + "tokens_seen": 1343750144 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994182547642929, + "loss": 1.3708, + "theoretical_loss": 3.579164064930811, + "tokens_seen": 1343815680 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002994082246740221, + "loss": 1.3146, + "theoretical_loss": 3.5791484739194317, + "tokens_seen": 1343881216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002993981945837513, + "loss": 1.3951, + "theoretical_loss": 3.5791328838812246, + "tokens_seen": 1343946752 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029938816449348046, + "loss": 1.3146, + "theoretical_loss": 3.579117294816082, + "tokens_seen": 1344012288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029937813440320964, + "loss": 1.4021, + "theoretical_loss": 3.579101706723895, + "tokens_seen": 1344077824 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002993681043129388, + "loss": 1.3688, + "theoretical_loss": 3.579086119604556, + "tokens_seen": 1344143360 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029935807422266805, + "loss": 1.3532, + "theoretical_loss": 3.5790705334579567, + "tokens_seen": 1344208896 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002993480441323972, + "loss": 1.3507, + "theoretical_loss": 3.5790549482839884, + "tokens_seen": 1344274432 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002993380140421264, + "loss": 1.3917, + "theoretical_loss": 3.579039364082544, + "tokens_seen": 1344339968 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029932798395185554, + "loss": 1.3631, + "theoretical_loss": 3.5790237808535146, + "tokens_seen": 1344405504 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002993179538615848, + "loss": 1.3512, + "theoretical_loss": 3.5790081985967928, + "tokens_seen": 1344471040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029930792377131396, + "loss": 1.3406, + "theoretical_loss": 3.57899261731227, + "tokens_seen": 1344536576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029929789368104314, + "loss": 1.373, + "theoretical_loss": 3.5789770369998384, + "tokens_seen": 1344602112 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992878635907723, + "loss": 1.3582, + "theoretical_loss": 3.57896145765939, + "tokens_seen": 1344667648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992778335005015, + "loss": 1.411, + "theoretical_loss": 3.5789458792908166, + "tokens_seen": 1344733184 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992678034102307, + "loss": 1.3113, + "theoretical_loss": 3.5789303018940104, + "tokens_seen": 1344798720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992577733199599, + "loss": 1.4317, + "theoretical_loss": 3.5789147254688634, + "tokens_seen": 1344864256 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029924774322968905, + "loss": 1.3683, + "theoretical_loss": 3.5788991500152676, + "tokens_seen": 1344929792 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992377131394183, + "loss": 1.3854, + "theoretical_loss": 3.5788835755331148, + "tokens_seen": 1344995328 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029922768304914746, + "loss": 1.2681, + "theoretical_loss": 3.578868002022298, + "tokens_seen": 1345060864 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029921765295887664, + "loss": 1.3126, + "theoretical_loss": 3.578852429482708, + "tokens_seen": 1345126400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002992076228686058, + "loss": 1.5062, + "theoretical_loss": 3.578836857914238, + "tokens_seen": 1345191936 + }, + { + "epoch": 0.41, + "learning_rate": 0.000299197592778335, + "loss": 1.4318, + "theoretical_loss": 3.5788212873167793, + "tokens_seen": 1345257472 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991875626880642, + "loss": 1.4702, + "theoretical_loss": 3.5788057176902246, + "tokens_seen": 1345323008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991775325977934, + "loss": 1.4041, + "theoretical_loss": 3.5787901490344662, + "tokens_seen": 1345388544 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029916750250752255, + "loss": 1.2624, + "theoretical_loss": 3.5787745813493954, + "tokens_seen": 1345454080 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029916750250752255, + "loss": 1.3724, + "theoretical_loss": 3.5787590146349055, + "tokens_seen": 1345519616 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991574724172518, + "loss": 1.2727, + "theoretical_loss": 3.5787434488908882, + "tokens_seen": 1345585152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991474423269809, + "loss": 1.4062, + "theoretical_loss": 3.578727884117236, + "tokens_seen": 1345650688 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029913741223671015, + "loss": 1.4037, + "theoretical_loss": 3.5787123203138402, + "tokens_seen": 1345716224 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029912738214643933, + "loss": 1.407, + "theoretical_loss": 3.5786967574805946, + "tokens_seen": 1345781760 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991173520561685, + "loss": 1.309, + "theoretical_loss": 3.57868119561739, + "tokens_seen": 1345847296 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002991073219658977, + "loss": 1.3204, + "theoretical_loss": 3.5786656347241195, + "tokens_seen": 1345912832 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029909729187562687, + "loss": 1.4531, + "theoretical_loss": 3.5786500748006755, + "tokens_seen": 1345978368 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029908726178535605, + "loss": 1.3926, + "theoretical_loss": 3.57863451584695, + "tokens_seen": 1346043904 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002990772316950853, + "loss": 1.4569, + "theoretical_loss": 3.578618957862836, + "tokens_seen": 1346109440 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002990672016048144, + "loss": 1.2732, + "theoretical_loss": 3.5786034008482246, + "tokens_seen": 1346174976 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029905717151454365, + "loss": 1.4287, + "theoretical_loss": 3.5785878448030095, + "tokens_seen": 1346240512 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029904714142427283, + "loss": 1.3526, + "theoretical_loss": 3.578572289727082, + "tokens_seen": 1346306048 + }, + { + "epoch": 0.41, + "learning_rate": 0.000299037111334002, + "loss": 1.378, + "theoretical_loss": 3.5785567356203356, + "tokens_seen": 1346371584 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002990270812437312, + "loss": 1.4235, + "theoretical_loss": 3.5785411824826623, + "tokens_seen": 1346437120 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002990170511534604, + "loss": 1.3621, + "theoretical_loss": 3.578525630313954, + "tokens_seen": 1346502656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029900702106318956, + "loss": 1.2925, + "theoretical_loss": 3.5785100791141042, + "tokens_seen": 1346568192 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989969909729188, + "loss": 1.3492, + "theoretical_loss": 3.578494528883004, + "tokens_seen": 1346633728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989869608826479, + "loss": 1.3352, + "theoretical_loss": 3.578478979620548, + "tokens_seen": 1346699264 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 492127, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3405084609985352, + "objective/train/theoretical_loss": 3.5784634313266266, + "objective/train/tokens_used": 1367224800, + "theoretical_loss": 3.5784634313266266, + "tokens_seen": 1346764800 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029897693079237715, + "loss": 1.3115, + "theoretical_loss": 3.5784634313266266, + "tokens_seen": 1346764800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989669007021063, + "loss": 1.3874, + "theoretical_loss": 3.578447884001134, + "tokens_seen": 1346830336 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989568706118355, + "loss": 1.2173, + "theoretical_loss": 3.578432337643961, + "tokens_seen": 1346895872 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989468405215647, + "loss": 1.425, + "theoretical_loss": 3.578416792255002, + "tokens_seen": 1346961408 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989368104312939, + "loss": 1.3898, + "theoretical_loss": 3.5784012478341483, + "tokens_seen": 1347026944 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029892678034102306, + "loss": 1.3133, + "theoretical_loss": 3.5783857043812937, + "tokens_seen": 1347092480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989167502507523, + "loss": 1.402, + "theoretical_loss": 3.5783701618963297, + "tokens_seen": 1347158016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002989067201604814, + "loss": 1.2632, + "theoretical_loss": 3.578354620379149, + "tokens_seen": 1347223552 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029889669007021066, + "loss": 1.401, + "theoretical_loss": 3.578339079829645, + "tokens_seen": 1347289088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002988866599799398, + "loss": 1.4805, + "theoretical_loss": 3.5783235402477103, + "tokens_seen": 1347354624 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298876629889669, + "loss": 1.241, + "theoretical_loss": 3.578308001633237, + "tokens_seen": 1347420160 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002988665997993982, + "loss": 1.3692, + "theoretical_loss": 3.578292463986118, + "tokens_seen": 1347485696 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002988565697091274, + "loss": 1.3552, + "theoretical_loss": 3.5782769273062467, + "tokens_seen": 1347551232 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029884653961885656, + "loss": 1.2953, + "theoretical_loss": 3.578261391593515, + "tokens_seen": 1347616768 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029883650952858574, + "loss": 1.3388, + "theoretical_loss": 3.5782458568478157, + "tokens_seen": 1347682304 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002988264794383149, + "loss": 1.3873, + "theoretical_loss": 3.578230323069042, + "tokens_seen": 1347747840 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029881644934804416, + "loss": 1.3853, + "theoretical_loss": 3.5782147902570873, + "tokens_seen": 1347813376 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002988064192577733, + "loss": 1.3016, + "theoretical_loss": 3.578199258411843, + "tokens_seen": 1347878912 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987963891675025, + "loss": 1.4024, + "theoretical_loss": 3.5781837275332027, + "tokens_seen": 1347944448 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987863590772317, + "loss": 1.3438, + "theoretical_loss": 3.5781681976210598, + "tokens_seen": 1348009984 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987763289869609, + "loss": 1.4646, + "theoretical_loss": 3.578152668675306, + "tokens_seen": 1348075520 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987662988966901, + "loss": 1.4602, + "theoretical_loss": 3.578137140695835, + "tokens_seen": 1348141056 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029875626880641925, + "loss": 1.3369, + "theoretical_loss": 3.578121613682539, + "tokens_seen": 1348206592 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987462387161485, + "loss": 1.3597, + "theoretical_loss": 3.578106087635312, + "tokens_seen": 1348272128 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029873620862587766, + "loss": 1.3473, + "theoretical_loss": 3.578090562554046, + "tokens_seen": 1348337664 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029872617853560684, + "loss": 1.3433, + "theoretical_loss": 3.578075038438634, + "tokens_seen": 1348403200 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298716148445336, + "loss": 1.4219, + "theoretical_loss": 3.5780595152889694, + "tokens_seen": 1348468736 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002987061183550652, + "loss": 1.3489, + "theoretical_loss": 3.5780439931049455, + "tokens_seen": 1348534272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002986960882647944, + "loss": 1.4589, + "theoretical_loss": 3.5780284718864546, + "tokens_seen": 1348599808 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002986860581745236, + "loss": 1.4792, + "theoretical_loss": 3.57801295163339, + "tokens_seen": 1348665344 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029867602808425275, + "loss": 1.3712, + "theoretical_loss": 3.5779974323456445, + "tokens_seen": 1348730880 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298665997993982, + "loss": 1.4583, + "theoretical_loss": 3.5779819140231117, + "tokens_seen": 1348796416 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002986559679037111, + "loss": 1.311, + "theoretical_loss": 3.577966396665684, + "tokens_seen": 1348861952 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029864593781344035, + "loss": 1.3623, + "theoretical_loss": 3.5779508802732547, + "tokens_seen": 1348927488 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029863590772316953, + "loss": 1.3917, + "theoretical_loss": 3.5779353648457173, + "tokens_seen": 1348993024 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002986258776328987, + "loss": 1.427, + "theoretical_loss": 3.577919850382965, + "tokens_seen": 1349058560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002986158475426279, + "loss": 1.393, + "theoretical_loss": 3.57790433688489, + "tokens_seen": 1349124096 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029860581745235707, + "loss": 1.3874, + "theoretical_loss": 3.577888824351387, + "tokens_seen": 1349189632 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029859578736208625, + "loss": 1.3935, + "theoretical_loss": 3.577873312782347, + "tokens_seen": 1349255168 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002985857572718155, + "loss": 1.3196, + "theoretical_loss": 3.5778578021776655, + "tokens_seen": 1349320704 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002985757271815446, + "loss": 1.2923, + "theoretical_loss": 3.577842292537234, + "tokens_seen": 1349386240 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029856569709127385, + "loss": 1.3276, + "theoretical_loss": 3.5778267838609468, + "tokens_seen": 1349451776 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029855566700100303, + "loss": 1.37, + "theoretical_loss": 3.5778112761486964, + "tokens_seen": 1349517312 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002985456369107322, + "loss": 1.4213, + "theoretical_loss": 3.577795769400377, + "tokens_seen": 1349582848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002985356068204614, + "loss": 1.3777, + "theoretical_loss": 3.577780263615881, + "tokens_seen": 1349648384 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002985255767301906, + "loss": 1.3635, + "theoretical_loss": 3.5777647587951016, + "tokens_seen": 1349713920 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029851554663991976, + "loss": 1.4077, + "theoretical_loss": 3.577749254937933, + "tokens_seen": 1349779456 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298505516549649, + "loss": 1.4732, + "theoretical_loss": 3.577733752044267, + "tokens_seen": 1349844992 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984954864593781, + "loss": 1.4207, + "theoretical_loss": 3.577718250113999, + "tokens_seen": 1349910528 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029848545636910735, + "loss": 1.392, + "theoretical_loss": 3.5777027491470212, + "tokens_seen": 1349976064 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 493399, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.433192253112793, + "objective/train/theoretical_loss": 3.577687249143227, + "objective/train/tokens_used": 1370501600, + "theoretical_loss": 3.577687249143227, + "tokens_seen": 1350041600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984754262788365, + "loss": 1.3274, + "theoretical_loss": 3.577687249143227, + "tokens_seen": 1350041600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984653961885657, + "loss": 1.3507, + "theoretical_loss": 3.57767175010251, + "tokens_seen": 1350107136 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984553660982949, + "loss": 1.4031, + "theoretical_loss": 3.577656252024763, + "tokens_seen": 1350172672 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984453360080241, + "loss": 1.4611, + "theoretical_loss": 3.5776407549098805, + "tokens_seen": 1350238208 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029843530591775326, + "loss": 1.3739, + "theoretical_loss": 3.577625258757755, + "tokens_seen": 1350303744 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984252758274825, + "loss": 1.3664, + "theoretical_loss": 3.5776097635682813, + "tokens_seen": 1350369280 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002984152457372116, + "loss": 1.3907, + "theoretical_loss": 3.577594269341351, + "tokens_seen": 1350434816 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029840521564694086, + "loss": 1.3707, + "theoretical_loss": 3.5775787760768587, + "tokens_seen": 1350500352 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029839518555667, + "loss": 1.374, + "theoretical_loss": 3.577563283774698, + "tokens_seen": 1350565888 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983851554663992, + "loss": 1.4349, + "theoretical_loss": 3.5775477924347623, + "tokens_seen": 1350631424 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983751253761284, + "loss": 1.3141, + "theoretical_loss": 3.5775323020569445, + "tokens_seen": 1350696960 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983650952858576, + "loss": 1.2407, + "theoretical_loss": 3.5775168126411394, + "tokens_seen": 1350762496 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029835506519558676, + "loss": 1.4449, + "theoretical_loss": 3.5775013241872395, + "tokens_seen": 1350828032 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029834503510531594, + "loss": 1.4012, + "theoretical_loss": 3.577485836695139, + "tokens_seen": 1350893568 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983350050150451, + "loss": 1.3757, + "theoretical_loss": 3.5774703501647314, + "tokens_seen": 1350959104 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029832497492477436, + "loss": 1.3978, + "theoretical_loss": 3.5774548645959103, + "tokens_seen": 1351024640 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983149448345035, + "loss": 1.3135, + "theoretical_loss": 3.5774393799885695, + "tokens_seen": 1351090176 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002983049147442327, + "loss": 1.3456, + "theoretical_loss": 3.577423896342602, + "tokens_seen": 1351155712 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029829488465396185, + "loss": 1.452, + "theoretical_loss": 3.577408413657902, + "tokens_seen": 1351221248 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002982848545636911, + "loss": 1.4677, + "theoretical_loss": 3.577392931934363, + "tokens_seen": 1351286784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029827482447342026, + "loss": 1.461, + "theoretical_loss": 3.5773774511718797, + "tokens_seen": 1351352320 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029826479438314945, + "loss": 1.506, + "theoretical_loss": 3.577361971370345, + "tokens_seen": 1351417856 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002982547642928786, + "loss": 1.332, + "theoretical_loss": 3.577346492529652, + "tokens_seen": 1351483392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029824473420260786, + "loss": 1.3921, + "theoretical_loss": 3.5773310146496953, + "tokens_seen": 1351548928 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298234704112337, + "loss": 1.3258, + "theoretical_loss": 3.5773155377303687, + "tokens_seen": 1351614464 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002982246740220662, + "loss": 1.3997, + "theoretical_loss": 3.577300061771566, + "tokens_seen": 1351680000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029821464393179535, + "loss": 1.3803, + "theoretical_loss": 3.577284586773181, + "tokens_seen": 1351745536 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002982046138415246, + "loss": 1.4606, + "theoretical_loss": 3.577269112735107, + "tokens_seen": 1351811072 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029819458375125377, + "loss": 1.3815, + "theoretical_loss": 3.577253639657238, + "tokens_seen": 1351876608 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029818455366098295, + "loss": 1.4233, + "theoretical_loss": 3.577238167539469, + "tokens_seen": 1351942144 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029817452357071213, + "loss": 1.54, + "theoretical_loss": 3.5772226963816927, + "tokens_seen": 1352007680 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002981644934804413, + "loss": 1.4597, + "theoretical_loss": 3.577207226183803, + "tokens_seen": 1352073216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002981544633901705, + "loss": 1.4771, + "theoretical_loss": 3.5771917569456946, + "tokens_seen": 1352138752 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029814443329989973, + "loss": 1.3602, + "theoretical_loss": 3.5771762886672605, + "tokens_seen": 1352204288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029813440320962885, + "loss": 1.4009, + "theoretical_loss": 3.5771608213483956, + "tokens_seen": 1352269824 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002981243731193581, + "loss": 1.3121, + "theoretical_loss": 3.5771453549889936, + "tokens_seen": 1352335360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002981143430290872, + "loss": 1.3455, + "theoretical_loss": 3.577129889588948, + "tokens_seen": 1352400896 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029810431293881645, + "loss": 1.3524, + "theoretical_loss": 3.577114425148153, + "tokens_seen": 1352466432 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029809428284854563, + "loss": 1.3453, + "theoretical_loss": 3.577098961666503, + "tokens_seen": 1352531968 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002980842527582748, + "loss": 1.2807, + "theoretical_loss": 3.577083499143892, + "tokens_seen": 1352597504 + }, + { + "epoch": 0.41, + "learning_rate": 0.000298074222668004, + "loss": 1.3141, + "theoretical_loss": 3.5770680375802137, + "tokens_seen": 1352663040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029806419257773323, + "loss": 1.371, + "theoretical_loss": 3.577052576975362, + "tokens_seen": 1352728576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029805416248746236, + "loss": 1.2941, + "theoretical_loss": 3.5770371173292315, + "tokens_seen": 1352794112 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002980441323971916, + "loss": 1.3224, + "theoretical_loss": 3.5770216586417165, + "tokens_seen": 1352859648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002980341023069208, + "loss": 1.3278, + "theoretical_loss": 3.5770062009127104, + "tokens_seen": 1352925184 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029802407221664996, + "loss": 1.4042, + "theoretical_loss": 3.576990744142108, + "tokens_seen": 1352990720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002980140421263792, + "loss": 1.3768, + "theoretical_loss": 3.5769752883298027, + "tokens_seen": 1353056256 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002980040120361083, + "loss": 1.3913, + "theoretical_loss": 3.576959833475689, + "tokens_seen": 1353121792 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029799398194583755, + "loss": 1.3124, + "theoretical_loss": 3.576944379579662, + "tokens_seen": 1353187328 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979839518555667, + "loss": 1.3518, + "theoretical_loss": 3.5769289266416147, + "tokens_seen": 1353252864 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 494475, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3703036308288574, + "objective/train/theoretical_loss": 3.5769134746614415, + "objective/train/tokens_used": 1373778400, + "theoretical_loss": 3.5769134746614415, + "tokens_seen": 1353318400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979739217652959, + "loss": 1.3747, + "theoretical_loss": 3.5769134746614415, + "tokens_seen": 1353318400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979638916750251, + "loss": 1.3748, + "theoretical_loss": 3.5768980236390373, + "tokens_seen": 1353383936 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979538615847543, + "loss": 1.4474, + "theoretical_loss": 3.576882573574296, + "tokens_seen": 1353449472 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029794383149448346, + "loss": 1.3929, + "theoretical_loss": 3.5768671244671113, + "tokens_seen": 1353515008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979338014042127, + "loss": 1.3873, + "theoretical_loss": 3.5768516763173785, + "tokens_seen": 1353580544 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979237713139418, + "loss": 1.4147, + "theoretical_loss": 3.5768362291249916, + "tokens_seen": 1353646080 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029791374122367106, + "loss": 1.3809, + "theoretical_loss": 3.576820782889844, + "tokens_seen": 1353711616 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002979037111334002, + "loss": 1.3875, + "theoretical_loss": 3.5768053376118316, + "tokens_seen": 1353777152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978936810431294, + "loss": 1.4087, + "theoretical_loss": 3.576789893290848, + "tokens_seen": 1353842688 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978836509528586, + "loss": 1.5326, + "theoretical_loss": 3.576774449926787, + "tokens_seen": 1353908224 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978736208625878, + "loss": 1.4735, + "theoretical_loss": 3.576759007519544, + "tokens_seen": 1353973760 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029786359077231696, + "loss": 1.3683, + "theoretical_loss": 3.5767435660690126, + "tokens_seen": 1354039296 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029785356068204614, + "loss": 1.4071, + "theoretical_loss": 3.5767281255750882, + "tokens_seen": 1354104832 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978435305917753, + "loss": 1.2619, + "theoretical_loss": 3.5767126860376637, + "tokens_seen": 1354170368 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029783350050150456, + "loss": 1.4226, + "theoretical_loss": 3.576697247456635, + "tokens_seen": 1354235904 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978234704112337, + "loss": 1.4409, + "theoretical_loss": 3.576681809831896, + "tokens_seen": 1354301440 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002978134403209629, + "loss": 1.3711, + "theoretical_loss": 3.5766663731633415, + "tokens_seen": 1354366976 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029780341023069205, + "loss": 1.3056, + "theoretical_loss": 3.5766509374508653, + "tokens_seen": 1354432512 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002977933801404213, + "loss": 1.4536, + "theoretical_loss": 3.5766355026943626, + "tokens_seen": 1354498048 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029778335005015046, + "loss": 1.4711, + "theoretical_loss": 3.576620068893728, + "tokens_seen": 1354563584 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029777331995987965, + "loss": 1.3676, + "theoretical_loss": 3.5766046360488555, + "tokens_seen": 1354629120 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029776328986960883, + "loss": 1.3553, + "theoretical_loss": 3.57658920415964, + "tokens_seen": 1354694656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029775325977933806, + "loss": 1.3313, + "theoretical_loss": 3.5765737732259764, + "tokens_seen": 1354760192 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002977432296890672, + "loss": 1.3681, + "theoretical_loss": 3.5765583432477586, + "tokens_seen": 1354825728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002977331995987964, + "loss": 1.4119, + "theoretical_loss": 3.5765429142248815, + "tokens_seen": 1354891264 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029772316950852555, + "loss": 1.4123, + "theoretical_loss": 3.5765274861572403, + "tokens_seen": 1354956800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002977131394182548, + "loss": 1.3311, + "theoretical_loss": 3.5765120590447292, + "tokens_seen": 1355022336 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029770310932798397, + "loss": 1.3827, + "theoretical_loss": 3.5764966328872427, + "tokens_seen": 1355087872 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029769307923771315, + "loss": 1.3212, + "theoretical_loss": 3.576481207684676, + "tokens_seen": 1355153408 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029768304914744233, + "loss": 1.4667, + "theoretical_loss": 3.576465783436923, + "tokens_seen": 1355218944 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002976730190571715, + "loss": 1.3785, + "theoretical_loss": 3.576450360143879, + "tokens_seen": 1355284480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002976629889669007, + "loss": 1.4059, + "theoretical_loss": 3.576434937805439, + "tokens_seen": 1355350016 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029765295887662993, + "loss": 1.3946, + "theoretical_loss": 3.576419516421497, + "tokens_seen": 1355415552 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029764292878635905, + "loss": 1.304, + "theoretical_loss": 3.5764040959919488, + "tokens_seen": 1355481088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002976328986960883, + "loss": 1.3226, + "theoretical_loss": 3.5763886765166886, + "tokens_seen": 1355546624 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002976228686058174, + "loss": 1.4716, + "theoretical_loss": 3.576373257995611, + "tokens_seen": 1355612160 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029761283851554665, + "loss": 1.4255, + "theoretical_loss": 3.576357840428611, + "tokens_seen": 1355677696 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029760280842527583, + "loss": 1.3745, + "theoretical_loss": 3.5763424238155834, + "tokens_seen": 1355743232 + }, + { + "epoch": 0.41, + "learning_rate": 0.000297592778335005, + "loss": 1.4182, + "theoretical_loss": 3.576327008156423, + "tokens_seen": 1355808768 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002975827482447342, + "loss": 1.4458, + "theoretical_loss": 3.5763115934510257, + "tokens_seen": 1355874304 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029757271815446343, + "loss": 1.3419, + "theoretical_loss": 3.576296179699285, + "tokens_seen": 1355939840 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029756268806419256, + "loss": 1.3955, + "theoretical_loss": 3.576280766901096, + "tokens_seen": 1356005376 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002975526579739218, + "loss": 1.2615, + "theoretical_loss": 3.5762653550563543, + "tokens_seen": 1356070912 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002975426278836509, + "loss": 1.2604, + "theoretical_loss": 3.576249944164955, + "tokens_seen": 1356136448 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029753259779338016, + "loss": 1.38, + "theoretical_loss": 3.576234534226792, + "tokens_seen": 1356201984 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029752256770310934, + "loss": 1.4028, + "theoretical_loss": 3.576219125241761, + "tokens_seen": 1356267520 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002975125376128385, + "loss": 1.3379, + "theoretical_loss": 3.576203717209757, + "tokens_seen": 1356333056 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002975025075225677, + "loss": 1.2543, + "theoretical_loss": 3.576188310130674, + "tokens_seen": 1356398592 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974924774322969, + "loss": 1.4076, + "theoretical_loss": 3.576172904004409, + "tokens_seen": 1356464128 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029748244734202606, + "loss": 1.3895, + "theoretical_loss": 3.576157498830856, + "tokens_seen": 1356529664 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 495730, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5578267574310303, + "objective/train/theoretical_loss": 3.576142094609909, + "objective/train/tokens_used": 1377055200, + "theoretical_loss": 3.576142094609909, + "tokens_seen": 1356595200 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974724172517553, + "loss": 1.4117, + "theoretical_loss": 3.576142094609909, + "tokens_seen": 1356595200 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974623871614844, + "loss": 1.4065, + "theoretical_loss": 3.5761266913414653, + "tokens_seen": 1356660736 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029745235707121366, + "loss": 1.4531, + "theoretical_loss": 3.576111289025418, + "tokens_seen": 1356726272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974423269809428, + "loss": 1.4974, + "theoretical_loss": 3.576095887661663, + "tokens_seen": 1356791808 + }, + { + "epoch": 0.41, + "learning_rate": 0.000297432296890672, + "loss": 1.3319, + "theoretical_loss": 3.5760804872500955, + "tokens_seen": 1356857344 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974222668004012, + "loss": 1.3996, + "theoretical_loss": 3.576065087790611, + "tokens_seen": 1356922880 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002974122367101304, + "loss": 1.4906, + "theoretical_loss": 3.5760496892831037, + "tokens_seen": 1356988416 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029740220661985956, + "loss": 1.3977, + "theoretical_loss": 3.5760342917274697, + "tokens_seen": 1357053952 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973921765295888, + "loss": 1.3962, + "theoretical_loss": 3.576018895123603, + "tokens_seen": 1357119488 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973821464393179, + "loss": 1.3541, + "theoretical_loss": 3.5760034994714003, + "tokens_seen": 1357185024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029737211634904716, + "loss": 1.3186, + "theoretical_loss": 3.5759881047707562, + "tokens_seen": 1357250560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973620862587763, + "loss": 1.4335, + "theoretical_loss": 3.5759727110215658, + "tokens_seen": 1357316096 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973520561685055, + "loss": 1.4332, + "theoretical_loss": 3.5759573182237245, + "tokens_seen": 1357381632 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973420260782347, + "loss": 1.2428, + "theoretical_loss": 3.5759419263771273, + "tokens_seen": 1357447168 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002973319959879639, + "loss": 1.4111, + "theoretical_loss": 3.57592653548167, + "tokens_seen": 1357512704 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029732196589769307, + "loss": 1.2381, + "theoretical_loss": 3.575911145537247, + "tokens_seen": 1357578240 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029731193580742225, + "loss": 1.4128, + "theoretical_loss": 3.5758957565437552, + "tokens_seen": 1357643776 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029730190571715143, + "loss": 1.4474, + "theoretical_loss": 3.5758803685010885, + "tokens_seen": 1357709312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029729187562688067, + "loss": 1.3043, + "theoretical_loss": 3.5758649814091426, + "tokens_seen": 1357774848 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029728184553660985, + "loss": 1.3708, + "theoretical_loss": 3.5758495952678135, + "tokens_seen": 1357840384 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029727181544633903, + "loss": 1.42, + "theoretical_loss": 3.5758342100769958, + "tokens_seen": 1357905920 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029726178535606826, + "loss": 1.3394, + "theoretical_loss": 3.5758188258365853, + "tokens_seen": 1357971456 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002972517552657974, + "loss": 1.3639, + "theoretical_loss": 3.5758034425464773, + "tokens_seen": 1358036992 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002972417251755266, + "loss": 1.4868, + "theoretical_loss": 3.5757880602065675, + "tokens_seen": 1358102528 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029723169508525575, + "loss": 1.419, + "theoretical_loss": 3.575772678816751, + "tokens_seen": 1358168064 + }, + { + "epoch": 0.41, + "learning_rate": 0.000297221664994985, + "loss": 1.4603, + "theoretical_loss": 3.5757572983769235, + "tokens_seen": 1358233600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029721163490471417, + "loss": 1.2949, + "theoretical_loss": 3.5757419188869806, + "tokens_seen": 1358299136 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029720160481444335, + "loss": 1.3429, + "theoretical_loss": 3.5757265403468175, + "tokens_seen": 1358364672 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029719157472417253, + "loss": 1.4442, + "theoretical_loss": 3.5757111627563294, + "tokens_seen": 1358430208 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002971815446339017, + "loss": 1.3501, + "theoretical_loss": 3.575695786115413, + "tokens_seen": 1358495744 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002971715145436309, + "loss": 1.3255, + "theoretical_loss": 3.5756804104239626, + "tokens_seen": 1358561280 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029716148445336013, + "loss": 1.435, + "theoretical_loss": 3.575665035681875, + "tokens_seen": 1358626816 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029715145436308926, + "loss": 1.3148, + "theoretical_loss": 3.575649661889045, + "tokens_seen": 1358692352 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002971414242728185, + "loss": 1.4074, + "theoretical_loss": 3.575634289045368, + "tokens_seen": 1358757888 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002971313941825476, + "loss": 1.4702, + "theoretical_loss": 3.5756189171507406, + "tokens_seen": 1358823424 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029712136409227685, + "loss": 1.3317, + "theoretical_loss": 3.5756035462050573, + "tokens_seen": 1358888960 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029711133400200603, + "loss": 1.3743, + "theoretical_loss": 3.575588176208214, + "tokens_seen": 1358954496 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002971013039117352, + "loss": 1.4101, + "theoretical_loss": 3.5755728071601074, + "tokens_seen": 1359020032 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002970912738214644, + "loss": 1.4229, + "theoretical_loss": 3.575557439060632, + "tokens_seen": 1359085568 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029708124373119363, + "loss": 1.3889, + "theoretical_loss": 3.575542071909684, + "tokens_seen": 1359151104 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029707121364092276, + "loss": 1.3828, + "theoretical_loss": 3.5755267057071594, + "tokens_seen": 1359216640 + }, + { + "epoch": 0.41, + "learning_rate": 0.000297061183550652, + "loss": 1.3577, + "theoretical_loss": 3.5755113404529535, + "tokens_seen": 1359282176 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002970511534603811, + "loss": 1.2518, + "theoretical_loss": 3.5754959761469616, + "tokens_seen": 1359347712 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029704112337011036, + "loss": 1.4278, + "theoretical_loss": 3.5754806127890806, + "tokens_seen": 1359413248 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029703109327983954, + "loss": 1.3828, + "theoretical_loss": 3.5754652503792057, + "tokens_seen": 1359478784 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002970210631895687, + "loss": 1.3908, + "theoretical_loss": 3.575449888917233, + "tokens_seen": 1359544320 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002970110330992979, + "loss": 1.2495, + "theoretical_loss": 3.5754345284030573, + "tokens_seen": 1359609856 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002970010030090271, + "loss": 1.4357, + "theoretical_loss": 3.575419168836576, + "tokens_seen": 1359675392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029699097291875626, + "loss": 1.3503, + "theoretical_loss": 3.5754038102176837, + "tokens_seen": 1359740928 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002969809428284855, + "loss": 1.4077, + "theoretical_loss": 3.575388452546277, + "tokens_seen": 1359806464 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 496861, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3311724662780762, + "objective/train/theoretical_loss": 3.5753730958222514, + "objective/train/tokens_used": 1380332000, + "theoretical_loss": 3.5753730958222514, + "tokens_seen": 1359872000 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002969709127382146, + "loss": 1.2331, + "theoretical_loss": 3.5753730958222514, + "tokens_seen": 1359872000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029696088264794386, + "loss": 1.4517, + "theoretical_loss": 3.5753577400455026, + "tokens_seen": 1359937536 + }, + { + "epoch": 0.41, + "learning_rate": 0.000296950852557673, + "loss": 1.3521, + "theoretical_loss": 3.575342385215927, + "tokens_seen": 1360003072 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002969408224674022, + "loss": 1.4218, + "theoretical_loss": 3.575327031333421, + "tokens_seen": 1360068608 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002969307923771314, + "loss": 1.3978, + "theoretical_loss": 3.5753116783978793, + "tokens_seen": 1360134144 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002969207622868606, + "loss": 1.331, + "theoretical_loss": 3.5752963264091986, + "tokens_seen": 1360199680 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029691073219658976, + "loss": 1.3955, + "theoretical_loss": 3.575280975367275, + "tokens_seen": 1360265216 + }, + { + "epoch": 0.41, + "learning_rate": 0.000296900702106319, + "loss": 1.431, + "theoretical_loss": 3.5752656252720048, + "tokens_seen": 1360330752 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002968906720160481, + "loss": 1.5343, + "theoretical_loss": 3.5752502761232825, + "tokens_seen": 1360396288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029688064192577736, + "loss": 1.3448, + "theoretical_loss": 3.575234927921006, + "tokens_seen": 1360461824 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002968706118355065, + "loss": 1.4551, + "theoretical_loss": 3.57521958066507, + "tokens_seen": 1360527360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002968605817452357, + "loss": 1.3671, + "theoretical_loss": 3.575204234355371, + "tokens_seen": 1360592896 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002968505516549649, + "loss": 1.3674, + "theoretical_loss": 3.5751888889918053, + "tokens_seen": 1360658432 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002968405215646941, + "loss": 1.3825, + "theoretical_loss": 3.575173544574269, + "tokens_seen": 1360723968 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029683049147442327, + "loss": 1.3393, + "theoretical_loss": 3.575158201102658, + "tokens_seen": 1360789504 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029682046138415245, + "loss": 1.3184, + "theoretical_loss": 3.5751428585768688, + "tokens_seen": 1360855040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029681043129388163, + "loss": 1.3676, + "theoretical_loss": 3.5751275169967975, + "tokens_seen": 1360920576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029680040120361087, + "loss": 1.4734, + "theoretical_loss": 3.5751121763623397, + "tokens_seen": 1360986112 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029679037111334, + "loss": 1.3065, + "theoretical_loss": 3.5750968366733917, + "tokens_seen": 1361051648 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029678034102306923, + "loss": 1.4234, + "theoretical_loss": 3.57508149792985, + "tokens_seen": 1361117184 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002967703109327984, + "loss": 1.4328, + "theoretical_loss": 3.5750661601316116, + "tokens_seen": 1361182720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002967602808425276, + "loss": 1.3193, + "theoretical_loss": 3.5750508232785707, + "tokens_seen": 1361248256 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029675025075225677, + "loss": 1.3306, + "theoretical_loss": 3.5750354873706254, + "tokens_seen": 1361313792 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029674022066198595, + "loss": 1.2841, + "theoretical_loss": 3.5750201524076712, + "tokens_seen": 1361379328 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029673019057171513, + "loss": 1.3177, + "theoretical_loss": 3.5750048183896044, + "tokens_seen": 1361444864 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029672016048144437, + "loss": 1.4552, + "theoretical_loss": 3.5749894853163218, + "tokens_seen": 1361510400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002967101303911735, + "loss": 1.271, + "theoretical_loss": 3.5749741531877186, + "tokens_seen": 1361575936 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029670010030090273, + "loss": 1.5177, + "theoretical_loss": 3.5749588220036923, + "tokens_seen": 1361641472 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029669007021063186, + "loss": 1.2912, + "theoretical_loss": 3.574943491764139, + "tokens_seen": 1361707008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002966800401203611, + "loss": 1.4166, + "theoretical_loss": 3.5749281624689546, + "tokens_seen": 1361772544 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002966700100300903, + "loss": 1.567, + "theoretical_loss": 3.5749128341180354, + "tokens_seen": 1361838080 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029665997993981946, + "loss": 1.3005, + "theoretical_loss": 3.5748975067112783, + "tokens_seen": 1361903616 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029664994984954864, + "loss": 1.4031, + "theoretical_loss": 3.5748821802485797, + "tokens_seen": 1361969152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002966399197592778, + "loss": 1.3675, + "theoretical_loss": 3.574866854729836, + "tokens_seen": 1362034688 + }, + { + "epoch": 0.41, + "learning_rate": 0.000296629889669007, + "loss": 1.3467, + "theoretical_loss": 3.5748515301549433, + "tokens_seen": 1362100224 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029661985957873623, + "loss": 1.3759, + "theoretical_loss": 3.5748362065237984, + "tokens_seen": 1362165760 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029660982948846536, + "loss": 1.5073, + "theoretical_loss": 3.5748208838362974, + "tokens_seen": 1362231296 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965997993981946, + "loss": 1.3389, + "theoretical_loss": 3.574805562092337, + "tokens_seen": 1362296832 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965897693079238, + "loss": 1.425, + "theoretical_loss": 3.574790241291814, + "tokens_seen": 1362362368 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029657973921765296, + "loss": 1.458, + "theoretical_loss": 3.574774921434625, + "tokens_seen": 1362427904 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029656970912738214, + "loss": 1.2891, + "theoretical_loss": 3.5747596025206656, + "tokens_seen": 1362493440 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965596790371113, + "loss": 1.368, + "theoretical_loss": 3.574744284549834, + "tokens_seen": 1362558976 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965496489468405, + "loss": 1.3336, + "theoretical_loss": 3.5747289675220246, + "tokens_seen": 1362624512 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029653961885656974, + "loss": 1.4895, + "theoretical_loss": 3.5747136514371363, + "tokens_seen": 1362690048 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965295887662989, + "loss": 1.506, + "theoretical_loss": 3.5746983362950635, + "tokens_seen": 1362755584 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965195586760281, + "loss": 1.3249, + "theoretical_loss": 3.5746830220957047, + "tokens_seen": 1362821120 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002965095285857573, + "loss": 1.2937, + "theoretical_loss": 3.5746677088389553, + "tokens_seen": 1362886656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029649949849548646, + "loss": 1.4751, + "theoretical_loss": 3.574652396524713, + "tokens_seen": 1362952192 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964894684052157, + "loss": 1.3411, + "theoretical_loss": 3.5746370851528733, + "tokens_seen": 1363017728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964794383149448, + "loss": 1.3906, + "theoretical_loss": 3.5746217747233335, + "tokens_seen": 1363083264 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 498147, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3767197132110596, + "objective/train/theoretical_loss": 3.5746064652359903, + "objective/train/tokens_used": 1383608800, + "theoretical_loss": 3.5746064652359903, + "tokens_seen": 1363148800 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029646940822467406, + "loss": 1.3619, + "theoretical_loss": 3.5746064652359903, + "tokens_seen": 1363148800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964593781344032, + "loss": 1.3244, + "theoretical_loss": 3.574591156690741, + "tokens_seen": 1363214336 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964493480441324, + "loss": 1.3171, + "theoretical_loss": 3.574575849087481, + "tokens_seen": 1363279872 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964393179538616, + "loss": 1.4064, + "theoretical_loss": 3.5745605424261084, + "tokens_seen": 1363345408 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964292878635908, + "loss": 1.3541, + "theoretical_loss": 3.574545236706519, + "tokens_seen": 1363410944 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029641925777331996, + "loss": 1.3868, + "theoretical_loss": 3.57452993192861, + "tokens_seen": 1363476480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002964092276830492, + "loss": 1.279, + "theoretical_loss": 3.5745146280922784, + "tokens_seen": 1363542016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002963991975927783, + "loss": 1.4878, + "theoretical_loss": 3.5744993251974204, + "tokens_seen": 1363607552 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029638916750250756, + "loss": 1.3055, + "theoretical_loss": 3.5744840232439334, + "tokens_seen": 1363673088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002963791374122367, + "loss": 1.3957, + "theoretical_loss": 3.5744687222317144, + "tokens_seen": 1363738624 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002963691073219659, + "loss": 1.4594, + "theoretical_loss": 3.5744534221606594, + "tokens_seen": 1363804160 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002963590772316951, + "loss": 1.416, + "theoretical_loss": 3.5744381230306663, + "tokens_seen": 1363869696 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002963490471414243, + "loss": 1.4207, + "theoretical_loss": 3.5744228248416317, + "tokens_seen": 1363935232 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029633901705115347, + "loss": 1.4707, + "theoretical_loss": 3.5744075275934515, + "tokens_seen": 1364000768 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029632898696088265, + "loss": 1.3214, + "theoretical_loss": 3.5743922312860246, + "tokens_seen": 1364066304 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029631895687061183, + "loss": 1.4343, + "theoretical_loss": 3.574376935919246, + "tokens_seen": 1364131840 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029630892678034107, + "loss": 1.4164, + "theoretical_loss": 3.574361641493014, + "tokens_seen": 1364197376 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002962988966900702, + "loss": 1.4447, + "theoretical_loss": 3.574346348007225, + "tokens_seen": 1364262912 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029628886659979943, + "loss": 1.3192, + "theoretical_loss": 3.574331055461776, + "tokens_seen": 1364328448 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002962788365095286, + "loss": 1.5172, + "theoretical_loss": 3.574315763856564, + "tokens_seen": 1364393984 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002962688064192578, + "loss": 1.3826, + "theoretical_loss": 3.5743004731914865, + "tokens_seen": 1364459520 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029625877632898697, + "loss": 1.4216, + "theoretical_loss": 3.57428518346644, + "tokens_seen": 1364525056 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029624874623871615, + "loss": 1.314, + "theoretical_loss": 3.574269894681322, + "tokens_seen": 1364590592 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029623871614844533, + "loss": 1.3464, + "theoretical_loss": 3.5742546068360292, + "tokens_seen": 1364656128 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029622868605817457, + "loss": 1.2874, + "theoretical_loss": 3.574239319930459, + "tokens_seen": 1364721664 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002962186559679037, + "loss": 1.4046, + "theoretical_loss": 3.5742240339645077, + "tokens_seen": 1364787200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029620862587763293, + "loss": 1.4471, + "theoretical_loss": 3.574208748938074, + "tokens_seen": 1364852736 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029619859578736206, + "loss": 1.3321, + "theoretical_loss": 3.574193464851054, + "tokens_seen": 1364918272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002961885656970913, + "loss": 1.3809, + "theoretical_loss": 3.5741781817033447, + "tokens_seen": 1364983808 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002961785356068205, + "loss": 1.3775, + "theoretical_loss": 3.5741628994948433, + "tokens_seen": 1365049344 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029616850551654966, + "loss": 1.3352, + "theoretical_loss": 3.5741476182254477, + "tokens_seen": 1365114880 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029615847542627884, + "loss": 1.5349, + "theoretical_loss": 3.5741323378950542, + "tokens_seen": 1365180416 + }, + { + "epoch": 0.41, + "learning_rate": 0.000296148445336008, + "loss": 1.3972, + "theoretical_loss": 3.5741170585035613, + "tokens_seen": 1365245952 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002961384152457372, + "loss": 1.4989, + "theoretical_loss": 3.5741017800508645, + "tokens_seen": 1365311488 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029612838515546643, + "loss": 1.4301, + "theoretical_loss": 3.574086502536863, + "tokens_seen": 1365377024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029611835506519556, + "loss": 1.5257, + "theoretical_loss": 3.574071225961452, + "tokens_seen": 1365442560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002961083249749248, + "loss": 1.2644, + "theoretical_loss": 3.574055950324531, + "tokens_seen": 1365508096 + }, + { + "epoch": 0.41, + "learning_rate": 0.000296098294884654, + "loss": 1.3589, + "theoretical_loss": 3.574040675625996, + "tokens_seen": 1365573632 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029608826479438316, + "loss": 1.3746, + "theoretical_loss": 3.574025401865744, + "tokens_seen": 1365639168 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029607823470411234, + "loss": 1.3895, + "theoretical_loss": 3.574010129043673, + "tokens_seen": 1365704704 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002960682046138415, + "loss": 1.4462, + "theoretical_loss": 3.5739948571596805, + "tokens_seen": 1365770240 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002960581745235707, + "loss": 1.4309, + "theoretical_loss": 3.573979586213663, + "tokens_seen": 1365835776 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029604814443329994, + "loss": 1.3631, + "theoretical_loss": 3.573964316205519, + "tokens_seen": 1365901312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029603811434302906, + "loss": 1.5099, + "theoretical_loss": 3.5739490471351454, + "tokens_seen": 1365966848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002960280842527583, + "loss": 1.4529, + "theoretical_loss": 3.5739337790024397, + "tokens_seen": 1366032384 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002960180541624874, + "loss": 1.5579, + "theoretical_loss": 3.5739185118072987, + "tokens_seen": 1366097920 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029600802407221666, + "loss": 1.3003, + "theoretical_loss": 3.5739032455496207, + "tokens_seen": 1366163456 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029599799398194584, + "loss": 1.3457, + "theoretical_loss": 3.573887980229303, + "tokens_seen": 1366228992 + }, + { + "epoch": 0.41, + "learning_rate": 0.000295987963891675, + "loss": 1.3699, + "theoretical_loss": 3.5738727158462424, + "tokens_seen": 1366294528 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002959779338014042, + "loss": 1.3769, + "theoretical_loss": 3.5738574524003375, + "tokens_seen": 1366360064 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 499217, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.311177134513855, + "objective/train/theoretical_loss": 3.573842189891485, + "objective/train/tokens_used": 1386885600, + "theoretical_loss": 3.573842189891485, + "tokens_seen": 1366425600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002959679037111334, + "loss": 1.3097, + "theoretical_loss": 3.573842189891485, + "tokens_seen": 1366425600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029595787362086257, + "loss": 1.3446, + "theoretical_loss": 3.573826928319583, + "tokens_seen": 1366491136 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002959478435305918, + "loss": 1.3388, + "theoretical_loss": 3.573811667684528, + "tokens_seen": 1366556672 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029593781344032093, + "loss": 1.4089, + "theoretical_loss": 3.5737964079862197, + "tokens_seen": 1366622208 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029592778335005016, + "loss": 1.3184, + "theoretical_loss": 3.573781149224553, + "tokens_seen": 1366687744 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029591775325977935, + "loss": 1.3579, + "theoretical_loss": 3.5737658913994275, + "tokens_seen": 1366753280 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029590772316950853, + "loss": 1.3943, + "theoretical_loss": 3.57375063451074, + "tokens_seen": 1366818816 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958976930792377, + "loss": 1.3842, + "theoretical_loss": 3.573735378558388, + "tokens_seen": 1366884352 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958876629889669, + "loss": 1.3832, + "theoretical_loss": 3.5737201235422695, + "tokens_seen": 1366949888 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029587763289869607, + "loss": 1.394, + "theoretical_loss": 3.573704869462282, + "tokens_seen": 1367015424 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958676028084253, + "loss": 1.4366, + "theoretical_loss": 3.5736896163183234, + "tokens_seen": 1367080960 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029585757271815443, + "loss": 1.5019, + "theoretical_loss": 3.573674364110291, + "tokens_seen": 1367146496 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029584754262788367, + "loss": 1.3041, + "theoretical_loss": 3.573659112838083, + "tokens_seen": 1367212032 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958375125376128, + "loss": 1.3628, + "theoretical_loss": 3.573643862501597, + "tokens_seen": 1367277568 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029582748244734203, + "loss": 1.4218, + "theoretical_loss": 3.5736286131007304, + "tokens_seen": 1367343104 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958174523570712, + "loss": 1.4396, + "theoretical_loss": 3.5736133646353805, + "tokens_seen": 1367408640 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002958074222668004, + "loss": 1.4104, + "theoretical_loss": 3.5735981171054467, + "tokens_seen": 1367474176 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002957973921765296, + "loss": 1.4407, + "theoretical_loss": 3.5735828705108252, + "tokens_seen": 1367539712 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002957873620862588, + "loss": 1.3542, + "theoretical_loss": 3.5735676248514148, + "tokens_seen": 1367605248 + }, + { + "epoch": 0.41, + "learning_rate": 0.000295777331995988, + "loss": 1.3845, + "theoretical_loss": 3.573552380127113, + "tokens_seen": 1367670784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029576730190571717, + "loss": 1.3137, + "theoretical_loss": 3.5735371363378174, + "tokens_seen": 1367736320 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029575727181544635, + "loss": 1.3972, + "theoretical_loss": 3.5735218934834263, + "tokens_seen": 1367801856 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029574724172517553, + "loss": 1.3634, + "theoretical_loss": 3.573506651563837, + "tokens_seen": 1367867392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029573721163490477, + "loss": 1.3751, + "theoretical_loss": 3.5734914105789484, + "tokens_seen": 1367932928 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002957271815446339, + "loss": 1.3171, + "theoretical_loss": 3.5734761705286573, + "tokens_seen": 1367998464 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029571715145436313, + "loss": 1.433, + "theoretical_loss": 3.573460931412862, + "tokens_seen": 1368064000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029570712136409226, + "loss": 1.3245, + "theoretical_loss": 3.57344569323146, + "tokens_seen": 1368129536 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002956970912738215, + "loss": 1.3588, + "theoretical_loss": 3.5734304559843504, + "tokens_seen": 1368195072 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002956870611835507, + "loss": 1.4773, + "theoretical_loss": 3.57341521967143, + "tokens_seen": 1368260608 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029567703109327986, + "loss": 1.2608, + "theoretical_loss": 3.573399984292598, + "tokens_seen": 1368326144 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029566700100300904, + "loss": 1.3987, + "theoretical_loss": 3.5733847498477513, + "tokens_seen": 1368391680 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002956569709127382, + "loss": 1.4247, + "theoretical_loss": 3.5733695163367885, + "tokens_seen": 1368457216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002956469408224674, + "loss": 1.4179, + "theoretical_loss": 3.573354283759607, + "tokens_seen": 1368522752 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029563691073219663, + "loss": 1.4366, + "theoretical_loss": 3.5733390521161055, + "tokens_seen": 1368588288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029562688064192576, + "loss": 1.4081, + "theoretical_loss": 3.573323821406182, + "tokens_seen": 1368653824 + }, + { + "epoch": 0.41, + "learning_rate": 0.000295616850551655, + "loss": 1.4372, + "theoretical_loss": 3.5733085916297345, + "tokens_seen": 1368719360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002956068204613842, + "loss": 1.4727, + "theoretical_loss": 3.5732933627866608, + "tokens_seen": 1368784896 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029559679037111336, + "loss": 1.2764, + "theoretical_loss": 3.573278134876859, + "tokens_seen": 1368850432 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029558676028084254, + "loss": 1.375, + "theoretical_loss": 3.5732629079002276, + "tokens_seen": 1368915968 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002955767301905717, + "loss": 1.413, + "theoretical_loss": 3.5732476818566647, + "tokens_seen": 1368981504 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002955667001003009, + "loss": 1.2988, + "theoretical_loss": 3.5732324567460685, + "tokens_seen": 1369047040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029555667001003014, + "loss": 1.3824, + "theoretical_loss": 3.573217232568337, + "tokens_seen": 1369112576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029554663991975926, + "loss": 1.3789, + "theoretical_loss": 3.5732020093233685, + "tokens_seen": 1369178112 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002955366098294885, + "loss": 1.3496, + "theoretical_loss": 3.573186787011061, + "tokens_seen": 1369243648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0002955265797392176, + "loss": 1.3746, + "theoretical_loss": 3.5731715656313128, + "tokens_seen": 1369309184 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029551654964894686, + "loss": 1.352, + "theoretical_loss": 3.5731563451840227, + "tokens_seen": 1369374720 + }, + { + "epoch": 0.41, + "learning_rate": 0.00029550651955867604, + "loss": 1.3744, + "theoretical_loss": 3.5731411256690877, + "tokens_seen": 1369440256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002954964894684052, + "loss": 1.3276, + "theoretical_loss": 3.573125907086407, + "tokens_seen": 1369505792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002954864593781344, + "loss": 1.3141, + "theoretical_loss": 3.5731106894358793, + "tokens_seen": 1369571328 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002954764292878636, + "loss": 1.4865, + "theoretical_loss": 3.5730954727174016, + "tokens_seen": 1369636864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 500422, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.444048523902893, + "objective/train/theoretical_loss": 3.573080256930874, + "objective/train/tokens_used": 1390162400, + "theoretical_loss": 3.573080256930874, + "tokens_seen": 1369702400 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029546639919759277, + "loss": 1.4387, + "theoretical_loss": 3.573080256930874, + "tokens_seen": 1369702400 + }, + { + "epoch": 0.42, + "learning_rate": 0.000295456369107322, + "loss": 1.4222, + "theoretical_loss": 3.573065042076193, + "tokens_seen": 1369767936 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029544633901705113, + "loss": 1.4884, + "theoretical_loss": 3.5730498281532577, + "tokens_seen": 1369833472 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029543630892678037, + "loss": 1.3268, + "theoretical_loss": 3.5730346151619665, + "tokens_seen": 1369899008 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029542627883650955, + "loss": 1.354, + "theoretical_loss": 3.5730194031022178, + "tokens_seen": 1369964544 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029541624874623873, + "loss": 1.4356, + "theoretical_loss": 3.57300419197391, + "tokens_seen": 1370030080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002954062186559679, + "loss": 1.3532, + "theoretical_loss": 3.572988981776942, + "tokens_seen": 1370095616 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002953961885656971, + "loss": 1.3419, + "theoretical_loss": 3.5729737725112116, + "tokens_seen": 1370161152 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029538615847542627, + "loss": 1.3233, + "theoretical_loss": 3.572958564176617, + "tokens_seen": 1370226688 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002953761283851555, + "loss": 1.378, + "theoretical_loss": 3.5729433567730573, + "tokens_seen": 1370292224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029536609829488463, + "loss": 1.4268, + "theoretical_loss": 3.5729281503004304, + "tokens_seen": 1370357760 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029535606820461387, + "loss": 1.3865, + "theoretical_loss": 3.5729129447586354, + "tokens_seen": 1370423296 + }, + { + "epoch": 0.42, + "learning_rate": 0.000295346038114343, + "loss": 1.3438, + "theoretical_loss": 3.5728977401475706, + "tokens_seen": 1370488832 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029533600802407223, + "loss": 1.3221, + "theoretical_loss": 3.572882536467134, + "tokens_seen": 1370554368 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002953259779338014, + "loss": 1.3706, + "theoretical_loss": 3.572867333717225, + "tokens_seen": 1370619904 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002953159478435306, + "loss": 1.4091, + "theoretical_loss": 3.5728521318977418, + "tokens_seen": 1370685440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002953059177532598, + "loss": 1.365, + "theoretical_loss": 3.572836931008583, + "tokens_seen": 1370750976 + }, + { + "epoch": 0.42, + "learning_rate": 0.000295295887662989, + "loss": 1.3843, + "theoretical_loss": 3.572821731049647, + "tokens_seen": 1370816512 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029528585757271814, + "loss": 1.3924, + "theoretical_loss": 3.5728065320208326, + "tokens_seen": 1370882048 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029527582748244737, + "loss": 1.4354, + "theoretical_loss": 3.572791333922038, + "tokens_seen": 1370947584 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002952657973921765, + "loss": 1.3275, + "theoretical_loss": 3.5727761367531627, + "tokens_seen": 1371013120 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029525576730190573, + "loss": 1.4484, + "theoretical_loss": 3.5727609405141045, + "tokens_seen": 1371078656 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002952457372116349, + "loss": 1.5861, + "theoretical_loss": 3.572745745204763, + "tokens_seen": 1371144192 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002952357071213641, + "loss": 1.4846, + "theoretical_loss": 3.5727305508250358, + "tokens_seen": 1371209728 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002952256770310933, + "loss": 1.3218, + "theoretical_loss": 3.572715357374822, + "tokens_seen": 1371275264 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029521564694082246, + "loss": 1.3718, + "theoretical_loss": 3.5727001648540204, + "tokens_seen": 1371340800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029520561685055164, + "loss": 1.3766, + "theoretical_loss": 3.57268497326253, + "tokens_seen": 1371406336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002951955867602809, + "loss": 1.4088, + "theoretical_loss": 3.5726697826002494, + "tokens_seen": 1371471872 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029518555667001, + "loss": 1.2787, + "theoretical_loss": 3.572654592867077, + "tokens_seen": 1371537408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029517552657973924, + "loss": 1.3867, + "theoretical_loss": 3.572639404062912, + "tokens_seen": 1371602944 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029516549648946836, + "loss": 1.3141, + "theoretical_loss": 3.572624216187653, + "tokens_seen": 1371668480 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002951554663991976, + "loss": 1.429, + "theoretical_loss": 3.5726090292411996, + "tokens_seen": 1371734016 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002951454363089268, + "loss": 1.2766, + "theoretical_loss": 3.5725938432234488, + "tokens_seen": 1371799552 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029513540621865596, + "loss": 1.3938, + "theoretical_loss": 3.572578658134301, + "tokens_seen": 1371865088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029512537612838514, + "loss": 1.2789, + "theoretical_loss": 3.5725634739736547, + "tokens_seen": 1371930624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002951153460381144, + "loss": 1.3818, + "theoretical_loss": 3.572548290741408, + "tokens_seen": 1371996160 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002951053159478435, + "loss": 1.3132, + "theoretical_loss": 3.5725331084374616, + "tokens_seen": 1372061696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029509528585757274, + "loss": 1.4095, + "theoretical_loss": 3.5725179270617127, + "tokens_seen": 1372127232 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029508525576730187, + "loss": 1.4833, + "theoretical_loss": 3.572502746614061, + "tokens_seen": 1372192768 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002950752256770311, + "loss": 1.4174, + "theoretical_loss": 3.572487567094405, + "tokens_seen": 1372258304 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002950651955867603, + "loss": 1.3554, + "theoretical_loss": 3.5724723885026437, + "tokens_seen": 1372323840 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029505516549648946, + "loss": 1.3073, + "theoretical_loss": 3.5724572108386767, + "tokens_seen": 1372389376 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029504513540621865, + "loss": 1.2526, + "theoretical_loss": 3.5724420341024024, + "tokens_seen": 1372454912 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002950351053159478, + "loss": 1.3364, + "theoretical_loss": 3.5724268582937198, + "tokens_seen": 1372520448 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029502507522567706, + "loss": 1.3236, + "theoretical_loss": 3.572411683412528, + "tokens_seen": 1372585984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029501504513540624, + "loss": 1.312, + "theoretical_loss": 3.5723965094587267, + "tokens_seen": 1372651520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002950050150451354, + "loss": 1.4325, + "theoretical_loss": 3.5723813364322137, + "tokens_seen": 1372717056 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002949949849548646, + "loss": 1.4121, + "theoretical_loss": 3.5723661643328892, + "tokens_seen": 1372782592 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002949849548645938, + "loss": 1.3814, + "theoretical_loss": 3.5723509931606516, + "tokens_seen": 1372848128 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029497492477432297, + "loss": 1.355, + "theoretical_loss": 3.5723358229154005, + "tokens_seen": 1372913664 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 501474, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2461267709732056, + "objective/train/theoretical_loss": 3.5723206535970347, + "objective/train/tokens_used": 1393439200, + "theoretical_loss": 3.5723206535970347, + "tokens_seen": 1372979200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002949648946840522, + "loss": 1.3261, + "theoretical_loss": 3.5723206535970347, + "tokens_seen": 1372979200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029495486459378133, + "loss": 1.4152, + "theoretical_loss": 3.5723054852054528, + "tokens_seen": 1373044736 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029494483450351057, + "loss": 1.3856, + "theoretical_loss": 3.572290317740555, + "tokens_seen": 1373110272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029493480441323975, + "loss": 1.3411, + "theoretical_loss": 3.57227515120224, + "tokens_seen": 1373175808 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029492477432296893, + "loss": 1.5791, + "theoretical_loss": 3.5722599855904065, + "tokens_seen": 1373241344 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002949147442326981, + "loss": 1.5266, + "theoretical_loss": 3.5722448209049547, + "tokens_seen": 1373306880 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002949047141424273, + "loss": 1.401, + "theoretical_loss": 3.5722296571457828, + "tokens_seen": 1373372416 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029489468405215647, + "loss": 1.3644, + "theoretical_loss": 3.5722144943127905, + "tokens_seen": 1373437952 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002948846539618857, + "loss": 1.4139, + "theoretical_loss": 3.5721993324058774, + "tokens_seen": 1373503488 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029487462387161483, + "loss": 1.4358, + "theoretical_loss": 3.572184171424942, + "tokens_seen": 1373569024 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029486459378134407, + "loss": 1.3383, + "theoretical_loss": 3.572169011369884, + "tokens_seen": 1373634560 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002948545636910732, + "loss": 1.5257, + "theoretical_loss": 3.5721538522406027, + "tokens_seen": 1373700096 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029484453360080243, + "loss": 1.4873, + "theoretical_loss": 3.572138694036997, + "tokens_seen": 1373765632 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002948345035105316, + "loss": 1.3984, + "theoretical_loss": 3.572123536758967, + "tokens_seen": 1373831168 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002948244734202608, + "loss": 1.381, + "theoretical_loss": 3.5721083804064113, + "tokens_seen": 1373896704 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029481444332999, + "loss": 1.414, + "theoretical_loss": 3.5720932249792297, + "tokens_seen": 1373962240 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002948044132397192, + "loss": 1.3406, + "theoretical_loss": 3.572078070477321, + "tokens_seen": 1374027776 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029479438314944834, + "loss": 1.3908, + "theoretical_loss": 3.5720629169005855, + "tokens_seen": 1374093312 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029478435305917757, + "loss": 1.3496, + "theoretical_loss": 3.572047764248922, + "tokens_seen": 1374158848 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002947743229689067, + "loss": 1.3125, + "theoretical_loss": 3.5720326125222295, + "tokens_seen": 1374224384 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029476429287863593, + "loss": 1.3303, + "theoretical_loss": 3.5720174617204083, + "tokens_seen": 1374289920 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002947542627883651, + "loss": 1.3461, + "theoretical_loss": 3.5720023118433577, + "tokens_seen": 1374355456 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002947442326980943, + "loss": 1.3171, + "theoretical_loss": 3.5719871628909763, + "tokens_seen": 1374420992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002947342026078235, + "loss": 1.3792, + "theoretical_loss": 3.5719720148631646, + "tokens_seen": 1374486528 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029472417251755266, + "loss": 1.3589, + "theoretical_loss": 3.5719568677598215, + "tokens_seen": 1374552064 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029471414242728184, + "loss": 1.3859, + "theoretical_loss": 3.5719417215808473, + "tokens_seen": 1374617600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002947041123370111, + "loss": 1.3014, + "theoretical_loss": 3.57192657632614, + "tokens_seen": 1374683136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002946940822467402, + "loss": 1.2556, + "theoretical_loss": 3.571911431995601, + "tokens_seen": 1374748672 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029468405215646944, + "loss": 1.4008, + "theoretical_loss": 3.5718962885891283, + "tokens_seen": 1374814208 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029467402206619856, + "loss": 1.2945, + "theoretical_loss": 3.571881146106622, + "tokens_seen": 1374879744 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002946639919759278, + "loss": 1.4051, + "theoretical_loss": 3.5718660045479824, + "tokens_seen": 1374945280 + }, + { + "epoch": 0.42, + "learning_rate": 0.000294653961885657, + "loss": 1.3038, + "theoretical_loss": 3.5718508639131077, + "tokens_seen": 1375010816 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029464393179538616, + "loss": 1.3566, + "theoretical_loss": 3.5718357242018985, + "tokens_seen": 1375076352 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029463390170511534, + "loss": 1.3636, + "theoretical_loss": 3.5718205854142546, + "tokens_seen": 1375141888 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002946238716148446, + "loss": 1.3337, + "theoretical_loss": 3.571805447550075, + "tokens_seen": 1375207424 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002946138415245737, + "loss": 1.5072, + "theoretical_loss": 3.57179031060926, + "tokens_seen": 1375272960 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029460381143430294, + "loss": 1.3369, + "theoretical_loss": 3.5717751745917083, + "tokens_seen": 1375338496 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029459378134403207, + "loss": 1.3493, + "theoretical_loss": 3.571760039497321, + "tokens_seen": 1375404032 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002945837512537613, + "loss": 1.4441, + "theoretical_loss": 3.571744905325996, + "tokens_seen": 1375469568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002945737211634905, + "loss": 1.3614, + "theoretical_loss": 3.571729772077635, + "tokens_seen": 1375535104 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029456369107321966, + "loss": 1.3698, + "theoretical_loss": 3.571714639752136, + "tokens_seen": 1375600640 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029455366098294885, + "loss": 1.5319, + "theoretical_loss": 3.5716995083494005, + "tokens_seen": 1375666176 + }, + { + "epoch": 0.42, + "learning_rate": 0.000294543630892678, + "loss": 1.3523, + "theoretical_loss": 3.5716843778693264, + "tokens_seen": 1375731712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002945336008024072, + "loss": 1.3167, + "theoretical_loss": 3.571669248311815, + "tokens_seen": 1375797248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029452357071213644, + "loss": 1.4226, + "theoretical_loss": 3.571654119676765, + "tokens_seen": 1375862784 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029451354062186557, + "loss": 1.555, + "theoretical_loss": 3.571638991964077, + "tokens_seen": 1375928320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002945035105315948, + "loss": 1.423, + "theoretical_loss": 3.571623865173651, + "tokens_seen": 1375993856 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029449348044132393, + "loss": 1.4094, + "theoretical_loss": 3.571608739305386, + "tokens_seen": 1376059392 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029448345035105317, + "loss": 1.3672, + "theoretical_loss": 3.5715936143591827, + "tokens_seen": 1376124928 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029447342026078235, + "loss": 1.3273, + "theoretical_loss": 3.5715784903349403, + "tokens_seen": 1376190464 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 502650, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2454359531402588, + "objective/train/theoretical_loss": 3.5715633672325593, + "objective/train/tokens_used": 1396716000, + "theoretical_loss": 3.5715633672325593, + "tokens_seen": 1376256000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029446339017051153, + "loss": 1.3078, + "theoretical_loss": 3.5715633672325593, + "tokens_seen": 1376256000 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002944533600802407, + "loss": 1.2641, + "theoretical_loss": 3.5715482450519387, + "tokens_seen": 1376321536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029444332998996995, + "loss": 1.4236, + "theoretical_loss": 3.5715331237929795, + "tokens_seen": 1376387072 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002944332998996991, + "loss": 1.3348, + "theoretical_loss": 3.571518003455581, + "tokens_seen": 1376452608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002944232698094283, + "loss": 1.2506, + "theoretical_loss": 3.5715028840396434, + "tokens_seen": 1376518144 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029441323971915744, + "loss": 1.3658, + "theoretical_loss": 3.5714877655450668, + "tokens_seen": 1376583680 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029440320962888667, + "loss": 1.3769, + "theoretical_loss": 3.571472647971751, + "tokens_seen": 1376649216 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029439317953861585, + "loss": 1.3314, + "theoretical_loss": 3.571457531319595, + "tokens_seen": 1376714752 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029438314944834503, + "loss": 1.4205, + "theoretical_loss": 3.571442415588501, + "tokens_seen": 1376780288 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943731193580742, + "loss": 1.4725, + "theoretical_loss": 3.5714273007783675, + "tokens_seen": 1376845824 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943630892678034, + "loss": 1.3516, + "theoretical_loss": 3.5714121868890953, + "tokens_seen": 1376911360 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943530591775326, + "loss": 1.3562, + "theoretical_loss": 3.571397073920584, + "tokens_seen": 1376976896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943430290872618, + "loss": 1.3708, + "theoretical_loss": 3.5713819618727336, + "tokens_seen": 1377042432 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029433299899699094, + "loss": 1.3252, + "theoretical_loss": 3.571366850745444, + "tokens_seen": 1377107968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943229689067202, + "loss": 1.3025, + "theoretical_loss": 3.5713517405386166, + "tokens_seen": 1377173504 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002943129388164493, + "loss": 1.2886, + "theoretical_loss": 3.57133663125215, + "tokens_seen": 1377239040 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029430290872617854, + "loss": 1.3592, + "theoretical_loss": 3.5713215228859454, + "tokens_seen": 1377304576 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029429287863590777, + "loss": 1.406, + "theoretical_loss": 3.571306415439903, + "tokens_seen": 1377370112 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942828485456369, + "loss": 1.3879, + "theoretical_loss": 3.5712913089139215, + "tokens_seen": 1377435648 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029427281845536613, + "loss": 1.333, + "theoretical_loss": 3.5712762033079026, + "tokens_seen": 1377501184 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942627883650953, + "loss": 1.5214, + "theoretical_loss": 3.571261098621746, + "tokens_seen": 1377566720 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942527582748245, + "loss": 1.3087, + "theoretical_loss": 3.571245994855352, + "tokens_seen": 1377632256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942427281845537, + "loss": 1.3876, + "theoretical_loss": 3.5712308920086215, + "tokens_seen": 1377697792 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029423269809428286, + "loss": 1.2841, + "theoretical_loss": 3.571215790081453, + "tokens_seen": 1377763328 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029422266800401204, + "loss": 1.3278, + "theoretical_loss": 3.5712006890737484, + "tokens_seen": 1377828864 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942126379137413, + "loss": 1.1797, + "theoretical_loss": 3.571185588985408, + "tokens_seen": 1377894400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002942026078234704, + "loss": 1.3503, + "theoretical_loss": 3.571170489816331, + "tokens_seen": 1377959936 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029419257773319964, + "loss": 1.3009, + "theoretical_loss": 3.571155391566418, + "tokens_seen": 1378025472 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029418254764292876, + "loss": 1.3391, + "theoretical_loss": 3.57114029423557, + "tokens_seen": 1378091008 + }, + { + "epoch": 0.42, + "learning_rate": 0.000294172517552658, + "loss": 1.4651, + "theoretical_loss": 3.5711251978236866, + "tokens_seen": 1378156544 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002941624874623872, + "loss": 1.3736, + "theoretical_loss": 3.5711101023306693, + "tokens_seen": 1378222080 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029415245737211636, + "loss": 1.4187, + "theoretical_loss": 3.5710950077564165, + "tokens_seen": 1378287616 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029414242728184554, + "loss": 1.3361, + "theoretical_loss": 3.571079914100831, + "tokens_seen": 1378353152 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002941323971915748, + "loss": 1.3996, + "theoretical_loss": 3.5710648213638114, + "tokens_seen": 1378418688 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002941223671013039, + "loss": 1.4555, + "theoretical_loss": 3.5710497295452592, + "tokens_seen": 1378484224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029411233701103314, + "loss": 1.3777, + "theoretical_loss": 3.5710346386450738, + "tokens_seen": 1378549760 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029410230692076227, + "loss": 1.3878, + "theoretical_loss": 3.571019548663157, + "tokens_seen": 1378615296 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002940922768304915, + "loss": 1.3495, + "theoretical_loss": 3.5710044595994077, + "tokens_seen": 1378680832 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002940822467402207, + "loss": 1.4022, + "theoretical_loss": 3.570989371453728, + "tokens_seen": 1378746368 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029407221664994986, + "loss": 1.4286, + "theoretical_loss": 3.5709742842260175, + "tokens_seen": 1378811904 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029406218655967905, + "loss": 1.3914, + "theoretical_loss": 3.570959197916176, + "tokens_seen": 1378877440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002940521564694082, + "loss": 1.4021, + "theoretical_loss": 3.570944112524106, + "tokens_seen": 1378942976 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002940421263791374, + "loss": 1.4069, + "theoretical_loss": 3.5709290280497066, + "tokens_seen": 1379008512 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029403209628886664, + "loss": 1.2478, + "theoretical_loss": 3.5709139444928786, + "tokens_seen": 1379074048 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029402206619859577, + "loss": 1.4725, + "theoretical_loss": 3.5708988618535225, + "tokens_seen": 1379139584 + }, + { + "epoch": 0.42, + "learning_rate": 0.000294012036108325, + "loss": 1.4858, + "theoretical_loss": 3.5708837801315396, + "tokens_seen": 1379205120 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029400200601805413, + "loss": 1.3533, + "theoretical_loss": 3.57086869932683, + "tokens_seen": 1379270656 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029399197592778337, + "loss": 1.4129, + "theoretical_loss": 3.570853619439294, + "tokens_seen": 1379336192 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029398194583751255, + "loss": 1.1683, + "theoretical_loss": 3.5708385404688325, + "tokens_seen": 1379401728 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029397191574724173, + "loss": 1.4125, + "theoretical_loss": 3.5708234624153468, + "tokens_seen": 1379467264 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 503826, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4077787399291992, + "objective/train/theoretical_loss": 3.5708083852787365, + "objective/train/tokens_used": 1399992800, + "theoretical_loss": 3.5708083852787365, + "tokens_seen": 1379532800 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002939618856569709, + "loss": 1.396, + "theoretical_loss": 3.5708083852787365, + "tokens_seen": 1379532800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029395185556670015, + "loss": 1.3232, + "theoretical_loss": 3.570793309058903, + "tokens_seen": 1379598336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002939418254764293, + "loss": 1.3811, + "theoretical_loss": 3.570778233755747, + "tokens_seen": 1379663872 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002939317953861585, + "loss": 1.4738, + "theoretical_loss": 3.5707631593691693, + "tokens_seen": 1379729408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029392176529588764, + "loss": 1.4145, + "theoretical_loss": 3.57074808589907, + "tokens_seen": 1379794944 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029391173520561687, + "loss": 1.4045, + "theoretical_loss": 3.5707330133453503, + "tokens_seen": 1379860480 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029390170511534605, + "loss": 1.2871, + "theoretical_loss": 3.570717941707911, + "tokens_seen": 1379926016 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029389167502507523, + "loss": 1.4134, + "theoretical_loss": 3.570702870986653, + "tokens_seen": 1379991552 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938816449348044, + "loss": 1.4282, + "theoretical_loss": 3.5706878011814767, + "tokens_seen": 1380057088 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938716148445336, + "loss": 1.424, + "theoretical_loss": 3.570672732292283, + "tokens_seen": 1380122624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938615847542628, + "loss": 1.364, + "theoretical_loss": 3.570657664318973, + "tokens_seen": 1380188160 + }, + { + "epoch": 0.42, + "learning_rate": 0.000293851554663992, + "loss": 1.4092, + "theoretical_loss": 3.5706425972614477, + "tokens_seen": 1380253696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029384152457372114, + "loss": 1.3612, + "theoretical_loss": 3.570627531119607, + "tokens_seen": 1380319232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938314944834504, + "loss": 1.3347, + "theoretical_loss": 3.570612465893353, + "tokens_seen": 1380384768 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938214643931795, + "loss": 1.4392, + "theoretical_loss": 3.570597401582586, + "tokens_seen": 1380450304 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029381143430290874, + "loss": 1.3511, + "theoretical_loss": 3.570582338187207, + "tokens_seen": 1380515840 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002938014042126379, + "loss": 1.3621, + "theoretical_loss": 3.5705672757071167, + "tokens_seen": 1380581376 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937913741223671, + "loss": 1.3846, + "theoretical_loss": 3.570552214142216, + "tokens_seen": 1380646912 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937813440320963, + "loss": 1.4568, + "theoretical_loss": 3.570537153492406, + "tokens_seen": 1380712448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937713139418255, + "loss": 1.3283, + "theoretical_loss": 3.5705220937575888, + "tokens_seen": 1380777984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029376128385155464, + "loss": 1.399, + "theoretical_loss": 3.5705070349376635, + "tokens_seen": 1380843520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937512537612839, + "loss": 1.2304, + "theoretical_loss": 3.570491977032532, + "tokens_seen": 1380909056 + }, + { + "epoch": 0.42, + "learning_rate": 0.000293741223671013, + "loss": 1.3645, + "theoretical_loss": 3.570476920042095, + "tokens_seen": 1380974592 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029373119358074224, + "loss": 1.3382, + "theoretical_loss": 3.5704618639662544, + "tokens_seen": 1381040128 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937211634904714, + "loss": 1.3461, + "theoretical_loss": 3.5704468088049106, + "tokens_seen": 1381105664 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937111334002006, + "loss": 1.4141, + "theoretical_loss": 3.5704317545579647, + "tokens_seen": 1381171200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002937011033099298, + "loss": 1.4254, + "theoretical_loss": 3.570416701225317, + "tokens_seen": 1381236736 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029369107321965896, + "loss": 1.3538, + "theoretical_loss": 3.57040164880687, + "tokens_seen": 1381302272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029368104312938815, + "loss": 1.3753, + "theoretical_loss": 3.570386597302524, + "tokens_seen": 1381367808 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002936710130391174, + "loss": 1.3122, + "theoretical_loss": 3.57037154671218, + "tokens_seen": 1381433344 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002936609829488465, + "loss": 1.4713, + "theoretical_loss": 3.57035649703574, + "tokens_seen": 1381498880 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029365095285857574, + "loss": 1.328, + "theoretical_loss": 3.570341448273105, + "tokens_seen": 1381564416 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029364092276830487, + "loss": 1.3703, + "theoretical_loss": 3.570326400424175, + "tokens_seen": 1381629952 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002936308926780341, + "loss": 1.3841, + "theoretical_loss": 3.570311353488852, + "tokens_seen": 1381695488 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002936208625877633, + "loss": 1.3335, + "theoretical_loss": 3.5702963074670375, + "tokens_seen": 1381761024 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029361083249749247, + "loss": 1.3977, + "theoretical_loss": 3.570281262358632, + "tokens_seen": 1381826560 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029360080240722165, + "loss": 1.3861, + "theoretical_loss": 3.5702662181635376, + "tokens_seen": 1381892096 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002935907723169509, + "loss": 1.2856, + "theoretical_loss": 3.5702511748816548, + "tokens_seen": 1381957632 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029358074222668, + "loss": 1.4613, + "theoretical_loss": 3.5702361325128846, + "tokens_seen": 1382023168 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029357071213640925, + "loss": 1.4333, + "theoretical_loss": 3.57022109105713, + "tokens_seen": 1382088704 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002935606820461384, + "loss": 1.3261, + "theoretical_loss": 3.5702060505142903, + "tokens_seen": 1382154240 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002935506519558676, + "loss": 1.3999, + "theoretical_loss": 3.570191010884267, + "tokens_seen": 1382219776 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029354062186559684, + "loss": 1.3588, + "theoretical_loss": 3.5701759721669633, + "tokens_seen": 1382285312 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029353059177532597, + "loss": 1.3555, + "theoretical_loss": 3.5701609343622787, + "tokens_seen": 1382350848 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002935205616850552, + "loss": 1.3147, + "theoretical_loss": 3.5701458974701152, + "tokens_seen": 1382416384 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029351053159478433, + "loss": 1.3901, + "theoretical_loss": 3.570130861490374, + "tokens_seen": 1382481920 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029350050150451357, + "loss": 1.2792, + "theoretical_loss": 3.570115826422956, + "tokens_seen": 1382547456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029349047141424275, + "loss": 1.3561, + "theoretical_loss": 3.5701007922677643, + "tokens_seen": 1382612992 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029348044132397193, + "loss": 1.4844, + "theoretical_loss": 3.570085759024699, + "tokens_seen": 1382678528 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002934704112337011, + "loss": 1.366, + "theoretical_loss": 3.570070726693661, + "tokens_seen": 1382744064 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 505055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.323938012123108, + "objective/train/theoretical_loss": 3.570055695274553, + "objective/train/tokens_used": 1403269600, + "theoretical_loss": 3.570055695274553, + "tokens_seen": 1382809600 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029346038114343035, + "loss": 1.3988, + "theoretical_loss": 3.570055695274553, + "tokens_seen": 1382809600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002934503510531595, + "loss": 1.3223, + "theoretical_loss": 3.570040664767276, + "tokens_seen": 1382875136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002934403209628887, + "loss": 1.3264, + "theoretical_loss": 3.570025635171731, + "tokens_seen": 1382940672 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029343029087261784, + "loss": 1.3829, + "theoretical_loss": 3.5700106064878203, + "tokens_seen": 1383006208 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029342026078234707, + "loss": 1.4483, + "theoretical_loss": 3.569995578715445, + "tokens_seen": 1383071744 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029341023069207625, + "loss": 1.3717, + "theoretical_loss": 3.5699805518545062, + "tokens_seen": 1383137280 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029340020060180543, + "loss": 1.3767, + "theoretical_loss": 3.569965525904906, + "tokens_seen": 1383202816 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933901705115346, + "loss": 1.3442, + "theoretical_loss": 3.569950500866547, + "tokens_seen": 1383268352 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933801404212638, + "loss": 1.2778, + "theoretical_loss": 3.569935476739328, + "tokens_seen": 1383333888 + }, + { + "epoch": 0.42, + "learning_rate": 0.000293370110330993, + "loss": 1.4576, + "theoretical_loss": 3.569920453523153, + "tokens_seen": 1383399424 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933600802407222, + "loss": 1.3992, + "theoretical_loss": 3.569905431217923, + "tokens_seen": 1383464960 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029335005015045134, + "loss": 1.3184, + "theoretical_loss": 3.5698904098235387, + "tokens_seen": 1383530496 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933400200601806, + "loss": 1.3497, + "theoretical_loss": 3.569875389339903, + "tokens_seen": 1383596032 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933299899699097, + "loss": 1.3334, + "theoretical_loss": 3.569860369766917, + "tokens_seen": 1383661568 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029331995987963894, + "loss": 1.3553, + "theoretical_loss": 3.5698453511044823, + "tokens_seen": 1383727104 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002933099297893681, + "loss": 1.3898, + "theoretical_loss": 3.5698303333525003, + "tokens_seen": 1383792640 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932998996990973, + "loss": 1.4371, + "theoretical_loss": 3.5698153165108732, + "tokens_seen": 1383858176 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932898696088265, + "loss": 1.3479, + "theoretical_loss": 3.5698003005795025, + "tokens_seen": 1383923712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932798395185557, + "loss": 1.4217, + "theoretical_loss": 3.5697852855582903, + "tokens_seen": 1383989248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029326980942828484, + "loss": 1.2558, + "theoretical_loss": 3.5697702714471378, + "tokens_seen": 1384054784 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932597793380141, + "loss": 1.325, + "theoretical_loss": 3.5697552582459466, + "tokens_seen": 1384120320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932497492477432, + "loss": 1.3302, + "theoretical_loss": 3.569740245954619, + "tokens_seen": 1384185856 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029323971915747244, + "loss": 1.3499, + "theoretical_loss": 3.5697252345730566, + "tokens_seen": 1384251392 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932296890672016, + "loss": 1.3523, + "theoretical_loss": 3.5697102241011613, + "tokens_seen": 1384316928 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002932196589769308, + "loss": 1.4799, + "theoretical_loss": 3.569695214538835, + "tokens_seen": 1384382464 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029320962888666, + "loss": 1.425, + "theoretical_loss": 3.569680205885979, + "tokens_seen": 1384448000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029319959879638916, + "loss": 1.4776, + "theoretical_loss": 3.5696651981424954, + "tokens_seen": 1384513536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029318956870611835, + "loss": 1.3932, + "theoretical_loss": 3.5696501913082863, + "tokens_seen": 1384579072 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002931795386158476, + "loss": 1.3372, + "theoretical_loss": 3.5696351853832535, + "tokens_seen": 1384644608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002931695085255767, + "loss": 1.426, + "theoretical_loss": 3.5696201803672984, + "tokens_seen": 1384710144 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029315947843530594, + "loss": 1.3561, + "theoretical_loss": 3.569605176260324, + "tokens_seen": 1384775680 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029314944834503507, + "loss": 1.3757, + "theoretical_loss": 3.569590173062231, + "tokens_seen": 1384841216 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002931394182547643, + "loss": 1.3143, + "theoretical_loss": 3.569575170772922, + "tokens_seen": 1384906752 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002931293881644935, + "loss": 1.4423, + "theoretical_loss": 3.5695601693922985, + "tokens_seen": 1384972288 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029311935807422267, + "loss": 1.3774, + "theoretical_loss": 3.5695451689202633, + "tokens_seen": 1385037824 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029310932798395185, + "loss": 1.4702, + "theoretical_loss": 3.569530169356718, + "tokens_seen": 1385103360 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002930992978936811, + "loss": 1.3255, + "theoretical_loss": 3.569515170701564, + "tokens_seen": 1385168896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002930892678034102, + "loss": 1.3182, + "theoretical_loss": 3.569500172954704, + "tokens_seen": 1385234432 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029307923771313945, + "loss": 1.4936, + "theoretical_loss": 3.5694851761160398, + "tokens_seen": 1385299968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002930692076228686, + "loss": 1.4612, + "theoretical_loss": 3.5694701801854736, + "tokens_seen": 1385365504 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002930591775325978, + "loss": 1.4409, + "theoretical_loss": 3.569455185162907, + "tokens_seen": 1385431040 + }, + { + "epoch": 0.42, + "learning_rate": 0.000293049147442327, + "loss": 1.3291, + "theoretical_loss": 3.5694401910482423, + "tokens_seen": 1385496576 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029303911735205617, + "loss": 1.3514, + "theoretical_loss": 3.569425197841382, + "tokens_seen": 1385562112 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029302908726178535, + "loss": 1.38, + "theoretical_loss": 3.5694102055422277, + "tokens_seen": 1385627648 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029301905717151453, + "loss": 1.3302, + "theoretical_loss": 3.5693952141506813, + "tokens_seen": 1385693184 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002930090270812437, + "loss": 1.3076, + "theoretical_loss": 3.569380223666646, + "tokens_seen": 1385758720 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029299899699097295, + "loss": 1.2907, + "theoretical_loss": 3.5693652340900233, + "tokens_seen": 1385824256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929889669007021, + "loss": 1.4109, + "theoretical_loss": 3.569350245420715, + "tokens_seen": 1385889792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929789368104313, + "loss": 1.4025, + "theoretical_loss": 3.5693352576586235, + "tokens_seen": 1385955328 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929689067201605, + "loss": 1.3911, + "theoretical_loss": 3.5693202708036513, + "tokens_seen": 1386020864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 506209, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.536266803741455, + "objective/train/theoretical_loss": 3.5693052848557003, + "objective/train/tokens_used": 1406546400, + "theoretical_loss": 3.5693052848557003, + "tokens_seen": 1386086400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929588766298897, + "loss": 1.4144, + "theoretical_loss": 3.5693052848557003, + "tokens_seen": 1386086400 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029294884653961885, + "loss": 1.3935, + "theoretical_loss": 3.569290299814673, + "tokens_seen": 1386151936 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029293881644934804, + "loss": 1.3449, + "theoretical_loss": 3.5692753156804713, + "tokens_seen": 1386217472 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929287863590772, + "loss": 1.2816, + "theoretical_loss": 3.569260332452998, + "tokens_seen": 1386283008 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029291875626880645, + "loss": 1.3474, + "theoretical_loss": 3.569245350132155, + "tokens_seen": 1386348544 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002929087261785356, + "loss": 1.2158, + "theoretical_loss": 3.569230368717844, + "tokens_seen": 1386414080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928986960882648, + "loss": 1.4378, + "theoretical_loss": 3.569215388209968, + "tokens_seen": 1386479616 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029288866599799394, + "loss": 1.3756, + "theoretical_loss": 3.56920040860843, + "tokens_seen": 1386545152 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928786359077232, + "loss": 1.399, + "theoretical_loss": 3.569185429913131, + "tokens_seen": 1386610688 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029286860581745236, + "loss": 1.3645, + "theoretical_loss": 3.5691704521239735, + "tokens_seen": 1386676224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029285857572718154, + "loss": 1.3461, + "theoretical_loss": 3.569155475240861, + "tokens_seen": 1386741760 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928485456369107, + "loss": 1.3331, + "theoretical_loss": 3.5691404992636944, + "tokens_seen": 1386807296 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928385155466399, + "loss": 1.2964, + "theoretical_loss": 3.569125524192377, + "tokens_seen": 1386872832 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928284854563691, + "loss": 1.3275, + "theoretical_loss": 3.5691105500268114, + "tokens_seen": 1386938368 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002928184553660983, + "loss": 1.3198, + "theoretical_loss": 3.5690955767669, + "tokens_seen": 1387003904 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029280842527582744, + "loss": 1.393, + "theoretical_loss": 3.569080604412544, + "tokens_seen": 1387069440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927983951855567, + "loss": 1.3888, + "theoretical_loss": 3.569065632963647, + "tokens_seen": 1387134976 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927883650952859, + "loss": 1.4397, + "theoretical_loss": 3.5690506624201115, + "tokens_seen": 1387200512 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029277833500501504, + "loss": 1.5466, + "theoretical_loss": 3.5690356927818394, + "tokens_seen": 1387266048 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927683049147443, + "loss": 1.3518, + "theoretical_loss": 3.5690207240487335, + "tokens_seen": 1387331584 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927582748244734, + "loss": 1.3708, + "theoretical_loss": 3.5690057562206965, + "tokens_seen": 1387397120 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029274824473420264, + "loss": 1.4008, + "theoretical_loss": 3.5689907892976303, + "tokens_seen": 1387462656 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927382146439318, + "loss": 1.209, + "theoretical_loss": 3.568975823279438, + "tokens_seen": 1387528192 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292728184553661, + "loss": 1.3838, + "theoretical_loss": 3.5689608581660224, + "tokens_seen": 1387593728 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002927181544633902, + "loss": 1.6254, + "theoretical_loss": 3.5689458939572853, + "tokens_seen": 1387659264 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029270812437311936, + "loss": 1.3853, + "theoretical_loss": 3.5689309306531296, + "tokens_seen": 1387724800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029269809428284855, + "loss": 1.2637, + "theoretical_loss": 3.5689159682534584, + "tokens_seen": 1387790336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002926880641925778, + "loss": 1.3828, + "theoretical_loss": 3.568901006758174, + "tokens_seen": 1387855872 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002926780341023069, + "loss": 1.2901, + "theoretical_loss": 3.5688860461671785, + "tokens_seen": 1387921408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029266800401203614, + "loss": 1.3665, + "theoretical_loss": 3.5688710864803745, + "tokens_seen": 1387986944 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029265797392176527, + "loss": 1.2838, + "theoretical_loss": 3.568856127697666, + "tokens_seen": 1388052480 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002926479438314945, + "loss": 1.2232, + "theoretical_loss": 3.5688411698189544, + "tokens_seen": 1388118016 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002926379137412237, + "loss": 1.417, + "theoretical_loss": 3.5688262128441433, + "tokens_seen": 1388183552 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029262788365095287, + "loss": 1.3023, + "theoretical_loss": 3.5688112567731345, + "tokens_seen": 1388249088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029261785356068205, + "loss": 1.3111, + "theoretical_loss": 3.5687963016058313, + "tokens_seen": 1388314624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002926078234704113, + "loss": 1.41, + "theoretical_loss": 3.568781347342136, + "tokens_seen": 1388380160 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002925977933801404, + "loss": 1.2571, + "theoretical_loss": 3.568766393981952, + "tokens_seen": 1388445696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029258776328986965, + "loss": 1.454, + "theoretical_loss": 3.5687514415251815, + "tokens_seen": 1388511232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002925777331995988, + "loss": 1.3622, + "theoretical_loss": 3.5687364899717275, + "tokens_seen": 1388576768 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292567703109328, + "loss": 1.3925, + "theoretical_loss": 3.5687215393214933, + "tokens_seen": 1388642304 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002925576730190572, + "loss": 1.3003, + "theoretical_loss": 3.5687065895743806, + "tokens_seen": 1388707840 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029254764292878637, + "loss": 1.4197, + "theoretical_loss": 3.568691640730293, + "tokens_seen": 1388773376 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029253761283851555, + "loss": 1.4206, + "theoretical_loss": 3.5686766927891336, + "tokens_seen": 1388838912 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029252758274824473, + "loss": 1.3539, + "theoretical_loss": 3.568661745750804, + "tokens_seen": 1388904448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002925175526579739, + "loss": 1.3825, + "theoretical_loss": 3.5686467996152085, + "tokens_seen": 1388969984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029250752256770315, + "loss": 1.3974, + "theoretical_loss": 3.5686318543822493, + "tokens_seen": 1389035520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924974924774323, + "loss": 1.4961, + "theoretical_loss": 3.5686169100518295, + "tokens_seen": 1389101056 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924874623871615, + "loss": 1.3811, + "theoretical_loss": 3.5686019666238513, + "tokens_seen": 1389166592 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924774322968907, + "loss": 1.3193, + "theoretical_loss": 3.568587024098219, + "tokens_seen": 1389232128 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924674022066199, + "loss": 1.5015, + "theoretical_loss": 3.5685720824748346, + "tokens_seen": 1389297664 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 507487, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3491973876953125, + "objective/train/theoretical_loss": 3.568557141753601, + "objective/train/tokens_used": 1409823200, + "theoretical_loss": 3.568557141753601, + "tokens_seen": 1389363200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029245737211634906, + "loss": 1.3113, + "theoretical_loss": 3.568557141753601, + "tokens_seen": 1389363200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029244734202607824, + "loss": 1.3645, + "theoretical_loss": 3.568542201934422, + "tokens_seen": 1389428736 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924373119358074, + "loss": 1.3571, + "theoretical_loss": 3.5685272630171996, + "tokens_seen": 1389494272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029242728184553665, + "loss": 1.2825, + "theoretical_loss": 3.568512325001837, + "tokens_seen": 1389559808 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002924172517552658, + "loss": 1.3417, + "theoretical_loss": 3.5684973878882382, + "tokens_seen": 1389625344 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292407221664995, + "loss": 1.484, + "theoretical_loss": 3.568482451676305, + "tokens_seen": 1389690880 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029239719157472414, + "loss": 1.4833, + "theoretical_loss": 3.568467516365941, + "tokens_seen": 1389756416 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923871614844534, + "loss": 1.3895, + "theoretical_loss": 3.5684525819570494, + "tokens_seen": 1389821952 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029237713139418256, + "loss": 1.3631, + "theoretical_loss": 3.5684376484495335, + "tokens_seen": 1389887488 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029236710130391174, + "loss": 1.259, + "theoretical_loss": 3.5684227158432953, + "tokens_seen": 1389953024 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923570712136409, + "loss": 1.3667, + "theoretical_loss": 3.5684077841382393, + "tokens_seen": 1390018560 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923470411233701, + "loss": 1.2769, + "theoretical_loss": 3.5683928533342675, + "tokens_seen": 1390084096 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923370110330993, + "loss": 1.3758, + "theoretical_loss": 3.5683779234312833, + "tokens_seen": 1390149632 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923269809428285, + "loss": 1.2269, + "theoretical_loss": 3.5683629944291906, + "tokens_seen": 1390215168 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029231695085255765, + "loss": 1.3069, + "theoretical_loss": 3.568348066327892, + "tokens_seen": 1390280704 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002923069207622869, + "loss": 1.2937, + "theoretical_loss": 3.56833313912729, + "tokens_seen": 1390346240 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029229689067201606, + "loss": 1.4146, + "theoretical_loss": 3.5683182128272897, + "tokens_seen": 1390411776 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029228686058174524, + "loss": 1.4568, + "theoretical_loss": 3.5683032874277925, + "tokens_seen": 1390477312 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002922768304914744, + "loss": 1.3558, + "theoretical_loss": 3.568288362928702, + "tokens_seen": 1390542848 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002922668004012036, + "loss": 1.4295, + "theoretical_loss": 3.5682734393299222, + "tokens_seen": 1390608384 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002922567703109328, + "loss": 1.3979, + "theoretical_loss": 3.568258516631356, + "tokens_seen": 1390673920 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292246740220662, + "loss": 1.3485, + "theoretical_loss": 3.5682435948329063, + "tokens_seen": 1390739456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029223671013039115, + "loss": 1.2323, + "theoretical_loss": 3.5682286739344766, + "tokens_seen": 1390804992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002922266800401204, + "loss": 1.3047, + "theoretical_loss": 3.5682137539359706, + "tokens_seen": 1390870528 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002922166499498495, + "loss": 1.3693, + "theoretical_loss": 3.5681988348372915, + "tokens_seen": 1390936064 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029220661985957875, + "loss": 1.3146, + "theoretical_loss": 3.568183916638342, + "tokens_seen": 1391001600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002921965897693079, + "loss": 1.3916, + "theoretical_loss": 3.5681689993390258, + "tokens_seen": 1391067136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002921865596790371, + "loss": 1.3361, + "theoretical_loss": 3.568154082939247, + "tokens_seen": 1391132672 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002921765295887663, + "loss": 1.2924, + "theoretical_loss": 3.5681391674389076, + "tokens_seen": 1391198208 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029216649949849547, + "loss": 1.2826, + "theoretical_loss": 3.568124252837912, + "tokens_seen": 1391263744 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029215646940822465, + "loss": 1.3885, + "theoretical_loss": 3.5681093391361633, + "tokens_seen": 1391329280 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002921464393179539, + "loss": 1.3487, + "theoretical_loss": 3.568094426333565, + "tokens_seen": 1391394816 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292136409227683, + "loss": 1.4443, + "theoretical_loss": 3.5680795144300212, + "tokens_seen": 1391460352 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029212637913741225, + "loss": 1.2709, + "theoretical_loss": 3.5680646034254337, + "tokens_seen": 1391525888 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029211634904714143, + "loss": 1.3588, + "theoretical_loss": 3.5680496933197077, + "tokens_seen": 1391591424 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002921063189568706, + "loss": 1.2229, + "theoretical_loss": 3.5680347841127453, + "tokens_seen": 1391656960 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002920962888665998, + "loss": 1.366, + "theoretical_loss": 3.5680198758044512, + "tokens_seen": 1391722496 + }, + { + "epoch": 0.42, + "learning_rate": 0.000292086258776329, + "loss": 1.3752, + "theoretical_loss": 3.568004968394728, + "tokens_seen": 1391788032 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029207622868605815, + "loss": 1.2861, + "theoretical_loss": 3.56799006188348, + "tokens_seen": 1391853568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002920661985957874, + "loss": 1.3198, + "theoretical_loss": 3.5679751562706103, + "tokens_seen": 1391919104 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002920561685055165, + "loss": 1.3935, + "theoretical_loss": 3.5679602515560225, + "tokens_seen": 1391984640 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029204613841524575, + "loss": 1.3628, + "theoretical_loss": 3.5679453477396197, + "tokens_seen": 1392050176 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029203610832497493, + "loss": 1.3241, + "theoretical_loss": 3.5679304448213065, + "tokens_seen": 1392115712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002920260782347041, + "loss": 1.2898, + "theoretical_loss": 3.5679155428009857, + "tokens_seen": 1392181248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029201604814443335, + "loss": 1.3542, + "theoretical_loss": 3.5679006416785617, + "tokens_seen": 1392246784 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002920060180541625, + "loss": 1.3609, + "theoretical_loss": 3.567885741453937, + "tokens_seen": 1392312320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002919959879638917, + "loss": 1.3776, + "theoretical_loss": 3.567870842127016, + "tokens_seen": 1392377856 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002919859578736209, + "loss": 1.278, + "theoretical_loss": 3.567855943697703, + "tokens_seen": 1392443392 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002919759277833501, + "loss": 1.4106, + "theoretical_loss": 3.5678410461659, + "tokens_seen": 1392508928 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029196589769307926, + "loss": 1.4167, + "theoretical_loss": 3.567826149531512, + "tokens_seen": 1392574464 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 508733, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2427304983139038, + "objective/train/theoretical_loss": 3.567811253794442, + "objective/train/tokens_used": 1413100000, + "theoretical_loss": 3.567811253794442, + "tokens_seen": 1392640000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029195586760280844, + "loss": 1.2811, + "theoretical_loss": 3.567811253794442, + "tokens_seen": 1392640000 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002919458375125376, + "loss": 1.4211, + "theoretical_loss": 3.5677963589545953, + "tokens_seen": 1392705536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029193580742226685, + "loss": 1.3912, + "theoretical_loss": 3.5677814650118735, + "tokens_seen": 1392771072 + }, + { + "epoch": 0.42, + "learning_rate": 0.000291925777331996, + "loss": 1.3873, + "theoretical_loss": 3.567766571966181, + "tokens_seen": 1392836608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002919157472417252, + "loss": 1.3779, + "theoretical_loss": 3.5677516798174223, + "tokens_seen": 1392902144 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029190571715145434, + "loss": 1.2787, + "theoretical_loss": 3.567736788565501, + "tokens_seen": 1392967680 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918956870611836, + "loss": 1.3515, + "theoretical_loss": 3.56772189821032, + "tokens_seen": 1393033216 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029188565697091276, + "loss": 1.3656, + "theoretical_loss": 3.567707008751784, + "tokens_seen": 1393098752 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029187562688064194, + "loss": 1.3377, + "theoretical_loss": 3.567692120189797, + "tokens_seen": 1393164288 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918655967903711, + "loss": 1.4593, + "theoretical_loss": 3.567677232524262, + "tokens_seen": 1393229824 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918555667001003, + "loss": 1.4214, + "theoretical_loss": 3.567662345755083, + "tokens_seen": 1393295360 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918455366098295, + "loss": 1.313, + "theoretical_loss": 3.5676474598821644, + "tokens_seen": 1393360896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918355065195587, + "loss": 1.3334, + "theoretical_loss": 3.56763257490541, + "tokens_seen": 1393426432 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029182547642928785, + "loss": 1.3146, + "theoretical_loss": 3.567617690824724, + "tokens_seen": 1393491968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002918154463390171, + "loss": 1.4617, + "theoretical_loss": 3.567602807640009, + "tokens_seen": 1393557504 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029180541624874626, + "loss": 1.3747, + "theoretical_loss": 3.5675879253511704, + "tokens_seen": 1393623040 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029179538615847544, + "loss": 1.3396, + "theoretical_loss": 3.5675730439581113, + "tokens_seen": 1393688576 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917853560682046, + "loss": 1.3599, + "theoretical_loss": 3.5675581634607356, + "tokens_seen": 1393754112 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917753259779338, + "loss": 1.3349, + "theoretical_loss": 3.567543283858948, + "tokens_seen": 1393819648 + }, + { + "epoch": 0.42, + "learning_rate": 0.000291765295887663, + "loss": 1.4545, + "theoretical_loss": 3.567528405152652, + "tokens_seen": 1393885184 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917552657973922, + "loss": 1.276, + "theoretical_loss": 3.5675135273417515, + "tokens_seen": 1393950720 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029174523570712135, + "loss": 1.4397, + "theoretical_loss": 3.5674986504261508, + "tokens_seen": 1394016256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917352056168506, + "loss": 1.2495, + "theoretical_loss": 3.5674837744057544, + "tokens_seen": 1394081792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917251755265797, + "loss": 1.4332, + "theoretical_loss": 3.567468899280465, + "tokens_seen": 1394147328 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029171514543630895, + "loss": 1.4191, + "theoretical_loss": 3.5674540250501883, + "tokens_seen": 1394212864 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002917051153460381, + "loss": 1.3362, + "theoretical_loss": 3.5674391517148267, + "tokens_seen": 1394278400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002916950852557673, + "loss": 1.4345, + "theoretical_loss": 3.567424279274286, + "tokens_seen": 1394343936 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002916850551654965, + "loss": 1.4292, + "theoretical_loss": 3.567409407728469, + "tokens_seen": 1394409472 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029167502507522567, + "loss": 1.3264, + "theoretical_loss": 3.56739453707728, + "tokens_seen": 1394475008 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029166499498495485, + "loss": 1.3838, + "theoretical_loss": 3.567379667320624, + "tokens_seen": 1394540544 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002916549648946841, + "loss": 1.4422, + "theoretical_loss": 3.5673647984584043, + "tokens_seen": 1394606080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002916449348044132, + "loss": 1.3048, + "theoretical_loss": 3.567349930490525, + "tokens_seen": 1394671616 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029163490471414245, + "loss": 1.4515, + "theoretical_loss": 3.567335063416891, + "tokens_seen": 1394737152 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029162487462387163, + "loss": 1.3582, + "theoretical_loss": 3.5673201972374065, + "tokens_seen": 1394802688 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002916148445336008, + "loss": 1.4067, + "theoretical_loss": 3.567305331951975, + "tokens_seen": 1394868224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029160481444333, + "loss": 1.4669, + "theoretical_loss": 3.5672904675605013, + "tokens_seen": 1394933760 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915947843530592, + "loss": 1.3084, + "theoretical_loss": 3.567275604062889, + "tokens_seen": 1394999296 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029158475426278835, + "loss": 1.4443, + "theoretical_loss": 3.5672607414590427, + "tokens_seen": 1395064832 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915747241725176, + "loss": 1.2922, + "theoretical_loss": 3.5672458797488673, + "tokens_seen": 1395130368 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915646940822467, + "loss": 1.4121, + "theoretical_loss": 3.567231018932266, + "tokens_seen": 1395195904 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029155466399197595, + "loss": 1.3951, + "theoretical_loss": 3.5672161590091442, + "tokens_seen": 1395261440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915446339017051, + "loss": 1.3492, + "theoretical_loss": 3.5672012999794056, + "tokens_seen": 1395326976 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915346038114343, + "loss": 1.5855, + "theoretical_loss": 3.5671864418429546, + "tokens_seen": 1395392512 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915245737211635, + "loss": 1.3431, + "theoretical_loss": 3.567171584599695, + "tokens_seen": 1395458048 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002915145436308927, + "loss": 1.3886, + "theoretical_loss": 3.5671567282495324, + "tokens_seen": 1395523584 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029150451354062186, + "loss": 1.4083, + "theoretical_loss": 3.5671418727923703, + "tokens_seen": 1395589120 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914944834503511, + "loss": 1.3878, + "theoretical_loss": 3.567127018228113, + "tokens_seen": 1395654656 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914844533600802, + "loss": 1.3882, + "theoretical_loss": 3.5671121645566655, + "tokens_seen": 1395720192 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029147442326980946, + "loss": 1.2726, + "theoretical_loss": 3.567097311777932, + "tokens_seen": 1395785728 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914643931795386, + "loss": 1.374, + "theoretical_loss": 3.5670824598918163, + "tokens_seen": 1395851264 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 509897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2565758228302002, + "objective/train/theoretical_loss": 3.5670676088982236, + "objective/train/tokens_used": 1416376800, + "theoretical_loss": 3.5670676088982236, + "tokens_seen": 1395916800 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914543630892678, + "loss": 1.3178, + "theoretical_loss": 3.5670676088982236, + "tokens_seen": 1395916800 + }, + { + "epoch": 0.42, + "learning_rate": 0.000291444332998997, + "loss": 1.5093, + "theoretical_loss": 3.5670527587970584, + "tokens_seen": 1395982336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914343029087262, + "loss": 1.4252, + "theoretical_loss": 3.5670379095882248, + "tokens_seen": 1396047872 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029142427281845536, + "loss": 1.3179, + "theoretical_loss": 3.5670230612716276, + "tokens_seen": 1396113408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029141424272818454, + "loss": 1.4475, + "theoretical_loss": 3.5670082138471706, + "tokens_seen": 1396178944 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002914042126379137, + "loss": 1.3243, + "theoretical_loss": 3.5669933673147596, + "tokens_seen": 1396244480 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029139418254764296, + "loss": 1.3782, + "theoretical_loss": 3.566978521674298, + "tokens_seen": 1396310016 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002913841524573721, + "loss": 1.4778, + "theoretical_loss": 3.566963676925691, + "tokens_seen": 1396375552 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002913741223671013, + "loss": 1.324, + "theoretical_loss": 3.5669488330688424, + "tokens_seen": 1396441088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029136409227683045, + "loss": 1.4174, + "theoretical_loss": 3.566933990103658, + "tokens_seen": 1396506624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002913540621865597, + "loss": 1.3395, + "theoretical_loss": 3.566919148030041, + "tokens_seen": 1396572160 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029134403209628886, + "loss": 1.4635, + "theoretical_loss": 3.5669043068478974, + "tokens_seen": 1396637696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029133400200601805, + "loss": 1.3491, + "theoretical_loss": 3.5668894665571305, + "tokens_seen": 1396703232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002913239719157472, + "loss": 1.4167, + "theoretical_loss": 3.566874627157646, + "tokens_seen": 1396768768 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029131394182547646, + "loss": 1.4631, + "theoretical_loss": 3.5668597886493485, + "tokens_seen": 1396834304 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002913039117352056, + "loss": 1.4071, + "theoretical_loss": 3.5668449510321416, + "tokens_seen": 1396899840 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912938816449348, + "loss": 1.4339, + "theoretical_loss": 3.5668301143059313, + "tokens_seen": 1396965376 + }, + { + "epoch": 0.42, + "learning_rate": 0.000291283851554664, + "loss": 1.2767, + "theoretical_loss": 3.5668152784706213, + "tokens_seen": 1397030912 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912738214643932, + "loss": 1.307, + "theoretical_loss": 3.566800443526117, + "tokens_seen": 1397096448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912637913741224, + "loss": 1.505, + "theoretical_loss": 3.5667856094723227, + "tokens_seen": 1397161984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029125376128385155, + "loss": 1.3429, + "theoretical_loss": 3.566770776309143, + "tokens_seen": 1397227520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912437311935808, + "loss": 1.4044, + "theoretical_loss": 3.5667559440364833, + "tokens_seen": 1397293056 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912337011033099, + "loss": 1.4434, + "theoretical_loss": 3.566741112654248, + "tokens_seen": 1397358592 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029122367101303915, + "loss": 1.5147, + "theoretical_loss": 3.566726282162342, + "tokens_seen": 1397424128 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029121364092276833, + "loss": 1.3618, + "theoretical_loss": 3.5667114525606696, + "tokens_seen": 1397489664 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002912036108324975, + "loss": 1.3218, + "theoretical_loss": 3.566696623849136, + "tokens_seen": 1397555200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002911935807422267, + "loss": 1.3617, + "theoretical_loss": 3.5666817960276465, + "tokens_seen": 1397620736 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029118355065195587, + "loss": 1.3742, + "theoretical_loss": 3.5666669690961057, + "tokens_seen": 1397686272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029117352056168505, + "loss": 1.3875, + "theoretical_loss": 3.566652143054418, + "tokens_seen": 1397751808 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002911634904714143, + "loss": 1.3516, + "theoretical_loss": 3.5666373179024884, + "tokens_seen": 1397817344 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002911534603811434, + "loss": 1.4692, + "theoretical_loss": 3.566622493640222, + "tokens_seen": 1397882880 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029114343029087265, + "loss": 1.35, + "theoretical_loss": 3.5666076702675236, + "tokens_seen": 1397948416 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029113340020060183, + "loss": 1.3043, + "theoretical_loss": 3.5665928477842987, + "tokens_seen": 1398013952 + }, + { + "epoch": 0.42, + "learning_rate": 0.000291123370110331, + "loss": 1.3613, + "theoretical_loss": 3.566578026190451, + "tokens_seen": 1398079488 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002911133400200602, + "loss": 1.3683, + "theoretical_loss": 3.5665632054858865, + "tokens_seen": 1398145024 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002911033099297894, + "loss": 1.4763, + "theoretical_loss": 3.56654838567051, + "tokens_seen": 1398210560 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029109327983951855, + "loss": 1.434, + "theoretical_loss": 3.5665335667442255, + "tokens_seen": 1398276096 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910832497492478, + "loss": 1.4237, + "theoretical_loss": 3.5665187487069394, + "tokens_seen": 1398341632 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910732196589769, + "loss": 1.3518, + "theoretical_loss": 3.5665039315585556, + "tokens_seen": 1398407168 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029106318956870615, + "loss": 1.4195, + "theoretical_loss": 3.56648911529898, + "tokens_seen": 1398472704 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910531594784353, + "loss": 1.3577, + "theoretical_loss": 3.5664742999281174, + "tokens_seen": 1398538240 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910431293881645, + "loss": 1.4252, + "theoretical_loss": 3.566459485445873, + "tokens_seen": 1398603776 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910330992978937, + "loss": 1.3544, + "theoretical_loss": 3.5664446718521505, + "tokens_seen": 1398669312 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910230692076229, + "loss": 1.3274, + "theoretical_loss": 3.566429859146857, + "tokens_seen": 1398734848 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029101303911735206, + "loss": 1.4136, + "theoretical_loss": 3.566415047329896, + "tokens_seen": 1398800384 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002910030090270813, + "loss": 1.3552, + "theoretical_loss": 3.566400236401174, + "tokens_seen": 1398865920 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002909929789368104, + "loss": 1.4347, + "theoretical_loss": 3.5663854263605943, + "tokens_seen": 1398931456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029098294884653966, + "loss": 1.2971, + "theoretical_loss": 3.5663706172080643, + "tokens_seen": 1398996992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002909729187562688, + "loss": 1.4041, + "theoretical_loss": 3.566355808943487, + "tokens_seen": 1399062528 + }, + { + "epoch": 0.42, + "learning_rate": 0.000290962888665998, + "loss": 1.3257, + "theoretical_loss": 3.5663410015667694, + "tokens_seen": 1399128064 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 511124, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3820679187774658, + "objective/train/theoretical_loss": 3.5663261950778153, + "objective/train/tokens_used": 1419653600, + "theoretical_loss": 3.5663261950778153, + "tokens_seen": 1399193600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002909528585757272, + "loss": 1.415, + "theoretical_loss": 3.5663261950778153, + "tokens_seen": 1399193600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002909428284854564, + "loss": 1.4307, + "theoretical_loss": 3.56631138947653, + "tokens_seen": 1399259136 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029093279839518556, + "loss": 1.459, + "theoretical_loss": 3.56629658476282, + "tokens_seen": 1399324672 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029092276830491474, + "loss": 1.2483, + "theoretical_loss": 3.566281780936589, + "tokens_seen": 1399390208 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002909127382146439, + "loss": 1.413, + "theoretical_loss": 3.566266977997743, + "tokens_seen": 1399455744 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029090270812437316, + "loss": 1.3272, + "theoretical_loss": 3.5662521759461874, + "tokens_seen": 1399521280 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002908926780341023, + "loss": 1.3497, + "theoretical_loss": 3.566237374781827, + "tokens_seen": 1399586816 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002908826479438315, + "loss": 1.3797, + "theoretical_loss": 3.5662225745045677, + "tokens_seen": 1399652352 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029087261785356065, + "loss": 1.2916, + "theoretical_loss": 3.566207775114314, + "tokens_seen": 1399717888 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002908625877632899, + "loss": 1.3975, + "theoretical_loss": 3.5661929766109717, + "tokens_seen": 1399783424 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029085255767301906, + "loss": 1.382, + "theoretical_loss": 3.5661781789944462, + "tokens_seen": 1399848960 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029084252758274825, + "loss": 1.3831, + "theoretical_loss": 3.5661633822646426, + "tokens_seen": 1399914496 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002908324974924774, + "loss": 1.4119, + "theoretical_loss": 3.566148586421466, + "tokens_seen": 1399980032 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029082246740220666, + "loss": 1.2671, + "theoretical_loss": 3.5661337914648223, + "tokens_seen": 1400045568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002908124373119358, + "loss": 1.3529, + "theoretical_loss": 3.566118997394617, + "tokens_seen": 1400111104 + }, + { + "epoch": 0.42, + "learning_rate": 0.000290802407221665, + "loss": 1.4034, + "theoretical_loss": 3.566104204210755, + "tokens_seen": 1400176640 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029079237713139415, + "loss": 1.4142, + "theoretical_loss": 3.566089411913142, + "tokens_seen": 1400242176 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002907823470411234, + "loss": 1.4216, + "theoretical_loss": 3.5660746205016833, + "tokens_seen": 1400307712 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029077231695085257, + "loss": 1.392, + "theoretical_loss": 3.5660598299762847, + "tokens_seen": 1400373248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029076228686058175, + "loss": 1.4063, + "theoretical_loss": 3.566045040336851, + "tokens_seen": 1400438784 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029075225677031093, + "loss": 1.3589, + "theoretical_loss": 3.566030251583288, + "tokens_seen": 1400504320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002907422266800401, + "loss": 1.4013, + "theoretical_loss": 3.5660154637155013, + "tokens_seen": 1400569856 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002907321965897693, + "loss": 1.5361, + "theoretical_loss": 3.5660006767333963, + "tokens_seen": 1400635392 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029072216649949853, + "loss": 1.4303, + "theoretical_loss": 3.5659858906368784, + "tokens_seen": 1400700928 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029071213640922765, + "loss": 1.48, + "theoretical_loss": 3.5659711054258536, + "tokens_seen": 1400766464 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002907021063189569, + "loss": 1.5155, + "theoretical_loss": 3.5659563211002268, + "tokens_seen": 1400832000 + }, + { + "epoch": 0.42, + "learning_rate": 0.000290692076228686, + "loss": 1.3777, + "theoretical_loss": 3.5659415376599037, + "tokens_seen": 1400897536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029068204613841525, + "loss": 1.4098, + "theoretical_loss": 3.5659267551047904, + "tokens_seen": 1400963072 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029067201604814443, + "loss": 1.2927, + "theoretical_loss": 3.565911973434792, + "tokens_seen": 1401028608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002906619859578736, + "loss": 1.272, + "theoretical_loss": 3.565897192649815, + "tokens_seen": 1401094144 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002906519558676028, + "loss": 1.2751, + "theoretical_loss": 3.565882412749763, + "tokens_seen": 1401159680 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029064192577733203, + "loss": 1.3781, + "theoretical_loss": 3.5658676337345434, + "tokens_seen": 1401225216 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029063189568706116, + "loss": 1.3989, + "theoretical_loss": 3.5658528556040614, + "tokens_seen": 1401290752 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002906218655967904, + "loss": 1.278, + "theoretical_loss": 3.5658380783582224, + "tokens_seen": 1401356288 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002906118355065195, + "loss": 1.3013, + "theoretical_loss": 3.5658233019969328, + "tokens_seen": 1401421824 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029060180541624876, + "loss": 1.3435, + "theoretical_loss": 3.5658085265200974, + "tokens_seen": 1401487360 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029059177532597794, + "loss": 1.3866, + "theoretical_loss": 3.5657937519276226, + "tokens_seen": 1401552896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905817452357071, + "loss": 1.3938, + "theoretical_loss": 3.565778978219413, + "tokens_seen": 1401618432 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905717151454363, + "loss": 1.3308, + "theoretical_loss": 3.565764205395376, + "tokens_seen": 1401683968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905616850551655, + "loss": 1.3613, + "theoretical_loss": 3.565749433455416, + "tokens_seen": 1401749504 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029055165496489466, + "loss": 1.4127, + "theoretical_loss": 3.5657346623994393, + "tokens_seen": 1401815040 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905416248746239, + "loss": 1.4738, + "theoretical_loss": 3.565719892227351, + "tokens_seen": 1401880576 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905315947843531, + "loss": 1.4108, + "theoretical_loss": 3.5657051229390584, + "tokens_seen": 1401946112 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029052156469408226, + "loss": 1.4012, + "theoretical_loss": 3.565690354534466, + "tokens_seen": 1402011648 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905115346038115, + "loss": 1.3651, + "theoretical_loss": 3.56567558701348, + "tokens_seen": 1402077184 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002905015045135406, + "loss": 1.2578, + "theoretical_loss": 3.5656608203760065, + "tokens_seen": 1402142720 + }, + { + "epoch": 0.42, + "learning_rate": 0.00029049147442326986, + "loss": 1.3946, + "theoretical_loss": 3.5656460546219515, + "tokens_seen": 1402208256 + }, + { + "epoch": 0.42, + "learning_rate": 0.000290481444332999, + "loss": 1.3358, + "theoretical_loss": 3.5656312897512197, + "tokens_seen": 1402273792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002904714142427282, + "loss": 1.4316, + "theoretical_loss": 3.5656165257637182, + "tokens_seen": 1402339328 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002904613841524574, + "loss": 1.5521, + "theoretical_loss": 3.5656017626593526, + "tokens_seen": 1402404864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 512300, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5069409608840942, + "objective/train/theoretical_loss": 3.565587000438028, + "objective/train/tokens_used": 1422930400, + "theoretical_loss": 3.565587000438028, + "tokens_seen": 1402470400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0002904513540621866, + "loss": 1.504, + "theoretical_loss": 3.565587000438028, + "tokens_seen": 1402470400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029044132397191576, + "loss": 1.3785, + "theoretical_loss": 3.5655722390996516, + "tokens_seen": 1402535936 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029043129388164494, + "loss": 1.2906, + "theoretical_loss": 3.5655574786441284, + "tokens_seen": 1402601472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002904212637913741, + "loss": 1.4528, + "theoretical_loss": 3.565542719071365, + "tokens_seen": 1402667008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029041123370110336, + "loss": 1.4866, + "theoretical_loss": 3.5655279603812664, + "tokens_seen": 1402732544 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002904012036108325, + "loss": 1.2783, + "theoretical_loss": 3.5655132025737393, + "tokens_seen": 1402798080 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002903911735205617, + "loss": 1.4945, + "theoretical_loss": 3.5654984456486902, + "tokens_seen": 1402863616 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029038114343029085, + "loss": 1.2813, + "theoretical_loss": 3.5654836896060242, + "tokens_seen": 1402929152 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002903711133400201, + "loss": 1.3263, + "theoretical_loss": 3.5654689344456476, + "tokens_seen": 1402994688 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029036108324974926, + "loss": 1.3832, + "theoretical_loss": 3.5654541801674666, + "tokens_seen": 1403060224 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029035105315947845, + "loss": 1.3657, + "theoretical_loss": 3.5654394267713867, + "tokens_seen": 1403125760 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002903410230692076, + "loss": 1.3867, + "theoretical_loss": 3.565424674257315, + "tokens_seen": 1403191296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029033099297893686, + "loss": 1.3606, + "theoretical_loss": 3.5654099226251565, + "tokens_seen": 1403256832 + }, + { + "epoch": 0.43, + "learning_rate": 0.000290320962888666, + "loss": 1.3709, + "theoretical_loss": 3.565395171874818, + "tokens_seen": 1403322368 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002903109327983952, + "loss": 1.4167, + "theoretical_loss": 3.5653804220062053, + "tokens_seen": 1403387904 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029030090270812435, + "loss": 1.3161, + "theoretical_loss": 3.565365673019225, + "tokens_seen": 1403453440 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002902908726178536, + "loss": 1.4148, + "theoretical_loss": 3.5653509249137825, + "tokens_seen": 1403518976 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029028084252758277, + "loss": 1.3617, + "theoretical_loss": 3.565336177689784, + "tokens_seen": 1403584512 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029027081243731195, + "loss": 1.454, + "theoretical_loss": 3.5653214313471366, + "tokens_seen": 1403650048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029026078234704113, + "loss": 1.3105, + "theoretical_loss": 3.565306685885745, + "tokens_seen": 1403715584 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002902507522567703, + "loss": 1.4372, + "theoretical_loss": 3.565291941305517, + "tokens_seen": 1403781120 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002902407221664995, + "loss": 1.4061, + "theoretical_loss": 3.5652771976063575, + "tokens_seen": 1403846656 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029023069207622873, + "loss": 1.3857, + "theoretical_loss": 3.5652624547881735, + "tokens_seen": 1403912192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029022066198595785, + "loss": 1.409, + "theoretical_loss": 3.565247712850871, + "tokens_seen": 1403977728 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002902106318956871, + "loss": 1.3274, + "theoretical_loss": 3.565232971794356, + "tokens_seen": 1404043264 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002902006018054162, + "loss": 1.3268, + "theoretical_loss": 3.5652182316185357, + "tokens_seen": 1404108800 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029019057171514545, + "loss": 1.3404, + "theoretical_loss": 3.565203492323315, + "tokens_seen": 1404174336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029018054162487463, + "loss": 1.3666, + "theoretical_loss": 3.5651887539086013, + "tokens_seen": 1404239872 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002901705115346038, + "loss": 1.378, + "theoretical_loss": 3.5651740163743004, + "tokens_seen": 1404305408 + }, + { + "epoch": 0.43, + "learning_rate": 0.000290160481444333, + "loss": 1.4005, + "theoretical_loss": 3.5651592797203184, + "tokens_seen": 1404370944 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029015045135406223, + "loss": 1.439, + "theoretical_loss": 3.5651445439465625, + "tokens_seen": 1404436480 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029014042126379136, + "loss": 1.392, + "theoretical_loss": 3.565129809052938, + "tokens_seen": 1404502016 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002901303911735206, + "loss": 1.4289, + "theoretical_loss": 3.565115075039352, + "tokens_seen": 1404567552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002901203610832497, + "loss": 1.3876, + "theoretical_loss": 3.565100341905711, + "tokens_seen": 1404633088 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029011033099297896, + "loss": 1.349, + "theoretical_loss": 3.565085609651921, + "tokens_seen": 1404698624 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029010030090270814, + "loss": 1.3008, + "theoretical_loss": 3.565070878277888, + "tokens_seen": 1404764160 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900902708124373, + "loss": 1.4374, + "theoretical_loss": 3.5650561477835194, + "tokens_seen": 1404829696 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900802407221665, + "loss": 1.3874, + "theoretical_loss": 3.565041418168721, + "tokens_seen": 1404895232 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900702106318957, + "loss": 1.2967, + "theoretical_loss": 3.5650266894333993, + "tokens_seen": 1404960768 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029006018054162486, + "loss": 1.4636, + "theoretical_loss": 3.565011961577461, + "tokens_seen": 1405026304 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900501504513541, + "loss": 1.5166, + "theoretical_loss": 3.5649972346008125, + "tokens_seen": 1405091840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900401203610832, + "loss": 1.484, + "theoretical_loss": 3.56498250850336, + "tokens_seen": 1405157376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029003009027081246, + "loss": 1.3587, + "theoretical_loss": 3.5649677832850104, + "tokens_seen": 1405222912 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900200601805416, + "loss": 1.4467, + "theoretical_loss": 3.5649530589456697, + "tokens_seen": 1405288448 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002900100300902708, + "loss": 1.3821, + "theoretical_loss": 3.5649383354852455, + "tokens_seen": 1405353984 + }, + { + "epoch": 0.43, + "learning_rate": 0.00029, + "loss": 1.4047, + "theoretical_loss": 3.564923612903643, + "tokens_seen": 1405419520 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899899699097292, + "loss": 1.4158, + "theoretical_loss": 3.56490889120077, + "tokens_seen": 1405485056 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028997993981945836, + "loss": 1.3265, + "theoretical_loss": 3.5648941703765322, + "tokens_seen": 1405550592 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899699097291876, + "loss": 1.3196, + "theoretical_loss": 3.5648794504308365, + "tokens_seen": 1405616128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899598796389167, + "loss": 1.5509, + "theoretical_loss": 3.5648647313635893, + "tokens_seen": 1405681664 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 513528, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3427006006240845, + "objective/train/theoretical_loss": 3.5648500131746976, + "objective/train/tokens_used": 1426207200, + "theoretical_loss": 3.5648500131746976, + "tokens_seen": 1405747200 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028994984954864596, + "loss": 1.3885, + "theoretical_loss": 3.5648500131746976, + "tokens_seen": 1405747200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899398194583751, + "loss": 1.4605, + "theoretical_loss": 3.5648352958640688, + "tokens_seen": 1405812736 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899297893681043, + "loss": 1.2841, + "theoretical_loss": 3.5648205794316077, + "tokens_seen": 1405878272 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899197592778335, + "loss": 1.4241, + "theoretical_loss": 3.564805863877222, + "tokens_seen": 1405943808 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002899097291875627, + "loss": 1.4085, + "theoretical_loss": 3.5647911492008184, + "tokens_seen": 1406009344 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028989969909729187, + "loss": 1.4382, + "theoretical_loss": 3.5647764354023037, + "tokens_seen": 1406074880 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028988966900702105, + "loss": 1.4638, + "theoretical_loss": 3.5647617224815837, + "tokens_seen": 1406140416 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028987963891675023, + "loss": 1.2213, + "theoretical_loss": 3.5647470104385666, + "tokens_seen": 1406205952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028986960882647946, + "loss": 1.4899, + "theoretical_loss": 3.564732299273158, + "tokens_seen": 1406271488 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002898595787362086, + "loss": 1.4606, + "theoretical_loss": 3.564717588985265, + "tokens_seen": 1406337024 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002898495486459378, + "loss": 1.4072, + "theoretical_loss": 3.564702879574795, + "tokens_seen": 1406402560 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289839518555667, + "loss": 1.38, + "theoretical_loss": 3.564688171041653, + "tokens_seen": 1406468096 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002898294884653962, + "loss": 1.2816, + "theoretical_loss": 3.5646734633857475, + "tokens_seen": 1406533632 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028981945837512537, + "loss": 1.3247, + "theoretical_loss": 3.5646587566069856, + "tokens_seen": 1406599168 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028980942828485455, + "loss": 1.2519, + "theoretical_loss": 3.5646440507052723, + "tokens_seen": 1406664704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028979939819458373, + "loss": 1.3733, + "theoretical_loss": 3.5646293456805154, + "tokens_seen": 1406730240 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028978936810431297, + "loss": 1.2939, + "theoretical_loss": 3.5646146415326223, + "tokens_seen": 1406795776 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028977933801404215, + "loss": 1.4567, + "theoretical_loss": 3.564599938261499, + "tokens_seen": 1406861312 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028976930792377133, + "loss": 1.3128, + "theoretical_loss": 3.564585235867053, + "tokens_seen": 1406926848 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002897592778335005, + "loss": 1.3767, + "theoretical_loss": 3.5645705343491914, + "tokens_seen": 1406992384 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002897492477432297, + "loss": 1.4261, + "theoretical_loss": 3.5645558337078196, + "tokens_seen": 1407057920 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028973921765295893, + "loss": 1.3915, + "theoretical_loss": 3.5645411339428463, + "tokens_seen": 1407123456 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028972918756268805, + "loss": 1.3353, + "theoretical_loss": 3.5645264350541774, + "tokens_seen": 1407188992 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002897191574724173, + "loss": 1.4014, + "theoretical_loss": 3.56451173704172, + "tokens_seen": 1407254528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002897091273821464, + "loss": 1.4632, + "theoretical_loss": 3.5644970399053815, + "tokens_seen": 1407320064 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028969909729187565, + "loss": 1.3657, + "theoretical_loss": 3.5644823436450688, + "tokens_seen": 1407385600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028968906720160483, + "loss": 1.378, + "theoretical_loss": 3.5644676482606883, + "tokens_seen": 1407451136 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289679037111334, + "loss": 1.5262, + "theoretical_loss": 3.5644529537521477, + "tokens_seen": 1407516672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002896690070210632, + "loss": 1.2547, + "theoretical_loss": 3.5644382601193536, + "tokens_seen": 1407582208 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028965897693079243, + "loss": 1.4375, + "theoretical_loss": 3.564423567362213, + "tokens_seen": 1407647744 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028964894684052156, + "loss": 1.3551, + "theoretical_loss": 3.564408875480633, + "tokens_seen": 1407713280 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002896389167502508, + "loss": 1.3038, + "theoretical_loss": 3.564394184474521, + "tokens_seen": 1407778816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002896288866599799, + "loss": 1.3565, + "theoretical_loss": 3.5643794943437843, + "tokens_seen": 1407844352 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028961885656970916, + "loss": 1.38, + "theoretical_loss": 3.5643648050883288, + "tokens_seen": 1407909888 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028960882647943834, + "loss": 1.3954, + "theoretical_loss": 3.564350116708063, + "tokens_seen": 1407975424 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895987963891675, + "loss": 1.388, + "theoretical_loss": 3.5643354292028935, + "tokens_seen": 1408040960 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895887662988967, + "loss": 1.5195, + "theoretical_loss": 3.564320742572727, + "tokens_seen": 1408106496 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895787362086259, + "loss": 1.3483, + "theoretical_loss": 3.564306056817471, + "tokens_seen": 1408172032 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028956870611835506, + "loss": 1.4946, + "theoretical_loss": 3.564291371937032, + "tokens_seen": 1408237568 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895586760280843, + "loss": 1.347, + "theoretical_loss": 3.5642766879313186, + "tokens_seen": 1408303104 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895486459378134, + "loss": 1.474, + "theoretical_loss": 3.5642620048002374, + "tokens_seen": 1408368640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028953861584754266, + "loss": 1.4507, + "theoretical_loss": 3.5642473225436953, + "tokens_seen": 1408434176 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895285857572718, + "loss": 1.3455, + "theoretical_loss": 3.5642326411615994, + "tokens_seen": 1408499712 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289518555667001, + "loss": 1.4855, + "theoretical_loss": 3.564217960653857, + "tokens_seen": 1408565248 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002895085255767302, + "loss": 1.3234, + "theoretical_loss": 3.564203281020376, + "tokens_seen": 1408630784 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894984954864594, + "loss": 1.411, + "theoretical_loss": 3.564188602261063, + "tokens_seen": 1408696320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028948846539618856, + "loss": 1.3087, + "theoretical_loss": 3.564173924375826, + "tokens_seen": 1408761856 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894784353059178, + "loss": 1.4014, + "theoretical_loss": 3.5641592473645707, + "tokens_seen": 1408827392 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894684052156469, + "loss": 1.2822, + "theoretical_loss": 3.564144571227206, + "tokens_seen": 1408892928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028945837512537616, + "loss": 1.3566, + "theoretical_loss": 3.5641298959636387, + "tokens_seen": 1408958464 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 514827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5627686977386475, + "objective/train/theoretical_loss": 3.5641152215737764, + "objective/train/tokens_used": 1429484000, + "theoretical_loss": 3.5641152215737764, + "tokens_seen": 1409024000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894483450351053, + "loss": 1.4981, + "theoretical_loss": 3.5641152215737764, + "tokens_seen": 1409024000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894383149448345, + "loss": 1.3572, + "theoretical_loss": 3.5641005480575263, + "tokens_seen": 1409089536 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894282848545637, + "loss": 1.3865, + "theoretical_loss": 3.5640858754147953, + "tokens_seen": 1409155072 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002894182547642929, + "loss": 1.3294, + "theoretical_loss": 3.564071203645491, + "tokens_seen": 1409220608 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028940822467402207, + "loss": 1.3494, + "theoretical_loss": 3.564056532749521, + "tokens_seen": 1409286144 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028939819458375125, + "loss": 1.4089, + "theoretical_loss": 3.5640418627267927, + "tokens_seen": 1409351680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028938816449348043, + "loss": 1.4101, + "theoretical_loss": 3.5640271935772136, + "tokens_seen": 1409417216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028937813440320966, + "loss": 1.3149, + "theoretical_loss": 3.564012525300691, + "tokens_seen": 1409482752 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002893681043129388, + "loss": 1.4324, + "theoretical_loss": 3.563997857897132, + "tokens_seen": 1409548288 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289358074222668, + "loss": 1.3918, + "theoretical_loss": 3.563983191366445, + "tokens_seen": 1409613824 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002893480441323972, + "loss": 1.4201, + "theoretical_loss": 3.5639685257085363, + "tokens_seen": 1409679360 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002893380140421264, + "loss": 1.3184, + "theoretical_loss": 3.563953860923314, + "tokens_seen": 1409744896 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028932798395185557, + "loss": 1.556, + "theoretical_loss": 3.5639391970106855, + "tokens_seen": 1409810432 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028931795386158475, + "loss": 1.3611, + "theoretical_loss": 3.5639245339705585, + "tokens_seen": 1409875968 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028930792377131393, + "loss": 1.4256, + "theoretical_loss": 3.5639098718028404, + "tokens_seen": 1409941504 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028929789368104317, + "loss": 1.3205, + "theoretical_loss": 3.563895210507439, + "tokens_seen": 1410007040 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002892878635907723, + "loss": 1.4025, + "theoretical_loss": 3.563880550084262, + "tokens_seen": 1410072576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028927783350050153, + "loss": 1.3247, + "theoretical_loss": 3.563865890533216, + "tokens_seen": 1410138112 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028926780341023066, + "loss": 1.3764, + "theoretical_loss": 3.5638512318542093, + "tokens_seen": 1410203648 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002892577733199599, + "loss": 1.472, + "theoretical_loss": 3.563836574047149, + "tokens_seen": 1410269184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002892477432296891, + "loss": 1.3661, + "theoretical_loss": 3.5638219171119436, + "tokens_seen": 1410334720 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028923771313941825, + "loss": 1.3726, + "theoretical_loss": 3.5638072610485, + "tokens_seen": 1410400256 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028922768304914744, + "loss": 1.3067, + "theoretical_loss": 3.563792605856726, + "tokens_seen": 1410465792 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002892176529588766, + "loss": 1.3679, + "theoretical_loss": 3.5637779515365295, + "tokens_seen": 1410531328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002892076228686058, + "loss": 1.3912, + "theoretical_loss": 3.563763298087818, + "tokens_seen": 1410596864 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028919759277833503, + "loss": 1.5275, + "theoretical_loss": 3.5637486455104996, + "tokens_seen": 1410662400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028918756268806416, + "loss": 1.3968, + "theoretical_loss": 3.5637339938044814, + "tokens_seen": 1410727936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002891775325977934, + "loss": 1.4175, + "theoretical_loss": 3.563719342969671, + "tokens_seen": 1410793472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002891675025075226, + "loss": 1.3124, + "theoretical_loss": 3.5637046930059766, + "tokens_seen": 1410859008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028915747241725176, + "loss": 1.4472, + "theoretical_loss": 3.563690043913306, + "tokens_seen": 1410924544 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028914744232698094, + "loss": 1.4591, + "theoretical_loss": 3.5636753956915665, + "tokens_seen": 1410990080 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002891374122367101, + "loss": 1.3149, + "theoretical_loss": 3.5636607483406664, + "tokens_seen": 1411055616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002891273821464393, + "loss": 1.3751, + "theoretical_loss": 3.5636461018605132, + "tokens_seen": 1411121152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028911735205616854, + "loss": 1.4408, + "theoretical_loss": 3.5636314562510147, + "tokens_seen": 1411186688 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028910732196589766, + "loss": 1.381, + "theoretical_loss": 3.5636168115120785, + "tokens_seen": 1411252224 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890972918756269, + "loss": 1.5036, + "theoretical_loss": 3.5636021676436123, + "tokens_seen": 1411317760 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289087261785356, + "loss": 1.4677, + "theoretical_loss": 3.5635875246455253, + "tokens_seen": 1411383296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028907723169508526, + "loss": 1.4182, + "theoretical_loss": 3.5635728825177235, + "tokens_seen": 1411448832 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028906720160481444, + "loss": 1.391, + "theoretical_loss": 3.5635582412601163, + "tokens_seen": 1411514368 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890571715145436, + "loss": 1.4479, + "theoretical_loss": 3.5635436008726105, + "tokens_seen": 1411579904 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890471414242728, + "loss": 1.3876, + "theoretical_loss": 3.5635289613551144, + "tokens_seen": 1411645440 + }, + { + "epoch": 0.43, + "learning_rate": 0.000289037111334002, + "loss": 1.3934, + "theoretical_loss": 3.563514322707536, + "tokens_seen": 1411710976 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890270812437312, + "loss": 1.6366, + "theoretical_loss": 3.5634996849297833, + "tokens_seen": 1411776512 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890170511534604, + "loss": 1.3679, + "theoretical_loss": 3.563485048021764, + "tokens_seen": 1411842048 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002890070210631896, + "loss": 1.3605, + "theoretical_loss": 3.5634704119833858, + "tokens_seen": 1411907584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028899699097291876, + "loss": 1.3906, + "theoretical_loss": 3.563455776814558, + "tokens_seen": 1411973120 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288986960882648, + "loss": 1.4921, + "theoretical_loss": 3.5634411425151864, + "tokens_seen": 1412038656 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002889769307923771, + "loss": 1.4254, + "theoretical_loss": 3.563426509085181, + "tokens_seen": 1412104192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028896690070210636, + "loss": 1.3522, + "theoretical_loss": 3.5634118765244485, + "tokens_seen": 1412169728 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002889568706118355, + "loss": 1.2849, + "theoretical_loss": 3.5633972448328977, + "tokens_seen": 1412235264 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 516049, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4431415796279907, + "objective/train/theoretical_loss": 3.563382614010436, + "objective/train/tokens_used": 1432760800, + "theoretical_loss": 3.563382614010436, + "tokens_seen": 1412300800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002889468405215647, + "loss": 1.4038, + "theoretical_loss": 3.563382614010436, + "tokens_seen": 1412300800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002889368104312939, + "loss": 1.4376, + "theoretical_loss": 3.5633679840569723, + "tokens_seen": 1412366336 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002889267803410231, + "loss": 1.4188, + "theoretical_loss": 3.5633533549724143, + "tokens_seen": 1412431872 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028891675025075227, + "loss": 1.3909, + "theoretical_loss": 3.56333872675667, + "tokens_seen": 1412497408 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028890672016048145, + "loss": 1.371, + "theoretical_loss": 3.563324099409647, + "tokens_seen": 1412562944 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028889669007021063, + "loss": 1.3174, + "theoretical_loss": 3.5633094729312536, + "tokens_seen": 1412628480 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028888665997993986, + "loss": 1.4857, + "theoretical_loss": 3.563294847321399, + "tokens_seen": 1412694016 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288876629889669, + "loss": 1.4007, + "theoretical_loss": 3.5632802225799898, + "tokens_seen": 1412759552 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028886659979939823, + "loss": 1.4659, + "theoretical_loss": 3.563265598706936, + "tokens_seen": 1412825088 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002888565697091274, + "loss": 1.3646, + "theoretical_loss": 3.563250975702143, + "tokens_seen": 1412890624 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002888465396188566, + "loss": 1.4218, + "theoretical_loss": 3.5632363535655216, + "tokens_seen": 1412956160 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028883650952858577, + "loss": 1.3798, + "theoretical_loss": 3.563221732296979, + "tokens_seen": 1413021696 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028882647943831495, + "loss": 1.3977, + "theoretical_loss": 3.563207111896424, + "tokens_seen": 1413087232 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028881644934804413, + "loss": 1.4072, + "theoretical_loss": 3.5631924923637635, + "tokens_seen": 1413152768 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028880641925777337, + "loss": 1.2831, + "theoretical_loss": 3.5631778736989066, + "tokens_seen": 1413218304 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002887963891675025, + "loss": 1.4105, + "theoretical_loss": 3.5631632559017614, + "tokens_seen": 1413283840 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028878635907723173, + "loss": 1.249, + "theoretical_loss": 3.563148638972236, + "tokens_seen": 1413349376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028877632898696086, + "loss": 1.4346, + "theoretical_loss": 3.563134022910239, + "tokens_seen": 1413414912 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002887662988966901, + "loss": 1.3355, + "theoretical_loss": 3.5631194077156785, + "tokens_seen": 1413480448 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002887562688064193, + "loss": 1.3894, + "theoretical_loss": 3.5631047933884634, + "tokens_seen": 1413545984 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028874623871614845, + "loss": 1.3519, + "theoretical_loss": 3.563090179928501, + "tokens_seen": 1413611520 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028873620862587764, + "loss": 1.4405, + "theoretical_loss": 3.5630755673357, + "tokens_seen": 1413677056 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002887261785356068, + "loss": 1.4793, + "theoretical_loss": 3.563060955609969, + "tokens_seen": 1413742592 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288716148445336, + "loss": 1.4827, + "theoretical_loss": 3.5630463447512164, + "tokens_seen": 1413808128 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028870611835506523, + "loss": 1.4373, + "theoretical_loss": 3.5630317347593503, + "tokens_seen": 1413873664 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028869608826479436, + "loss": 1.4295, + "theoretical_loss": 3.5630171256342793, + "tokens_seen": 1413939200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002886860581745236, + "loss": 1.2916, + "theoretical_loss": 3.563002517375911, + "tokens_seen": 1414004736 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002886760280842528, + "loss": 1.4328, + "theoretical_loss": 3.5629879099841553, + "tokens_seen": 1414070272 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028866599799398196, + "loss": 1.3212, + "theoretical_loss": 3.5629733034589197, + "tokens_seen": 1414135808 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028865596790371114, + "loss": 1.4251, + "theoretical_loss": 3.5629586978001124, + "tokens_seen": 1414201344 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002886459378134403, + "loss": 1.4612, + "theoretical_loss": 3.5629440930076424, + "tokens_seen": 1414266880 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002886359077231695, + "loss": 1.491, + "theoretical_loss": 3.562929489081418, + "tokens_seen": 1414332416 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028862587763289874, + "loss": 1.3551, + "theoretical_loss": 3.562914886021348, + "tokens_seen": 1414397952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028861584754262786, + "loss": 1.3123, + "theoretical_loss": 3.5629002838273403, + "tokens_seen": 1414463488 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002886058174523571, + "loss": 1.383, + "theoretical_loss": 3.5628856824993034, + "tokens_seen": 1414529024 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002885957873620862, + "loss": 1.476, + "theoretical_loss": 3.5628710820371463, + "tokens_seen": 1414594560 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028858575727181546, + "loss": 1.4951, + "theoretical_loss": 3.5628564824407776, + "tokens_seen": 1414660096 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028857572718154464, + "loss": 1.4332, + "theoretical_loss": 3.5628418837101057, + "tokens_seen": 1414725632 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002885656970912738, + "loss": 1.3512, + "theoretical_loss": 3.5628272858450387, + "tokens_seen": 1414791168 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288555667001003, + "loss": 1.4377, + "theoretical_loss": 3.5628126888454856, + "tokens_seen": 1414856704 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002885456369107322, + "loss": 1.3553, + "theoretical_loss": 3.562798092711355, + "tokens_seen": 1414922240 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028853560682046137, + "loss": 1.3044, + "theoretical_loss": 3.5627834974425556, + "tokens_seen": 1414987776 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002885255767301906, + "loss": 1.5384, + "theoretical_loss": 3.5627689030389957, + "tokens_seen": 1415053312 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028851554663991973, + "loss": 1.3314, + "theoretical_loss": 3.5627543095005842, + "tokens_seen": 1415118848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028850551654964896, + "loss": 1.3966, + "theoretical_loss": 3.56273971682723, + "tokens_seen": 1415184384 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028849548645937815, + "loss": 1.4571, + "theoretical_loss": 3.562725125018841, + "tokens_seen": 1415249920 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002884854563691073, + "loss": 1.4602, + "theoretical_loss": 3.5627105340753267, + "tokens_seen": 1415315456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002884754262788365, + "loss": 1.3748, + "theoretical_loss": 3.5626959439965953, + "tokens_seen": 1415380992 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002884653961885657, + "loss": 1.3928, + "theoretical_loss": 3.5626813547825558, + "tokens_seen": 1415446528 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028845536609829487, + "loss": 1.4206, + "theoretical_loss": 3.562666766433116, + "tokens_seen": 1415512064 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 517252, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.445289134979248, + "objective/train/theoretical_loss": 3.5626521789481864, + "objective/train/tokens_used": 1436037600, + "theoretical_loss": 3.5626521789481864, + "tokens_seen": 1415577600 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002884453360080241, + "loss": 1.4326, + "theoretical_loss": 3.5626521789481864, + "tokens_seen": 1415577600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028843530591775323, + "loss": 1.2362, + "theoretical_loss": 3.5626375923276745, + "tokens_seen": 1415643136 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028842527582748247, + "loss": 1.2908, + "theoretical_loss": 3.562623006571489, + "tokens_seen": 1415708672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002884152457372116, + "loss": 1.3794, + "theoretical_loss": 3.5626084216795393, + "tokens_seen": 1415774208 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028840521564694083, + "loss": 1.3925, + "theoretical_loss": 3.5625938376517334, + "tokens_seen": 1415839744 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028839518555667, + "loss": 1.3164, + "theoretical_loss": 3.562579254487981, + "tokens_seen": 1415905280 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883851554663992, + "loss": 1.5236, + "theoretical_loss": 3.5625646721881905, + "tokens_seen": 1415970816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883751253761284, + "loss": 1.4421, + "theoretical_loss": 3.56255009075227, + "tokens_seen": 1416036352 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883650952858576, + "loss": 1.5012, + "theoretical_loss": 3.5625355101801297, + "tokens_seen": 1416101888 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028835506519558674, + "loss": 1.4053, + "theoretical_loss": 3.5625209304716776, + "tokens_seen": 1416167424 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028834503510531597, + "loss": 1.2958, + "theoretical_loss": 3.562506351626823, + "tokens_seen": 1416232960 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883350050150451, + "loss": 1.4992, + "theoretical_loss": 3.5624917736454744, + "tokens_seen": 1416298496 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028832497492477433, + "loss": 1.464, + "theoretical_loss": 3.5624771965275412, + "tokens_seen": 1416364032 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883149448345035, + "loss": 1.3848, + "theoretical_loss": 3.5624626202729317, + "tokens_seen": 1416429568 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002883049147442327, + "loss": 1.502, + "theoretical_loss": 3.5624480448815556, + "tokens_seen": 1416495104 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002882948846539619, + "loss": 1.3032, + "theoretical_loss": 3.562433470353321, + "tokens_seen": 1416560640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028828485456369106, + "loss": 1.3288, + "theoretical_loss": 3.562418896688137, + "tokens_seen": 1416626176 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002882748244734203, + "loss": 1.4728, + "theoretical_loss": 3.5624043238859135, + "tokens_seen": 1416691712 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002882647943831495, + "loss": 1.4551, + "theoretical_loss": 3.5623897519465584, + "tokens_seen": 1416757248 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028825476429287866, + "loss": 1.364, + "theoretical_loss": 3.5623751808699806, + "tokens_seen": 1416822784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028824473420260784, + "loss": 1.3967, + "theoretical_loss": 3.5623606106560906, + "tokens_seen": 1416888320 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288234704112337, + "loss": 1.2313, + "theoretical_loss": 3.562346041304796, + "tokens_seen": 1416953856 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002882246740220662, + "loss": 1.4189, + "theoretical_loss": 3.562331472816006, + "tokens_seen": 1417019392 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028821464393179543, + "loss": 1.304, + "theoretical_loss": 3.56231690518963, + "tokens_seen": 1417084928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028820461384152456, + "loss": 1.3839, + "theoretical_loss": 3.562302338425577, + "tokens_seen": 1417150464 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002881945837512538, + "loss": 1.4278, + "theoretical_loss": 3.5622877725237565, + "tokens_seen": 1417216000 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288184553660983, + "loss": 1.292, + "theoretical_loss": 3.5622732074840764, + "tokens_seen": 1417281536 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028817452357071216, + "loss": 1.5788, + "theoretical_loss": 3.562258643306447, + "tokens_seen": 1417347072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028816449348044134, + "loss": 1.4199, + "theoretical_loss": 3.562244079990777, + "tokens_seen": 1417412608 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002881544633901705, + "loss": 1.3381, + "theoretical_loss": 3.5622295175369754, + "tokens_seen": 1417478144 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002881444332998997, + "loss": 1.4177, + "theoretical_loss": 3.5622149559449516, + "tokens_seen": 1417543680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028813440320962894, + "loss": 1.3501, + "theoretical_loss": 3.5622003952146146, + "tokens_seen": 1417609216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028812437311935806, + "loss": 1.4547, + "theoretical_loss": 3.5621858353458737, + "tokens_seen": 1417674752 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002881143430290873, + "loss": 1.3887, + "theoretical_loss": 3.5621712763386375, + "tokens_seen": 1417740288 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002881043129388164, + "loss": 1.4621, + "theoretical_loss": 3.562156718192816, + "tokens_seen": 1417805824 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028809428284854566, + "loss": 1.3401, + "theoretical_loss": 3.5621421609083184, + "tokens_seen": 1417871360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028808425275827484, + "loss": 1.2462, + "theoretical_loss": 3.5621276044850534, + "tokens_seen": 1417936896 + }, + { + "epoch": 0.43, + "learning_rate": 0.000288074222668004, + "loss": 1.3062, + "theoretical_loss": 3.5621130489229307, + "tokens_seen": 1418002432 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002880641925777332, + "loss": 1.4003, + "theoretical_loss": 3.5620984942218596, + "tokens_seen": 1418067968 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002880541624874624, + "loss": 1.3978, + "theoretical_loss": 3.562083940381749, + "tokens_seen": 1418133504 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028804413239719157, + "loss": 1.2981, + "theoretical_loss": 3.5620693874025076, + "tokens_seen": 1418199040 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002880341023069208, + "loss": 1.377, + "theoretical_loss": 3.562054835284046, + "tokens_seen": 1418264576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028802407221664993, + "loss": 1.2934, + "theoretical_loss": 3.5620402840262733, + "tokens_seen": 1418330112 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028801404212637916, + "loss": 1.3644, + "theoretical_loss": 3.5620257336290977, + "tokens_seen": 1418395648 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028800401203610835, + "loss": 1.4256, + "theoretical_loss": 3.5620111840924302, + "tokens_seen": 1418461184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879939819458375, + "loss": 1.4524, + "theoretical_loss": 3.5619966354161785, + "tokens_seen": 1418526720 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879839518555667, + "loss": 1.4006, + "theoretical_loss": 3.561982087600253, + "tokens_seen": 1418592256 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879739217652959, + "loss": 1.3434, + "theoretical_loss": 3.561967540644563, + "tokens_seen": 1418657792 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028796389167502507, + "loss": 1.3879, + "theoretical_loss": 3.561952994549017, + "tokens_seen": 1418723328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879538615847543, + "loss": 1.4378, + "theoretical_loss": 3.5619384493135264, + "tokens_seen": 1418788864 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 518371, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3722832202911377, + "objective/train/theoretical_loss": 3.5619239049379985, + "objective/train/tokens_used": 1439314400, + "theoretical_loss": 3.5619239049379985, + "tokens_seen": 1418854400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028794383149448343, + "loss": 1.3787, + "theoretical_loss": 3.5619239049379985, + "tokens_seen": 1418854400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028793380140421267, + "loss": 1.3417, + "theoretical_loss": 3.5619093614223436, + "tokens_seen": 1418919936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879237713139418, + "loss": 1.533, + "theoretical_loss": 3.561894818766471, + "tokens_seen": 1418985472 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028791374122367103, + "loss": 1.3738, + "theoretical_loss": 3.5618802769702906, + "tokens_seen": 1419051008 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002879037111334002, + "loss": 1.4338, + "theoretical_loss": 3.561865736033712, + "tokens_seen": 1419116544 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878936810431294, + "loss": 1.3974, + "theoretical_loss": 3.5618511959566437, + "tokens_seen": 1419182080 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878836509528586, + "loss": 1.4338, + "theoretical_loss": 3.561836656738996, + "tokens_seen": 1419247616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878736208625878, + "loss": 1.4349, + "theoretical_loss": 3.5618221183806784, + "tokens_seen": 1419313152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028786359077231694, + "loss": 1.3112, + "theoretical_loss": 3.5618075808815997, + "tokens_seen": 1419378688 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028785356068204617, + "loss": 1.3864, + "theoretical_loss": 3.5617930442416705, + "tokens_seen": 1419444224 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878435305917753, + "loss": 1.4028, + "theoretical_loss": 3.5617785084608, + "tokens_seen": 1419509760 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028783350050150453, + "loss": 1.4491, + "theoretical_loss": 3.5617639735388975, + "tokens_seen": 1419575296 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878234704112337, + "loss": 1.3396, + "theoretical_loss": 3.5617494394758724, + "tokens_seen": 1419640832 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878134403209629, + "loss": 1.5099, + "theoretical_loss": 3.561734906271635, + "tokens_seen": 1419706368 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002878034102306921, + "loss": 1.4125, + "theoretical_loss": 3.561720373926094, + "tokens_seen": 1419771904 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028779338014042126, + "loss": 1.4419, + "theoretical_loss": 3.56170584243916, + "tokens_seen": 1419837440 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028778335005015044, + "loss": 1.3963, + "theoretical_loss": 3.5616913118107423, + "tokens_seen": 1419902976 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002877733199598797, + "loss": 1.3468, + "theoretical_loss": 3.56167678204075, + "tokens_seen": 1419968512 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002877632898696088, + "loss": 1.3381, + "theoretical_loss": 3.561662253129094, + "tokens_seen": 1420034048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028775325977933804, + "loss": 1.4165, + "theoretical_loss": 3.5616477250756824, + "tokens_seen": 1420099584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028774322968906716, + "loss": 1.351, + "theoretical_loss": 3.5616331978804263, + "tokens_seen": 1420165120 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002877331995987964, + "loss": 1.3453, + "theoretical_loss": 3.5616186715432345, + "tokens_seen": 1420230656 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002877231695085256, + "loss": 1.4195, + "theoretical_loss": 3.561604146064017, + "tokens_seen": 1420296192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028771313941825476, + "loss": 1.3893, + "theoretical_loss": 3.5615896214426837, + "tokens_seen": 1420361728 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028770310932798394, + "loss": 1.3313, + "theoretical_loss": 3.5615750976791447, + "tokens_seen": 1420427264 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876930792377132, + "loss": 1.3749, + "theoretical_loss": 3.561560574773309, + "tokens_seen": 1420492800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876830491474423, + "loss": 1.3686, + "theoretical_loss": 3.5615460527250864, + "tokens_seen": 1420558336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028767301905717154, + "loss": 1.4079, + "theoretical_loss": 3.5615315315343876, + "tokens_seen": 1420623872 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028766298896690067, + "loss": 1.4498, + "theoretical_loss": 3.561517011201121, + "tokens_seen": 1420689408 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876529588766299, + "loss": 1.3925, + "theoretical_loss": 3.5615024917251983, + "tokens_seen": 1420754944 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876429287863591, + "loss": 1.388, + "theoretical_loss": 3.5614879731065274, + "tokens_seen": 1420820480 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028763289869608826, + "loss": 1.4224, + "theoretical_loss": 3.56147345534502, + "tokens_seen": 1420886016 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028762286860581745, + "loss": 1.3851, + "theoretical_loss": 3.561458938440584, + "tokens_seen": 1420951552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876128385155466, + "loss": 1.4137, + "theoretical_loss": 3.5614444223931305, + "tokens_seen": 1421017088 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002876028084252758, + "loss": 1.3986, + "theoretical_loss": 3.561429907202569, + "tokens_seen": 1421082624 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028759277833500504, + "loss": 1.3112, + "theoretical_loss": 3.5614153928688097, + "tokens_seen": 1421148160 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028758274824473417, + "loss": 1.395, + "theoretical_loss": 3.561400879391763, + "tokens_seen": 1421213696 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002875727181544634, + "loss": 1.4203, + "theoretical_loss": 3.5613863667713375, + "tokens_seen": 1421279232 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028756268806419253, + "loss": 1.2949, + "theoretical_loss": 3.561371855007444, + "tokens_seen": 1421344768 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028755265797392177, + "loss": 1.4314, + "theoretical_loss": 3.561357344099992, + "tokens_seen": 1421410304 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028754262788365095, + "loss": 1.4741, + "theoretical_loss": 3.561342834048892, + "tokens_seen": 1421475840 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028753259779338013, + "loss": 1.3676, + "theoretical_loss": 3.5613283248540535, + "tokens_seen": 1421541376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028752256770310936, + "loss": 1.2762, + "theoretical_loss": 3.561313816515387, + "tokens_seen": 1421606912 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028751253761283855, + "loss": 1.3192, + "theoretical_loss": 3.5612993090328025, + "tokens_seen": 1421672448 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002875025075225677, + "loss": 1.3516, + "theoretical_loss": 3.5612848024062096, + "tokens_seen": 1421737984 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002874924774322969, + "loss": 1.3764, + "theoretical_loss": 3.5612702966355183, + "tokens_seen": 1421803520 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002874824473420261, + "loss": 1.4862, + "theoretical_loss": 3.5612557917206393, + "tokens_seen": 1421869056 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028747241725175527, + "loss": 1.4508, + "theoretical_loss": 3.561241287661482, + "tokens_seen": 1421934592 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002874623871614845, + "loss": 1.3382, + "theoretical_loss": 3.561226784457957, + "tokens_seen": 1422000128 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028745235707121363, + "loss": 1.39, + "theoretical_loss": 3.561212282109974, + "tokens_seen": 1422065664 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 519513, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5280126333236694, + "objective/train/theoretical_loss": 3.5611977806174426, + "objective/train/tokens_used": 1442591200, + "theoretical_loss": 3.5611977806174426, + "tokens_seen": 1422131200 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028744232698094287, + "loss": 1.4709, + "theoretical_loss": 3.5611977806174426, + "tokens_seen": 1422131200 + }, + { + "epoch": 0.43, + "learning_rate": 0.000287432296890672, + "loss": 1.3869, + "theoretical_loss": 3.5611832799802743, + "tokens_seen": 1422196736 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028742226680040123, + "loss": 1.3866, + "theoretical_loss": 3.5611687801983782, + "tokens_seen": 1422262272 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002874122367101304, + "loss": 1.3181, + "theoretical_loss": 3.561154281271665, + "tokens_seen": 1422327808 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002874022066198596, + "loss": 1.4743, + "theoretical_loss": 3.561139783200044, + "tokens_seen": 1422393344 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002873921765295888, + "loss": 1.3544, + "theoretical_loss": 3.5611252859834264, + "tokens_seen": 1422458880 + }, + { + "epoch": 0.43, + "learning_rate": 0.000287382146439318, + "loss": 1.3338, + "theoretical_loss": 3.561110789621722, + "tokens_seen": 1422524416 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028737211634904714, + "loss": 1.3364, + "theoretical_loss": 3.561096294114841, + "tokens_seen": 1422589952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028736208625877637, + "loss": 1.4429, + "theoretical_loss": 3.561081799462693, + "tokens_seen": 1422655488 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002873520561685055, + "loss": 1.4543, + "theoretical_loss": 3.5610673056651896, + "tokens_seen": 1422721024 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028734202607823473, + "loss": 1.3045, + "theoretical_loss": 3.56105281272224, + "tokens_seen": 1422786560 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002873319959879639, + "loss": 1.3115, + "theoretical_loss": 3.561038320633754, + "tokens_seen": 1422852096 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002873219658976931, + "loss": 1.3351, + "theoretical_loss": 3.5610238293996437, + "tokens_seen": 1422917632 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002873119358074223, + "loss": 1.3535, + "theoretical_loss": 3.561009339019818, + "tokens_seen": 1422983168 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028730190571715146, + "loss": 1.347, + "theoretical_loss": 3.560994849494187, + "tokens_seen": 1423048704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028729187562688064, + "loss": 1.3873, + "theoretical_loss": 3.5609803608226613, + "tokens_seen": 1423114240 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002872818455366099, + "loss": 1.3653, + "theoretical_loss": 3.560965873005152, + "tokens_seen": 1423179776 + }, + { + "epoch": 0.43, + "learning_rate": 0.000287271815446339, + "loss": 1.4181, + "theoretical_loss": 3.5609513860415687, + "tokens_seen": 1423245312 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028726178535606824, + "loss": 1.2882, + "theoretical_loss": 3.5609368999318214, + "tokens_seen": 1423310848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028725175526579736, + "loss": 1.4252, + "theoretical_loss": 3.5609224146758214, + "tokens_seen": 1423376384 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002872417251755266, + "loss": 1.3285, + "theoretical_loss": 3.560907930273479, + "tokens_seen": 1423441920 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002872316950852558, + "loss": 1.3738, + "theoretical_loss": 3.5608934467247035, + "tokens_seen": 1423507456 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028722166499498496, + "loss": 1.3092, + "theoretical_loss": 3.5608789640294063, + "tokens_seen": 1423572992 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028721163490471414, + "loss": 1.494, + "theoretical_loss": 3.5608644821874975, + "tokens_seen": 1423638528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002872016048144434, + "loss": 1.3029, + "theoretical_loss": 3.5608500011988875, + "tokens_seen": 1423704064 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002871915747241725, + "loss": 1.3994, + "theoretical_loss": 3.5608355210634874, + "tokens_seen": 1423769600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028718154463390174, + "loss": 1.2388, + "theoretical_loss": 3.5608210417812067, + "tokens_seen": 1423835136 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028717151454363087, + "loss": 1.3904, + "theoretical_loss": 3.560806563351956, + "tokens_seen": 1423900672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002871614844533601, + "loss": 1.4264, + "theoretical_loss": 3.5607920857756463, + "tokens_seen": 1423966208 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002871514543630893, + "loss": 1.4877, + "theoretical_loss": 3.5607776090521877, + "tokens_seen": 1424031744 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028714142427281846, + "loss": 1.3947, + "theoretical_loss": 3.5607631331814913, + "tokens_seen": 1424097280 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028713139418254765, + "loss": 1.3398, + "theoretical_loss": 3.5607486581634666, + "tokens_seen": 1424162816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002871213640922768, + "loss": 1.3393, + "theoretical_loss": 3.560734183998025, + "tokens_seen": 1424228352 + }, + { + "epoch": 0.43, + "learning_rate": 0.000287111334002006, + "loss": 1.3338, + "theoretical_loss": 3.5607197106850768, + "tokens_seen": 1424293888 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028710130391173524, + "loss": 1.4461, + "theoretical_loss": 3.560705238224532, + "tokens_seen": 1424359424 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028709127382146437, + "loss": 1.4373, + "theoretical_loss": 3.560690766616303, + "tokens_seen": 1424424960 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002870812437311936, + "loss": 1.4833, + "theoretical_loss": 3.560676295860298, + "tokens_seen": 1424490496 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028707121364092273, + "loss": 1.3855, + "theoretical_loss": 3.560661825956429, + "tokens_seen": 1424556032 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028706118355065197, + "loss": 1.3783, + "theoretical_loss": 3.5606473569046067, + "tokens_seen": 1424621568 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028705115346038115, + "loss": 1.329, + "theoretical_loss": 3.560632888704741, + "tokens_seen": 1424687104 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028704112337011033, + "loss": 1.3411, + "theoretical_loss": 3.560618421356743, + "tokens_seen": 1424752640 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002870310932798395, + "loss": 1.3387, + "theoretical_loss": 3.5606039548605235, + "tokens_seen": 1424818176 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028702106318956875, + "loss": 1.362, + "theoretical_loss": 3.560589489215993, + "tokens_seen": 1424883712 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002870110330992979, + "loss": 1.3323, + "theoretical_loss": 3.560575024423062, + "tokens_seen": 1424949248 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002870010030090271, + "loss": 1.4729, + "theoretical_loss": 3.5605605604816413, + "tokens_seen": 1425014784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028699097291875624, + "loss": 1.3985, + "theoretical_loss": 3.560546097391642, + "tokens_seen": 1425080320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028698094282848547, + "loss": 1.3474, + "theoretical_loss": 3.560531635152974, + "tokens_seen": 1425145856 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028697091273821465, + "loss": 1.4094, + "theoretical_loss": 3.5605171737655485, + "tokens_seen": 1425211392 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028696088264794383, + "loss": 1.3796, + "theoretical_loss": 3.560502713229277, + "tokens_seen": 1425276928 + }, + { + "epoch": 0.43, + "learning_rate": 0.000286950852557673, + "loss": 1.4853, + "theoretical_loss": 3.5604882535440696, + "tokens_seen": 1425342464 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 520731, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2737388610839844, + "objective/train/theoretical_loss": 3.5604737947098366, + "objective/train/tokens_used": 1445868000, + "theoretical_loss": 3.5604737947098366, + "tokens_seen": 1425408000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002869408224674022, + "loss": 1.2786, + "theoretical_loss": 3.5604737947098366, + "tokens_seen": 1425408000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002869307923771314, + "loss": 1.3647, + "theoretical_loss": 3.560459336726489, + "tokens_seen": 1425473536 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002869207622868606, + "loss": 1.4497, + "theoretical_loss": 3.5604448795939385, + "tokens_seen": 1425539072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028691073219658974, + "loss": 1.3701, + "theoretical_loss": 3.560430423312095, + "tokens_seen": 1425604608 + }, + { + "epoch": 0.43, + "learning_rate": 0.000286900702106319, + "loss": 1.3858, + "theoretical_loss": 3.56041596788087, + "tokens_seen": 1425670144 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868906720160481, + "loss": 1.3076, + "theoretical_loss": 3.5604015133001736, + "tokens_seen": 1425735680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028688064192577734, + "loss": 1.4474, + "theoretical_loss": 3.5603870595699174, + "tokens_seen": 1425801216 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868706118355065, + "loss": 1.3495, + "theoretical_loss": 3.560372606690011, + "tokens_seen": 1425866752 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868605817452357, + "loss": 1.3589, + "theoretical_loss": 3.560358154660367, + "tokens_seen": 1425932288 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868505516549649, + "loss": 1.4209, + "theoretical_loss": 3.560343703480896, + "tokens_seen": 1425997824 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868405215646941, + "loss": 1.3033, + "theoretical_loss": 3.560329253151508, + "tokens_seen": 1426063360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028683049147442324, + "loss": 1.4289, + "theoretical_loss": 3.560314803672114, + "tokens_seen": 1426128896 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868204613841525, + "loss": 1.3047, + "theoretical_loss": 3.5603003550426258, + "tokens_seen": 1426194432 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002868104312938816, + "loss": 1.2975, + "theoretical_loss": 3.560285907262953, + "tokens_seen": 1426259968 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028680040120361084, + "loss": 1.2944, + "theoretical_loss": 3.5602714603330083, + "tokens_seen": 1426325504 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028679037111334, + "loss": 1.3821, + "theoretical_loss": 3.560257014252702, + "tokens_seen": 1426391040 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002867803410230692, + "loss": 1.3248, + "theoretical_loss": 3.560242569021945, + "tokens_seen": 1426456576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028677031093279844, + "loss": 1.4074, + "theoretical_loss": 3.5602281246406475, + "tokens_seen": 1426522112 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028676028084252756, + "loss": 1.4065, + "theoretical_loss": 3.5602136811087215, + "tokens_seen": 1426587648 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002867502507522568, + "loss": 1.2801, + "theoretical_loss": 3.560199238426078, + "tokens_seen": 1426653184 + }, + { + "epoch": 0.43, + "learning_rate": 0.000286740220661986, + "loss": 1.368, + "theoretical_loss": 3.560184796592628, + "tokens_seen": 1426718720 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028673019057171516, + "loss": 1.4346, + "theoretical_loss": 3.5601703556082818, + "tokens_seen": 1426784256 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028672016048144434, + "loss": 1.4255, + "theoretical_loss": 3.5601559154729516, + "tokens_seen": 1426849792 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002867101303911736, + "loss": 1.3792, + "theoretical_loss": 3.5601414761865477, + "tokens_seen": 1426915328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002867001003009027, + "loss": 1.3301, + "theoretical_loss": 3.560127037748982, + "tokens_seen": 1426980864 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028669007021063194, + "loss": 1.4247, + "theoretical_loss": 3.560112600160165, + "tokens_seen": 1427046400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028668004012036107, + "loss": 1.4232, + "theoretical_loss": 3.5600981634200073, + "tokens_seen": 1427111936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002866700100300903, + "loss": 1.3991, + "theoretical_loss": 3.5600837275284216, + "tokens_seen": 1427177472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002866599799398195, + "loss": 1.443, + "theoretical_loss": 3.5600692924853177, + "tokens_seen": 1427243008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028664994984954866, + "loss": 1.3303, + "theoretical_loss": 3.5600548582906075, + "tokens_seen": 1427308544 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028663991975927785, + "loss": 1.4593, + "theoretical_loss": 3.5600404249442015, + "tokens_seen": 1427374080 + }, + { + "epoch": 0.43, + "learning_rate": 0.000286629889669007, + "loss": 1.5004, + "theoretical_loss": 3.5600259924460116, + "tokens_seen": 1427439616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002866198595787362, + "loss": 1.54, + "theoretical_loss": 3.5600115607959486, + "tokens_seen": 1427505152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028660982948846544, + "loss": 1.5189, + "theoretical_loss": 3.559997129993924, + "tokens_seen": 1427570688 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028659979939819457, + "loss": 1.3354, + "theoretical_loss": 3.559982700039849, + "tokens_seen": 1427636224 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002865897693079238, + "loss": 1.3401, + "theoretical_loss": 3.559968270933634, + "tokens_seen": 1427701760 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028657973921765293, + "loss": 1.3441, + "theoretical_loss": 3.559953842675192, + "tokens_seen": 1427767296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028656970912738217, + "loss": 1.416, + "theoretical_loss": 3.5599394152644326, + "tokens_seen": 1427832832 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028655967903711135, + "loss": 1.3638, + "theoretical_loss": 3.559924988701268, + "tokens_seen": 1427898368 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028654964894684053, + "loss": 1.4043, + "theoretical_loss": 3.5599105629856096, + "tokens_seen": 1427963904 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002865396188565697, + "loss": 1.2897, + "theoretical_loss": 3.5598961381173684, + "tokens_seen": 1428029440 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028652958876629895, + "loss": 1.4495, + "theoretical_loss": 3.5598817140964556, + "tokens_seen": 1428094976 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002865195586760281, + "loss": 1.3472, + "theoretical_loss": 3.5598672909227824, + "tokens_seen": 1428160512 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002865095285857573, + "loss": 1.5441, + "theoretical_loss": 3.559852868596261, + "tokens_seen": 1428226048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028649949849548644, + "loss": 1.3728, + "theoretical_loss": 3.5598384471168014, + "tokens_seen": 1428291584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028648946840521567, + "loss": 1.3818, + "theoretical_loss": 3.559824026484317, + "tokens_seen": 1428357120 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028647943831494485, + "loss": 1.3469, + "theoretical_loss": 3.5598096066987166, + "tokens_seen": 1428422656 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028646940822467403, + "loss": 1.3819, + "theoretical_loss": 3.5597951877599137, + "tokens_seen": 1428488192 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002864593781344032, + "loss": 1.4497, + "theoretical_loss": 3.559780769667819, + "tokens_seen": 1428553728 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002864493480441324, + "loss": 1.4843, + "theoretical_loss": 3.5597663524223444, + "tokens_seen": 1428619264 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 521864, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3886792659759521, + "objective/train/theoretical_loss": 3.559751936023401, + "objective/train/tokens_used": 1449144800, + "theoretical_loss": 3.559751936023401, + "tokens_seen": 1428684800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002864393179538616, + "loss": 1.4029, + "theoretical_loss": 3.559751936023401, + "tokens_seen": 1428684800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002864292878635908, + "loss": 1.4341, + "theoretical_loss": 3.5597375204709, + "tokens_seen": 1428750336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028641925777331994, + "loss": 1.4165, + "theoretical_loss": 3.5597231057647525, + "tokens_seen": 1428815872 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002864092276830492, + "loss": 1.4974, + "theoretical_loss": 3.5597086919048713, + "tokens_seen": 1428881408 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863991975927783, + "loss": 1.5366, + "theoretical_loss": 3.5596942788911665, + "tokens_seen": 1428946944 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028638916750250754, + "loss": 1.4708, + "theoretical_loss": 3.5596798667235507, + "tokens_seen": 1429012480 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863791374122367, + "loss": 1.3165, + "theoretical_loss": 3.5596654554019356, + "tokens_seen": 1429078016 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863691073219659, + "loss": 1.4662, + "theoretical_loss": 3.5596510449262313, + "tokens_seen": 1429143552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863590772316951, + "loss": 1.3778, + "theoretical_loss": 3.559636635296351, + "tokens_seen": 1429209088 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863490471414243, + "loss": 1.3539, + "theoretical_loss": 3.559622226512205, + "tokens_seen": 1429274624 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028633901705115344, + "loss": 1.2751, + "theoretical_loss": 3.5596078185737055, + "tokens_seen": 1429340160 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863289869608827, + "loss": 1.397, + "theoretical_loss": 3.559593411480764, + "tokens_seen": 1429405696 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002863189568706118, + "loss": 1.438, + "theoretical_loss": 3.5595790052332923, + "tokens_seen": 1429471232 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028630892678034104, + "loss": 1.4283, + "theoretical_loss": 3.5595645998312015, + "tokens_seen": 1429536768 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862988966900702, + "loss": 1.3929, + "theoretical_loss": 3.559550195274404, + "tokens_seen": 1429602304 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862888665997994, + "loss": 1.3671, + "theoretical_loss": 3.5595357915628107, + "tokens_seen": 1429667840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862788365095286, + "loss": 1.3343, + "theoretical_loss": 3.559521388696334, + "tokens_seen": 1429733376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028626880641925776, + "loss": 1.4291, + "theoretical_loss": 3.5595069866748847, + "tokens_seen": 1429798912 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028625877632898694, + "loss": 1.356, + "theoretical_loss": 3.559492585498375, + "tokens_seen": 1429864448 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862487462387162, + "loss": 1.4223, + "theoretical_loss": 3.559478185166717, + "tokens_seen": 1429929984 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862387161484453, + "loss": 1.3519, + "theoretical_loss": 3.5594637856798217, + "tokens_seen": 1429995520 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028622868605817454, + "loss": 1.2849, + "theoretical_loss": 3.5594493870376014, + "tokens_seen": 1430061056 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028621865596790367, + "loss": 1.4222, + "theoretical_loss": 3.5594349892399673, + "tokens_seen": 1430126592 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002862086258776329, + "loss": 1.4351, + "theoretical_loss": 3.5594205922868314, + "tokens_seen": 1430192128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002861985957873621, + "loss": 1.3549, + "theoretical_loss": 3.559406196178106, + "tokens_seen": 1430257664 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028618856569709127, + "loss": 1.3576, + "theoretical_loss": 3.5593918009137018, + "tokens_seen": 1430323200 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028617853560682045, + "loss": 1.3173, + "theoretical_loss": 3.5593774064935317, + "tokens_seen": 1430388736 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002861685055165497, + "loss": 1.4067, + "theoretical_loss": 3.5593630129175065, + "tokens_seen": 1430454272 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002861584754262788, + "loss": 1.4516, + "theoretical_loss": 3.559348620185539, + "tokens_seen": 1430519808 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028614844533600805, + "loss": 1.3488, + "theoretical_loss": 3.55933422829754, + "tokens_seen": 1430585344 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028613841524573717, + "loss": 1.4985, + "theoretical_loss": 3.5593198372534225, + "tokens_seen": 1430650880 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002861283851554664, + "loss": 1.4512, + "theoretical_loss": 3.5593054470530974, + "tokens_seen": 1430716416 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002861183550651956, + "loss": 1.3142, + "theoretical_loss": 3.559291057696477, + "tokens_seen": 1430781952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028610832497492477, + "loss": 1.2828, + "theoretical_loss": 3.559276669183473, + "tokens_seen": 1430847488 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028609829488465395, + "loss": 1.4084, + "theoretical_loss": 3.559262281513998, + "tokens_seen": 1430913024 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028608826479438313, + "loss": 1.3921, + "theoretical_loss": 3.559247894687963, + "tokens_seen": 1430978560 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860782347041123, + "loss": 1.4636, + "theoretical_loss": 3.5592335087052804, + "tokens_seen": 1431044096 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028606820461384155, + "loss": 1.375, + "theoretical_loss": 3.559219123565862, + "tokens_seen": 1431109632 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860581745235707, + "loss": 1.4489, + "theoretical_loss": 3.55920473926962, + "tokens_seen": 1431175168 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860481444332999, + "loss": 1.4039, + "theoretical_loss": 3.559190355816466, + "tokens_seen": 1431240704 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860381143430291, + "loss": 1.4117, + "theoretical_loss": 3.5591759732063117, + "tokens_seen": 1431306240 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860280842527583, + "loss": 1.3911, + "theoretical_loss": 3.5591615914390706, + "tokens_seen": 1431371776 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002860180541624875, + "loss": 1.2825, + "theoretical_loss": 3.559147210514653, + "tokens_seen": 1431437312 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028600802407221664, + "loss": 1.3889, + "theoretical_loss": 3.5591328304329712, + "tokens_seen": 1431502848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028599799398194587, + "loss": 1.3661, + "theoretical_loss": 3.559118451193938, + "tokens_seen": 1431568384 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028598796389167505, + "loss": 1.4395, + "theoretical_loss": 3.5591040727974654, + "tokens_seen": 1431633920 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028597793380140423, + "loss": 1.4638, + "theoretical_loss": 3.559089695243465, + "tokens_seen": 1431699456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002859679037111334, + "loss": 1.4217, + "theoretical_loss": 3.559075318531849, + "tokens_seen": 1431764992 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002859578736208626, + "loss": 1.4282, + "theoretical_loss": 3.5590609426625295, + "tokens_seen": 1431830528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002859478435305918, + "loss": 1.3236, + "theoretical_loss": 3.5590465676354186, + "tokens_seen": 1431896064 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 522998, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2735655307769775, + "objective/train/theoretical_loss": 3.5590321934504283, + "objective/train/tokens_used": 1452421600, + "theoretical_loss": 3.5590321934504283, + "tokens_seen": 1431961600 + }, + { + "epoch": 0.43, + "learning_rate": 0.000285937813440321, + "loss": 1.3707, + "theoretical_loss": 3.5590321934504283, + "tokens_seen": 1431961600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028592778335005014, + "loss": 1.3392, + "theoretical_loss": 3.559017820107471, + "tokens_seen": 1432027136 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002859177532597794, + "loss": 1.4506, + "theoretical_loss": 3.5590034476064587, + "tokens_seen": 1432092672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002859077231695085, + "loss": 1.4592, + "theoretical_loss": 3.558989075947303, + "tokens_seen": 1432158208 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028589769307923774, + "loss": 1.3316, + "theoretical_loss": 3.558974705129917, + "tokens_seen": 1432223744 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858876629889669, + "loss": 1.3209, + "theoretical_loss": 3.558960335154213, + "tokens_seen": 1432289280 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858776328986961, + "loss": 1.3656, + "theoretical_loss": 3.5589459660201026, + "tokens_seen": 1432354816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858676028084253, + "loss": 1.336, + "theoretical_loss": 3.558931597727498, + "tokens_seen": 1432420352 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858575727181545, + "loss": 1.3768, + "theoretical_loss": 3.558917230276311, + "tokens_seen": 1432485888 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028584754262788364, + "loss": 1.3918, + "theoretical_loss": 3.5589028636664546, + "tokens_seen": 1432551424 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858375125376129, + "loss": 1.442, + "theoretical_loss": 3.558888497897841, + "tokens_seen": 1432616960 + }, + { + "epoch": 0.43, + "learning_rate": 0.000285827482447342, + "loss": 1.4232, + "theoretical_loss": 3.5588741329703826, + "tokens_seen": 1432682496 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028581745235707124, + "loss": 1.4748, + "theoretical_loss": 3.558859768883991, + "tokens_seen": 1432748032 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002858074222668004, + "loss": 1.4045, + "theoretical_loss": 3.5588454056385785, + "tokens_seen": 1432813568 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857973921765296, + "loss": 1.3112, + "theoretical_loss": 3.5588310432340577, + "tokens_seen": 1432879104 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857873620862588, + "loss": 1.2311, + "theoretical_loss": 3.5588166816703417, + "tokens_seen": 1432944640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028577733199598796, + "loss": 1.339, + "theoretical_loss": 3.5588023209473416, + "tokens_seen": 1433010176 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028576730190571714, + "loss": 1.335, + "theoretical_loss": 3.5587879610649704, + "tokens_seen": 1433075712 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857572718154464, + "loss": 1.3395, + "theoretical_loss": 3.55877360202314, + "tokens_seen": 1433141248 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857472417251755, + "loss": 1.3135, + "theoretical_loss": 3.558759243821763, + "tokens_seen": 1433206784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028573721163490474, + "loss": 1.4883, + "theoretical_loss": 3.558744886460752, + "tokens_seen": 1433272320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028572718154463387, + "loss": 1.3031, + "theoretical_loss": 3.558730529940019, + "tokens_seen": 1433337856 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857171514543631, + "loss": 1.2327, + "theoretical_loss": 3.5587161742594766, + "tokens_seen": 1433403392 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002857071213640923, + "loss": 1.399, + "theoretical_loss": 3.5587018194190376, + "tokens_seen": 1433468928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028569709127382147, + "loss": 1.2148, + "theoretical_loss": 3.558687465418614, + "tokens_seen": 1433534464 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028568706118355065, + "loss": 1.3402, + "theoretical_loss": 3.558673112258118, + "tokens_seen": 1433600000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002856770310932799, + "loss": 1.4171, + "theoretical_loss": 3.5586587599374626, + "tokens_seen": 1433665536 + }, + { + "epoch": 0.43, + "learning_rate": 0.000285667001003009, + "loss": 1.439, + "theoretical_loss": 3.5586444084565603, + "tokens_seen": 1433731072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028565697091273825, + "loss": 1.4408, + "theoretical_loss": 3.5586300578153227, + "tokens_seen": 1433796608 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028564694082246737, + "loss": 1.4714, + "theoretical_loss": 3.5586157080136633, + "tokens_seen": 1433862144 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002856369107321966, + "loss": 1.4323, + "theoretical_loss": 3.558601359051494, + "tokens_seen": 1433927680 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002856268806419258, + "loss": 1.3947, + "theoretical_loss": 3.5585870109287274, + "tokens_seen": 1433993216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028561685055165497, + "loss": 1.4698, + "theoretical_loss": 3.558572663645277, + "tokens_seen": 1434058752 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028560682046138415, + "loss": 1.3924, + "theoretical_loss": 3.5585583172010535, + "tokens_seen": 1434124288 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028559679037111333, + "loss": 1.3706, + "theoretical_loss": 3.5585439715959715, + "tokens_seen": 1434189824 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002855867602808425, + "loss": 1.4307, + "theoretical_loss": 3.558529626829942, + "tokens_seen": 1434255360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028557673019057175, + "loss": 1.3813, + "theoretical_loss": 3.558515282902878, + "tokens_seen": 1434320896 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002855667001003009, + "loss": 1.488, + "theoretical_loss": 3.5585009398146927, + "tokens_seen": 1434386432 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002855566700100301, + "loss": 1.3378, + "theoretical_loss": 3.558486597565298, + "tokens_seen": 1434451968 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002855466399197593, + "loss": 1.4256, + "theoretical_loss": 3.558472256154607, + "tokens_seen": 1434517504 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002855366098294885, + "loss": 1.4105, + "theoretical_loss": 3.558457915582532, + "tokens_seen": 1434583040 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028552657973921765, + "loss": 1.3788, + "theoretical_loss": 3.558443575848986, + "tokens_seen": 1434648576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028551654964894684, + "loss": 1.3422, + "theoretical_loss": 3.558429236953881, + "tokens_seen": 1434714112 + }, + { + "epoch": 0.43, + "learning_rate": 0.000285506519558676, + "loss": 1.3679, + "theoretical_loss": 3.558414898897131, + "tokens_seen": 1434779648 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028549648946840525, + "loss": 1.3435, + "theoretical_loss": 3.558400561678647, + "tokens_seen": 1434845184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002854864593781344, + "loss": 1.3482, + "theoretical_loss": 3.558386225298343, + "tokens_seen": 1434910720 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002854764292878636, + "loss": 1.2387, + "theoretical_loss": 3.558371889756131, + "tokens_seen": 1434976256 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028546639919759274, + "loss": 1.2286, + "theoretical_loss": 3.5583575550519244, + "tokens_seen": 1435041792 + }, + { + "epoch": 0.43, + "learning_rate": 0.000285456369107322, + "loss": 1.4109, + "theoretical_loss": 3.558343221185635, + "tokens_seen": 1435107328 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028544633901705116, + "loss": 1.3747, + "theoretical_loss": 3.5583288881571766, + "tokens_seen": 1435172864 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 524207, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.443185567855835, + "objective/train/theoretical_loss": 3.5583145559664615, + "objective/train/tokens_used": 1455698400, + "theoretical_loss": 3.5583145559664615, + "tokens_seen": 1435238400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00028543630892678034, + "loss": 1.4376, + "theoretical_loss": 3.5583145559664615, + "tokens_seen": 1435238400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002854262788365095, + "loss": 1.4347, + "theoretical_loss": 3.5583002246134026, + "tokens_seen": 1435303936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002854162487462387, + "loss": 1.4034, + "theoretical_loss": 3.558285894097912, + "tokens_seen": 1435369472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0002854062186559679, + "loss": 1.3959, + "theoretical_loss": 3.558271564419903, + "tokens_seen": 1435435008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002853961885656971, + "loss": 1.341, + "theoretical_loss": 3.558257235579289, + "tokens_seen": 1435500544 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028538615847542624, + "loss": 1.3751, + "theoretical_loss": 3.5582429075759823, + "tokens_seen": 1435566080 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002853761283851555, + "loss": 1.325, + "theoretical_loss": 3.558228580409896, + "tokens_seen": 1435631616 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028536609829488466, + "loss": 1.3103, + "theoretical_loss": 3.5582142540809425, + "tokens_seen": 1435697152 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028535606820461384, + "loss": 1.4275, + "theoretical_loss": 3.558199928589035, + "tokens_seen": 1435762688 + }, + { + "epoch": 0.44, + "learning_rate": 0.000285346038114343, + "loss": 1.4215, + "theoretical_loss": 3.5581856039340862, + "tokens_seen": 1435828224 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002853360080240722, + "loss": 1.3188, + "theoretical_loss": 3.5581712801160092, + "tokens_seen": 1435893760 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002853259779338014, + "loss": 1.3141, + "theoretical_loss": 3.5581569571347167, + "tokens_seen": 1435959296 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002853159478435306, + "loss": 1.382, + "theoretical_loss": 3.558142634990122, + "tokens_seen": 1436024832 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028530591775325975, + "loss": 1.4151, + "theoretical_loss": 3.558128313682138, + "tokens_seen": 1436090368 + }, + { + "epoch": 0.44, + "learning_rate": 0.000285295887662989, + "loss": 1.3486, + "theoretical_loss": 3.558113993210677, + "tokens_seen": 1436155904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002852858575727181, + "loss": 1.4347, + "theoretical_loss": 3.5580996735756525, + "tokens_seen": 1436221440 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028527582748244735, + "loss": 1.4989, + "theoretical_loss": 3.5580853547769777, + "tokens_seen": 1436286976 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002852657973921766, + "loss": 1.3346, + "theoretical_loss": 3.558071036814565, + "tokens_seen": 1436352512 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002852557673019057, + "loss": 1.3625, + "theoretical_loss": 3.558056719688328, + "tokens_seen": 1436418048 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028524573721163494, + "loss": 1.3138, + "theoretical_loss": 3.5580424033981792, + "tokens_seen": 1436483584 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028523570712136407, + "loss": 1.4291, + "theoretical_loss": 3.5580280879440322, + "tokens_seen": 1436549120 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002852256770310933, + "loss": 1.466, + "theoretical_loss": 3.5580137733257997, + "tokens_seen": 1436614656 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002852156469408225, + "loss": 1.2654, + "theoretical_loss": 3.557999459543394, + "tokens_seen": 1436680192 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028520561685055167, + "loss": 1.574, + "theoretical_loss": 3.55798514659673, + "tokens_seen": 1436745728 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028519558676028085, + "loss": 1.3945, + "theoretical_loss": 3.5579708344857197, + "tokens_seen": 1436811264 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002851855566700101, + "loss": 1.3288, + "theoretical_loss": 3.557956523210276, + "tokens_seen": 1436876800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002851755265797392, + "loss": 1.2389, + "theoretical_loss": 3.5579422127703113, + "tokens_seen": 1436942336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028516549648946845, + "loss": 1.3748, + "theoretical_loss": 3.5579279031657407, + "tokens_seen": 1437007872 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028515546639919757, + "loss": 1.2863, + "theoretical_loss": 3.557913594396476, + "tokens_seen": 1437073408 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002851454363089268, + "loss": 1.2804, + "theoretical_loss": 3.5578992864624306, + "tokens_seen": 1437138944 + }, + { + "epoch": 0.44, + "learning_rate": 0.000285135406218656, + "loss": 1.3868, + "theoretical_loss": 3.557884979363518, + "tokens_seen": 1437204480 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028512537612838517, + "loss": 1.467, + "theoretical_loss": 3.557870673099651, + "tokens_seen": 1437270016 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028511534603811435, + "loss": 1.3245, + "theoretical_loss": 3.557856367670743, + "tokens_seen": 1437335552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028510531594784353, + "loss": 1.3489, + "theoretical_loss": 3.557842063076707, + "tokens_seen": 1437401088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850952858575727, + "loss": 1.37, + "theoretical_loss": 3.5578277593174565, + "tokens_seen": 1437466624 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028508525576730195, + "loss": 1.4, + "theoretical_loss": 3.5578134563929042, + "tokens_seen": 1437532160 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850752256770311, + "loss": 1.304, + "theoretical_loss": 3.5577991543029635, + "tokens_seen": 1437597696 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850651955867603, + "loss": 1.2768, + "theoretical_loss": 3.557784853047548, + "tokens_seen": 1437663232 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850551654964895, + "loss": 1.3782, + "theoretical_loss": 3.557770552626571, + "tokens_seen": 1437728768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850551654964895, + "loss": 1.3799, + "theoretical_loss": 3.557756253039946, + "tokens_seen": 1437794304 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850451354062187, + "loss": 1.4047, + "theoretical_loss": 3.557741954287585, + "tokens_seen": 1437859840 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028503510531594785, + "loss": 1.4532, + "theoretical_loss": 3.5577276563694022, + "tokens_seen": 1437925376 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028502507522567704, + "loss": 1.4826, + "theoretical_loss": 3.5577133592853114, + "tokens_seen": 1437990912 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002850150451354062, + "loss": 1.4253, + "theoretical_loss": 3.557699063035225, + "tokens_seen": 1438056448 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028500501504513545, + "loss": 1.3305, + "theoretical_loss": 3.5576847676190573, + "tokens_seen": 1438121984 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849949849548646, + "loss": 1.3507, + "theoretical_loss": 3.557670473036721, + "tokens_seen": 1438187520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849849548645938, + "loss": 1.4336, + "theoretical_loss": 3.557656179288129, + "tokens_seen": 1438253056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028497492477432294, + "loss": 1.3932, + "theoretical_loss": 3.5576418863731956, + "tokens_seen": 1438318592 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849648946840522, + "loss": 1.4228, + "theoretical_loss": 3.5576275942918345, + "tokens_seen": 1438384128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028495486459378136, + "loss": 1.4795, + "theoretical_loss": 3.557613303043958, + "tokens_seen": 1438449664 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 525418, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3827131986618042, + "objective/train/theoretical_loss": 3.5575990126294794, + "objective/train/tokens_used": 1458975200, + "theoretical_loss": 3.5575990126294794, + "tokens_seen": 1438515200 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028494483450351054, + "loss": 1.3201, + "theoretical_loss": 3.5575990126294794, + "tokens_seen": 1438515200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849348044132397, + "loss": 1.4815, + "theoretical_loss": 3.5575847230483135, + "tokens_seen": 1438580736 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849247743229689, + "loss": 1.4403, + "theoretical_loss": 3.5575704343003727, + "tokens_seen": 1438646272 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849147442326981, + "loss": 1.5051, + "theoretical_loss": 3.557556146385571, + "tokens_seen": 1438711808 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002849047141424273, + "loss": 1.3981, + "theoretical_loss": 3.5575418593038215, + "tokens_seen": 1438777344 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028489468405215644, + "loss": 1.3426, + "theoretical_loss": 3.557527573055038, + "tokens_seen": 1438842880 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848846539618857, + "loss": 1.3808, + "theoretical_loss": 3.5575132876391335, + "tokens_seen": 1438908416 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028487462387161486, + "loss": 1.3208, + "theoretical_loss": 3.557499003056022, + "tokens_seen": 1438973952 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028486459378134404, + "loss": 1.378, + "theoretical_loss": 3.557484719305617, + "tokens_seen": 1439039488 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848545636910732, + "loss": 1.4067, + "theoretical_loss": 3.5574704363878316, + "tokens_seen": 1439105024 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848445336008024, + "loss": 1.4273, + "theoretical_loss": 3.5574561543025798, + "tokens_seen": 1439170560 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848345035105316, + "loss": 1.3435, + "theoretical_loss": 3.5574418730497745, + "tokens_seen": 1439236096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848244734202608, + "loss": 1.3466, + "theoretical_loss": 3.5574275926293306, + "tokens_seen": 1439301632 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028481444332998995, + "loss": 1.2849, + "theoretical_loss": 3.5574133130411605, + "tokens_seen": 1439367168 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002848044132397192, + "loss": 1.427, + "theoretical_loss": 3.5573990342851785, + "tokens_seen": 1439432704 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847943831494483, + "loss": 1.5112, + "theoretical_loss": 3.5573847563612975, + "tokens_seen": 1439498240 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028478435305917755, + "loss": 1.4662, + "theoretical_loss": 3.557370479269432, + "tokens_seen": 1439563776 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847743229689067, + "loss": 1.4464, + "theoretical_loss": 3.5573562030094945, + "tokens_seen": 1439629312 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847642928786359, + "loss": 1.4558, + "theoretical_loss": 3.5573419275813993, + "tokens_seen": 1439694848 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847542627883651, + "loss": 1.343, + "theoretical_loss": 3.5573276529850606, + "tokens_seen": 1439760384 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028474423269809427, + "loss": 1.3606, + "theoretical_loss": 3.5573133792203917, + "tokens_seen": 1439825920 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028473420260782345, + "loss": 1.4299, + "theoretical_loss": 3.5572991062873056, + "tokens_seen": 1439891456 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847241725175527, + "loss": 1.4251, + "theoretical_loss": 3.557284834185717, + "tokens_seen": 1439956992 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002847141424272818, + "loss": 1.4462, + "theoretical_loss": 3.557270562915539, + "tokens_seen": 1440022528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028470411233701105, + "loss": 1.3704, + "theoretical_loss": 3.5572562924766853, + "tokens_seen": 1440088064 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028469408224674023, + "loss": 1.4157, + "theoretical_loss": 3.5572420228690707, + "tokens_seen": 1440153600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846840521564694, + "loss": 1.465, + "theoretical_loss": 3.557227754092607, + "tokens_seen": 1440219136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846740220661986, + "loss": 1.3595, + "theoretical_loss": 3.55721348614721, + "tokens_seen": 1440284672 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846639919759278, + "loss": 1.5785, + "theoretical_loss": 3.5571992190327926, + "tokens_seen": 1440350208 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028465396188565695, + "loss": 1.4361, + "theoretical_loss": 3.557184952749268, + "tokens_seen": 1440415744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846439317953862, + "loss": 1.4077, + "theoretical_loss": 3.5571706872965505, + "tokens_seen": 1440481280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846339017051153, + "loss": 1.3774, + "theoretical_loss": 3.5571564226745545, + "tokens_seen": 1440546816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028462387161484455, + "loss": 1.4118, + "theoretical_loss": 3.5571421588831935, + "tokens_seen": 1440612352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846138415245737, + "loss": 1.332, + "theoretical_loss": 3.5571278959223807, + "tokens_seen": 1440677888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002846038114343029, + "loss": 1.3509, + "theoretical_loss": 3.5571136337920306, + "tokens_seen": 1440743424 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845937813440321, + "loss": 1.2275, + "theoretical_loss": 3.5570993724920568, + "tokens_seen": 1440808960 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845837512537613, + "loss": 1.3571, + "theoretical_loss": 3.5570851120223734, + "tokens_seen": 1440874496 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028457372116349046, + "loss": 1.4788, + "theoretical_loss": 3.5570708523828944, + "tokens_seen": 1440940032 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845636910732197, + "loss": 1.2997, + "theoretical_loss": 3.5570565935735337, + "tokens_seen": 1441005568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845536609829488, + "loss": 1.4816, + "theoretical_loss": 3.5570423355942045, + "tokens_seen": 1441071104 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028454363089267805, + "loss": 1.3925, + "theoretical_loss": 3.5570280784448216, + "tokens_seen": 1441136640 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028453360080240724, + "loss": 1.4686, + "theoretical_loss": 3.557013822125299, + "tokens_seen": 1441202176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845235707121364, + "loss": 1.4393, + "theoretical_loss": 3.5569995666355494, + "tokens_seen": 1441267712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028451354062186565, + "loss": 1.3843, + "theoretical_loss": 3.556985311975488, + "tokens_seen": 1441333248 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002845035105315948, + "loss": 1.2938, + "theoretical_loss": 3.556971058145029, + "tokens_seen": 1441398784 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284493480441324, + "loss": 1.3911, + "theoretical_loss": 3.556956805144085, + "tokens_seen": 1441464320 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028448345035105314, + "loss": 1.301, + "theoretical_loss": 3.556942552972572, + "tokens_seen": 1441529856 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002844734202607824, + "loss": 1.4, + "theoretical_loss": 3.556928301630402, + "tokens_seen": 1441595392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028446339017051156, + "loss": 1.4284, + "theoretical_loss": 3.5569140511174906, + "tokens_seen": 1441660928 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028445336008024074, + "loss": 1.3357, + "theoretical_loss": 3.556899801433751, + "tokens_seen": 1441726464 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 526671, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.291252851486206, + "objective/train/theoretical_loss": 3.5568855525790974, + "objective/train/tokens_used": 1462252000, + "theoretical_loss": 3.5568855525790974, + "tokens_seen": 1441792000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002844433299899699, + "loss": 1.3507, + "theoretical_loss": 3.5568855525790974, + "tokens_seen": 1441792000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002844332998996991, + "loss": 1.5702, + "theoretical_loss": 3.556871304553444, + "tokens_seen": 1441857536 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002844232698094283, + "loss": 1.4809, + "theoretical_loss": 3.556857057356704, + "tokens_seen": 1441923072 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002844132397191575, + "loss": 1.3562, + "theoretical_loss": 3.556842810988794, + "tokens_seen": 1441988608 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028440320962888664, + "loss": 1.3616, + "theoretical_loss": 3.5568285654496252, + "tokens_seen": 1442054144 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843931795386159, + "loss": 1.3265, + "theoretical_loss": 3.5568143207391136, + "tokens_seen": 1442119680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028438314944834506, + "loss": 1.5211, + "theoretical_loss": 3.556800076857172, + "tokens_seen": 1442185216 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028437311935807424, + "loss": 1.39, + "theoretical_loss": 3.5567858338037164, + "tokens_seen": 1442250752 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843630892678034, + "loss": 1.3429, + "theoretical_loss": 3.5567715915786593, + "tokens_seen": 1442316288 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843530591775326, + "loss": 1.3437, + "theoretical_loss": 3.556757350181915, + "tokens_seen": 1442381824 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843430290872618, + "loss": 1.4509, + "theoretical_loss": 3.556743109613399, + "tokens_seen": 1442447360 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284332998996991, + "loss": 1.4783, + "theoretical_loss": 3.556728869873024, + "tokens_seen": 1442512896 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028432296890672015, + "loss": 1.4038, + "theoretical_loss": 3.5567146309607054, + "tokens_seen": 1442578432 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843129388164494, + "loss": 1.4573, + "theoretical_loss": 3.5567003928763565, + "tokens_seen": 1442643968 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002843029087261785, + "loss": 1.4082, + "theoretical_loss": 3.556686155619892, + "tokens_seen": 1442709504 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028429287863590775, + "loss": 1.351, + "theoretical_loss": 3.556671919191226, + "tokens_seen": 1442775040 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002842828485456369, + "loss": 1.2894, + "theoretical_loss": 3.556657683590273, + "tokens_seen": 1442840576 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002842728184553661, + "loss": 1.3088, + "theoretical_loss": 3.5566434488169474, + "tokens_seen": 1442906112 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002842627883650953, + "loss": 1.4001, + "theoretical_loss": 3.5566292148711627, + "tokens_seen": 1442971648 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028425275827482447, + "loss": 1.3407, + "theoretical_loss": 3.5566149817528343, + "tokens_seen": 1443037184 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028424272818455365, + "loss": 1.3353, + "theoretical_loss": 3.556600749461876, + "tokens_seen": 1443102720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002842326980942829, + "loss": 1.2663, + "theoretical_loss": 3.556586517998202, + "tokens_seen": 1443168256 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284222668004012, + "loss": 1.5015, + "theoretical_loss": 3.5565722873617265, + "tokens_seen": 1443233792 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028421263791374125, + "loss": 1.4193, + "theoretical_loss": 3.5565580575523645, + "tokens_seen": 1443299328 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028420260782347043, + "loss": 1.3773, + "theoretical_loss": 3.55654382857003, + "tokens_seen": 1443364864 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841925777331996, + "loss": 1.4486, + "theoretical_loss": 3.5565296004146365, + "tokens_seen": 1443430400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841825476429288, + "loss": 1.3363, + "theoretical_loss": 3.5565153730861, + "tokens_seen": 1443495936 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284172517552658, + "loss": 1.3258, + "theoretical_loss": 3.5565011465843344, + "tokens_seen": 1443561472 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028416248746238715, + "loss": 1.4219, + "theoretical_loss": 3.556486920909254, + "tokens_seen": 1443627008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841524573721164, + "loss": 1.3436, + "theoretical_loss": 3.556472696060773, + "tokens_seen": 1443692544 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841424272818455, + "loss": 1.3171, + "theoretical_loss": 3.5564584720388055, + "tokens_seen": 1443758080 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028413239719157475, + "loss": 1.3847, + "theoretical_loss": 3.556444248843267, + "tokens_seen": 1443823616 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841223671013039, + "loss": 1.3289, + "theoretical_loss": 3.556430026474071, + "tokens_seen": 1443889152 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841123370110331, + "loss": 1.379, + "theoretical_loss": 3.5564158049311327, + "tokens_seen": 1443954688 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002841023069207623, + "loss": 1.3282, + "theoretical_loss": 3.556401584214367, + "tokens_seen": 1444020224 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002840922768304915, + "loss": 1.2446, + "theoretical_loss": 3.5563873643236867, + "tokens_seen": 1444085760 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028408224674022066, + "loss": 1.3885, + "theoretical_loss": 3.5563731452590077, + "tokens_seen": 1444151296 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002840722166499499, + "loss": 1.3753, + "theoretical_loss": 3.5563589270202445, + "tokens_seen": 1444216832 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284062186559679, + "loss": 1.3915, + "theoretical_loss": 3.556344709607311, + "tokens_seen": 1444282368 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028405215646940825, + "loss": 1.3909, + "theoretical_loss": 3.5563304930201225, + "tokens_seen": 1444347904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002840421263791374, + "loss": 1.4595, + "theoretical_loss": 3.556316277258593, + "tokens_seen": 1444413440 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002840320962888666, + "loss": 1.407, + "theoretical_loss": 3.5563020623226373, + "tokens_seen": 1444478976 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002840220661985958, + "loss": 1.4853, + "theoretical_loss": 3.55628784821217, + "tokens_seen": 1444544512 + }, + { + "epoch": 0.44, + "learning_rate": 0.000284012036108325, + "loss": 1.2908, + "theoretical_loss": 3.5562736349271056, + "tokens_seen": 1444610048 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028400200601805416, + "loss": 1.483, + "theoretical_loss": 3.5562594224673587, + "tokens_seen": 1444675584 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028399197592778334, + "loss": 1.4166, + "theoretical_loss": 3.556245210832844, + "tokens_seen": 1444741120 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002839819458375125, + "loss": 1.3603, + "theoretical_loss": 3.556231000023477, + "tokens_seen": 1444806656 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028397191574724176, + "loss": 1.4507, + "theoretical_loss": 3.556216790039171, + "tokens_seen": 1444872192 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002839618856569709, + "loss": 1.4016, + "theoretical_loss": 3.5562025808798414, + "tokens_seen": 1444937728 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002839518555667001, + "loss": 1.3825, + "theoretical_loss": 3.556188372545403, + "tokens_seen": 1445003264 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 528002, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2996197938919067, + "objective/train/theoretical_loss": 3.55617416503577, + "objective/train/tokens_used": 1465528800, + "theoretical_loss": 3.55617416503577, + "tokens_seen": 1445068800 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028394182547642925, + "loss": 1.2966, + "theoretical_loss": 3.55617416503577, + "tokens_seen": 1445068800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002839317953861585, + "loss": 1.2849, + "theoretical_loss": 3.5561599583508574, + "tokens_seen": 1445134336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028392176529588766, + "loss": 1.5217, + "theoretical_loss": 3.55614575249058, + "tokens_seen": 1445199872 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028391173520561684, + "loss": 1.2612, + "theoretical_loss": 3.5561315474548527, + "tokens_seen": 1445265408 + }, + { + "epoch": 0.44, + "learning_rate": 0.000283901705115346, + "loss": 1.355, + "theoretical_loss": 3.5561173432435895, + "tokens_seen": 1445330944 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028389167502507526, + "loss": 1.2209, + "theoretical_loss": 3.556103139856706, + "tokens_seen": 1445396480 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002838816449348044, + "loss": 1.4291, + "theoretical_loss": 3.5560889372941165, + "tokens_seen": 1445462016 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002838716148445336, + "loss": 1.1744, + "theoretical_loss": 3.556074735555736, + "tokens_seen": 1445527552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028386158475426275, + "loss": 1.4452, + "theoretical_loss": 3.556060534641479, + "tokens_seen": 1445593088 + }, + { + "epoch": 0.44, + "learning_rate": 0.000283851554663992, + "loss": 1.4651, + "theoretical_loss": 3.5560463345512607, + "tokens_seen": 1445658624 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028384152457372117, + "loss": 1.3553, + "theoretical_loss": 3.556032135284996, + "tokens_seen": 1445724160 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028383149448345035, + "loss": 1.43, + "theoretical_loss": 3.5560179368425997, + "tokens_seen": 1445789696 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028382146439317953, + "loss": 1.4502, + "theoretical_loss": 3.5560037392239865, + "tokens_seen": 1445855232 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002838114343029087, + "loss": 1.3619, + "theoretical_loss": 3.555989542429071, + "tokens_seen": 1445920768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002838014042126379, + "loss": 1.3531, + "theoretical_loss": 3.5559753464577684, + "tokens_seen": 1445986304 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837913741223671, + "loss": 1.3706, + "theoretical_loss": 3.5559611513099934, + "tokens_seen": 1446051840 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837813440320963, + "loss": 1.408, + "theoretical_loss": 3.555946956985661, + "tokens_seen": 1446117376 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837713139418255, + "loss": 1.3946, + "theoretical_loss": 3.5559327634846865, + "tokens_seen": 1446182912 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028376128385155467, + "loss": 1.3373, + "theoretical_loss": 3.5559185708069845, + "tokens_seen": 1446248448 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028375125376128385, + "loss": 1.4262, + "theoretical_loss": 3.55590437895247, + "tokens_seen": 1446313984 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837412236710131, + "loss": 1.39, + "theoretical_loss": 3.5558901879210576, + "tokens_seen": 1446379520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837311935807422, + "loss": 1.3454, + "theoretical_loss": 3.555875997712663, + "tokens_seen": 1446445056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028372116349047145, + "loss": 1.3115, + "theoretical_loss": 3.5558618083272, + "tokens_seen": 1446510592 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028371113340020063, + "loss": 1.3556, + "theoretical_loss": 3.5558476197645854, + "tokens_seen": 1446576128 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002837011033099298, + "loss": 1.4568, + "theoretical_loss": 3.5558334320247322, + "tokens_seen": 1446641664 + }, + { + "epoch": 0.44, + "learning_rate": 0.000283691073219659, + "loss": 1.347, + "theoretical_loss": 3.555819245107557, + "tokens_seen": 1446707200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836810431293882, + "loss": 1.3576, + "theoretical_loss": 3.555805059012974, + "tokens_seen": 1446772736 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028367101303911735, + "loss": 1.4638, + "theoretical_loss": 3.5557908737408983, + "tokens_seen": 1446838272 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836609829488466, + "loss": 1.2892, + "theoretical_loss": 3.5557766892912452, + "tokens_seen": 1446903808 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836509528585757, + "loss": 1.3512, + "theoretical_loss": 3.55576250566393, + "tokens_seen": 1446969344 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028364092276830495, + "loss": 1.3467, + "theoretical_loss": 3.555748322858867, + "tokens_seen": 1447034880 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836308926780341, + "loss": 1.2798, + "theoretical_loss": 3.555734140875972, + "tokens_seen": 1447100416 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836208625877633, + "loss": 1.3404, + "theoretical_loss": 3.5557199597151596, + "tokens_seen": 1447165952 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836108324974925, + "loss": 1.2898, + "theoretical_loss": 3.555705779376346, + "tokens_seen": 1447231488 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002836008024072217, + "loss": 1.4138, + "theoretical_loss": 3.5556915998594447, + "tokens_seen": 1447297024 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028359077231695086, + "loss": 1.3598, + "theoretical_loss": 3.555677421164372, + "tokens_seen": 1447362560 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002835807422266801, + "loss": 1.2356, + "theoretical_loss": 3.555663243291043, + "tokens_seen": 1447428096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002835707121364092, + "loss": 1.3802, + "theoretical_loss": 3.555649066239372, + "tokens_seen": 1447493632 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028356068204613846, + "loss": 1.2935, + "theoretical_loss": 3.555634890009275, + "tokens_seen": 1447559168 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002835506519558676, + "loss": 1.3688, + "theoretical_loss": 3.5556207146006673, + "tokens_seen": 1447624704 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002835406218655968, + "loss": 1.4041, + "theoretical_loss": 3.5556065400134633, + "tokens_seen": 1447690240 + }, + { + "epoch": 0.44, + "learning_rate": 0.000283530591775326, + "loss": 1.3752, + "theoretical_loss": 3.5555923662475792, + "tokens_seen": 1447755776 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002835205616850552, + "loss": 1.3167, + "theoretical_loss": 3.55557819330293, + "tokens_seen": 1447821312 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028351053159478436, + "loss": 1.4182, + "theoretical_loss": 3.55556402117943, + "tokens_seen": 1447886848 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028350050150451354, + "loss": 1.3607, + "theoretical_loss": 3.5555498498769955, + "tokens_seen": 1447952384 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002834904714142427, + "loss": 1.3634, + "theoretical_loss": 3.555535679395541, + "tokens_seen": 1448017920 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028348044132397196, + "loss": 1.3296, + "theoretical_loss": 3.555521509734983, + "tokens_seen": 1448083456 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002834704112337011, + "loss": 1.41, + "theoretical_loss": 3.555507340895235, + "tokens_seen": 1448148992 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002834603811434303, + "loss": 1.4134, + "theoretical_loss": 3.555493172876214, + "tokens_seen": 1448214528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028345035105315945, + "loss": 1.3235, + "theoretical_loss": 3.555479005677835, + "tokens_seen": 1448280064 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 529166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4000072479248047, + "objective/train/theoretical_loss": 3.5554648393000123, + "objective/train/tokens_used": 1468805600, + "theoretical_loss": 3.5554648393000123, + "tokens_seen": 1448345600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002834403209628887, + "loss": 1.408, + "theoretical_loss": 3.5554648393000123, + "tokens_seen": 1448345600 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028343029087261786, + "loss": 1.3262, + "theoretical_loss": 3.555450673742662, + "tokens_seen": 1448411136 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028342026078234705, + "loss": 1.4067, + "theoretical_loss": 3.5554365090057, + "tokens_seen": 1448476672 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002834102306920762, + "loss": 1.4807, + "theoretical_loss": 3.5554223450890405, + "tokens_seen": 1448542208 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028340020060180546, + "loss": 1.3958, + "theoretical_loss": 3.5554081819925996, + "tokens_seen": 1448607744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002833901705115346, + "loss": 1.3491, + "theoretical_loss": 3.5553940197162923, + "tokens_seen": 1448673280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002833801404212638, + "loss": 1.4055, + "theoretical_loss": 3.555379858260035, + "tokens_seen": 1448738816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028337011033099295, + "loss": 1.4212, + "theoretical_loss": 3.5553656976237415, + "tokens_seen": 1448804352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002833600802407222, + "loss": 1.2727, + "theoretical_loss": 3.5553515378073284, + "tokens_seen": 1448869888 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028335005015045137, + "loss": 1.3543, + "theoretical_loss": 3.5553373788107114, + "tokens_seen": 1448935424 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028334002006018055, + "loss": 1.3774, + "theoretical_loss": 3.555323220633805, + "tokens_seen": 1449000960 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028332998996990973, + "loss": 1.4245, + "theoretical_loss": 3.5553090632765247, + "tokens_seen": 1449066496 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002833199598796389, + "loss": 1.5125, + "theoretical_loss": 3.555294906738787, + "tokens_seen": 1449132032 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002833099297893681, + "loss": 1.2953, + "theoretical_loss": 3.5552807510205064, + "tokens_seen": 1449197568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002832998996990973, + "loss": 1.3795, + "theoretical_loss": 3.555266596121599, + "tokens_seen": 1449263104 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028328986960882645, + "loss": 1.4518, + "theoretical_loss": 3.55525244204198, + "tokens_seen": 1449328640 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002832798395185557, + "loss": 1.5063, + "theoretical_loss": 3.555238288781565, + "tokens_seen": 1449394176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002832698094282848, + "loss": 1.3533, + "theoretical_loss": 3.5552241363402697, + "tokens_seen": 1449459712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028325977933801405, + "loss": 1.3679, + "theoretical_loss": 3.55520998471801, + "tokens_seen": 1449525248 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028324974924774323, + "loss": 1.4283, + "theoretical_loss": 3.5551958339147003, + "tokens_seen": 1449590784 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002832397191574724, + "loss": 1.3234, + "theoretical_loss": 3.5551816839302575, + "tokens_seen": 1449656320 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002832296890672016, + "loss": 1.4567, + "theoretical_loss": 3.555167534764596, + "tokens_seen": 1449721856 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028321965897693083, + "loss": 1.2226, + "theoretical_loss": 3.555153386417633, + "tokens_seen": 1449787392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028320962888665996, + "loss": 1.4485, + "theoretical_loss": 3.555139238889282, + "tokens_seen": 1449852928 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831995987963892, + "loss": 1.33, + "theoretical_loss": 3.5551250921794604, + "tokens_seen": 1449918464 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831895687061183, + "loss": 1.3825, + "theoretical_loss": 3.5551109462880834, + "tokens_seen": 1449984000 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028317953861584755, + "loss": 1.4322, + "theoretical_loss": 3.555096801215066, + "tokens_seen": 1450049536 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028316950852557674, + "loss": 1.474, + "theoretical_loss": 3.555082656960325, + "tokens_seen": 1450115072 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831594784353059, + "loss": 1.3307, + "theoretical_loss": 3.555068513523775, + "tokens_seen": 1450180608 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831494483450351, + "loss": 1.4324, + "theoretical_loss": 3.555054370905332, + "tokens_seen": 1450246144 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831394182547643, + "loss": 1.4388, + "theoretical_loss": 3.5550402291049124, + "tokens_seen": 1450311680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028312938816449346, + "loss": 1.4894, + "theoretical_loss": 3.555026088122431, + "tokens_seen": 1450377216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831193580742227, + "loss": 1.3957, + "theoretical_loss": 3.555011947957804, + "tokens_seen": 1450442752 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002831093279839518, + "loss": 1.3018, + "theoretical_loss": 3.5549978086109473, + "tokens_seen": 1450508288 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028309929789368106, + "loss": 1.3299, + "theoretical_loss": 3.554983670081776, + "tokens_seen": 1450573824 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830892678034102, + "loss": 1.2803, + "theoretical_loss": 3.554969532370207, + "tokens_seen": 1450639360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830792377131394, + "loss": 1.426, + "theoretical_loss": 3.5549553954761546, + "tokens_seen": 1450704896 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830692076228686, + "loss": 1.4137, + "theoretical_loss": 3.554941259399536, + "tokens_seen": 1450770432 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830591775325978, + "loss": 1.3416, + "theoretical_loss": 3.5549271241402662, + "tokens_seen": 1450835968 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028304914744232696, + "loss": 1.3546, + "theoretical_loss": 3.554912989698261, + "tokens_seen": 1450901504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830391173520562, + "loss": 1.4287, + "theoretical_loss": 3.5548988560734367, + "tokens_seen": 1450967040 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002830290872617854, + "loss": 1.3264, + "theoretical_loss": 3.554884723265709, + "tokens_seen": 1451032576 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028301905717151456, + "loss": 1.3666, + "theoretical_loss": 3.5548705912749936, + "tokens_seen": 1451098112 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028300902708124374, + "loss": 1.3921, + "theoretical_loss": 3.554856460101206, + "tokens_seen": 1451163648 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002829989969909729, + "loss": 1.3728, + "theoretical_loss": 3.5548423297442633, + "tokens_seen": 1451229184 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028298896690070216, + "loss": 1.3913, + "theoretical_loss": 3.5548282002040805, + "tokens_seen": 1451294720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002829789368104313, + "loss": 1.4521, + "theoretical_loss": 3.554814071480573, + "tokens_seen": 1451360256 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002829689067201605, + "loss": 1.3392, + "theoretical_loss": 3.554799943573658, + "tokens_seen": 1451425792 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028295887662988965, + "loss": 1.4502, + "theoretical_loss": 3.5547858164832506, + "tokens_seen": 1451491328 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002829488465396189, + "loss": 1.4614, + "theoretical_loss": 3.554771690209267, + "tokens_seen": 1451556864 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 530451, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5022616386413574, + "objective/train/theoretical_loss": 3.5547575647516227, + "objective/train/tokens_used": 1472082400, + "theoretical_loss": 3.5547575647516227, + "tokens_seen": 1451622400 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028293881644934806, + "loss": 1.524, + "theoretical_loss": 3.5547575647516227, + "tokens_seen": 1451622400 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028292878635907725, + "loss": 1.3359, + "theoretical_loss": 3.5547434401102347, + "tokens_seen": 1451687936 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002829187562688064, + "loss": 1.3487, + "theoretical_loss": 3.554729316285018, + "tokens_seen": 1451753472 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028290872617853566, + "loss": 1.3239, + "theoretical_loss": 3.5547151932758894, + "tokens_seen": 1451819008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002828986960882648, + "loss": 1.4384, + "theoretical_loss": 3.554701071082764, + "tokens_seen": 1451884544 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282888665997994, + "loss": 1.329, + "theoretical_loss": 3.5546869497055584, + "tokens_seen": 1451950080 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028287863590772315, + "loss": 1.3936, + "theoretical_loss": 3.554672829144189, + "tokens_seen": 1452015616 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002828686058174524, + "loss": 1.3388, + "theoretical_loss": 3.5546587093985704, + "tokens_seen": 1452081152 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028285857572718157, + "loss": 1.4235, + "theoretical_loss": 3.5546445904686204, + "tokens_seen": 1452146688 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028284854563691075, + "loss": 1.4562, + "theoretical_loss": 3.554630472354254, + "tokens_seen": 1452212224 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028283851554663993, + "loss": 1.3108, + "theoretical_loss": 3.554616355055388, + "tokens_seen": 1452277760 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002828284854563691, + "loss": 1.4225, + "theoretical_loss": 3.5546022385719374, + "tokens_seen": 1452343296 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002828184553660983, + "loss": 1.3693, + "theoretical_loss": 3.5545881229038194, + "tokens_seen": 1452408832 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002828084252758275, + "loss": 1.3533, + "theoretical_loss": 3.55457400805095, + "tokens_seen": 1452474368 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028279839518555665, + "loss": 1.4573, + "theoretical_loss": 3.554559894013245, + "tokens_seen": 1452539904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002827883650952859, + "loss": 1.462, + "theoretical_loss": 3.5545457807906202, + "tokens_seen": 1452605440 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282778335005015, + "loss": 1.407, + "theoretical_loss": 3.5545316683829924, + "tokens_seen": 1452670976 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028276830491474425, + "loss": 1.4047, + "theoretical_loss": 3.5545175567902776, + "tokens_seen": 1452736512 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028275827482447343, + "loss": 1.366, + "theoretical_loss": 3.554503446012392, + "tokens_seen": 1452802048 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002827482447342026, + "loss": 1.3395, + "theoretical_loss": 3.5544893360492513, + "tokens_seen": 1452867584 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002827382146439318, + "loss": 1.3701, + "theoretical_loss": 3.5544752269007724, + "tokens_seen": 1452933120 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028272818455366103, + "loss": 1.3572, + "theoretical_loss": 3.5544611185668713, + "tokens_seen": 1452998656 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028271815446339016, + "loss": 1.4497, + "theoretical_loss": 3.554447011047464, + "tokens_seen": 1453064192 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002827081243731194, + "loss": 1.349, + "theoretical_loss": 3.554432904342467, + "tokens_seen": 1453129728 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002826980942828485, + "loss": 1.4031, + "theoretical_loss": 3.5544187984517968, + "tokens_seen": 1453195264 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028268806419257775, + "loss": 1.367, + "theoretical_loss": 3.5544046933753695, + "tokens_seen": 1453260800 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028267803410230694, + "loss": 1.3605, + "theoretical_loss": 3.5543905891131007, + "tokens_seen": 1453326336 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002826680040120361, + "loss": 1.4038, + "theoretical_loss": 3.554376485664907, + "tokens_seen": 1453391872 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002826579739217653, + "loss": 1.3335, + "theoretical_loss": 3.5543623830307056, + "tokens_seen": 1453457408 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002826479438314945, + "loss": 1.4131, + "theoretical_loss": 3.554348281210412, + "tokens_seen": 1453522944 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028263791374122366, + "loss": 1.4557, + "theoretical_loss": 3.5543341802039423, + "tokens_seen": 1453588480 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002826278836509529, + "loss": 1.3779, + "theoretical_loss": 3.5543200800112134, + "tokens_seen": 1453654016 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282617853560682, + "loss": 1.2651, + "theoretical_loss": 3.554305980632142, + "tokens_seen": 1453719552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028260782347041126, + "loss": 1.3566, + "theoretical_loss": 3.554291882066644, + "tokens_seen": 1453785088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825977933801404, + "loss": 1.4282, + "theoretical_loss": 3.554277784314635, + "tokens_seen": 1453850624 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825877632898696, + "loss": 1.3393, + "theoretical_loss": 3.5542636873760323, + "tokens_seen": 1453916160 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825777331995988, + "loss": 1.4351, + "theoretical_loss": 3.554249591250753, + "tokens_seen": 1453981696 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282567703109328, + "loss": 1.2875, + "theoretical_loss": 3.5542354959387117, + "tokens_seen": 1454047232 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028255767301905716, + "loss": 1.531, + "theoretical_loss": 3.5542214014398263, + "tokens_seen": 1454112768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825476429287864, + "loss": 1.2675, + "theoretical_loss": 3.5542073077540124, + "tokens_seen": 1454178304 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825376128385155, + "loss": 1.3855, + "theoretical_loss": 3.5541932148811872, + "tokens_seen": 1454243840 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028252758274824476, + "loss": 1.5439, + "theoretical_loss": 3.5541791228212665, + "tokens_seen": 1454309376 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825175526579739, + "loss": 1.4491, + "theoretical_loss": 3.554165031574167, + "tokens_seen": 1454374912 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002825075225677031, + "loss": 1.3037, + "theoretical_loss": 3.5541509411398056, + "tokens_seen": 1454440448 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002824974924774323, + "loss": 1.3688, + "theoretical_loss": 3.5541368515180984, + "tokens_seen": 1454505984 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002824874623871615, + "loss": 1.4356, + "theoretical_loss": 3.5541227627089613, + "tokens_seen": 1454571520 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028247743229689067, + "loss": 1.3304, + "theoretical_loss": 3.554108674712312, + "tokens_seen": 1454637056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028246740220661985, + "loss": 1.3839, + "theoretical_loss": 3.554094587528067, + "tokens_seen": 1454702592 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028245737211634903, + "loss": 1.3585, + "theoretical_loss": 3.5540805011561423, + "tokens_seen": 1454768128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028244734202607826, + "loss": 1.3923, + "theoretical_loss": 3.554066415596454, + "tokens_seen": 1454833664 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 531642, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4474467039108276, + "objective/train/theoretical_loss": 3.5540523308489194, + "objective/train/tokens_used": 1475359200, + "theoretical_loss": 3.5540523308489194, + "tokens_seen": 1454899200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002824373119358074, + "loss": 1.3341, + "theoretical_loss": 3.5540523308489194, + "tokens_seen": 1454899200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002824272818455366, + "loss": 1.3882, + "theoretical_loss": 3.5540382469134553, + "tokens_seen": 1454964736 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002824172517552658, + "loss": 1.2952, + "theoretical_loss": 3.5540241637899777, + "tokens_seen": 1455030272 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282407221664995, + "loss": 1.4228, + "theoretical_loss": 3.5540100814784035, + "tokens_seen": 1455095808 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028239719157472417, + "loss": 1.3995, + "theoretical_loss": 3.553995999978649, + "tokens_seen": 1455161344 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028238716148445335, + "loss": 1.3229, + "theoretical_loss": 3.5539819192906315, + "tokens_seen": 1455226880 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028237713139418253, + "loss": 1.3014, + "theoretical_loss": 3.553967839414267, + "tokens_seen": 1455292416 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028236710130391177, + "loss": 1.3293, + "theoretical_loss": 3.5539537603494726, + "tokens_seen": 1455357952 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002823570712136409, + "loss": 1.5394, + "theoretical_loss": 3.553939682096165, + "tokens_seen": 1455423488 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028234704112337013, + "loss": 1.2165, + "theoretical_loss": 3.5539256046542604, + "tokens_seen": 1455489024 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028233701103309926, + "loss": 1.4438, + "theoretical_loss": 3.553911528023676, + "tokens_seen": 1455554560 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002823269809428285, + "loss": 1.3491, + "theoretical_loss": 3.5538974522043283, + "tokens_seen": 1455620096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002823169508525577, + "loss": 1.3659, + "theoretical_loss": 3.553883377196134, + "tokens_seen": 1455685632 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028230692076228685, + "loss": 1.3744, + "theoretical_loss": 3.55386930299901, + "tokens_seen": 1455751168 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028229689067201604, + "loss": 1.3269, + "theoretical_loss": 3.553855229612873, + "tokens_seen": 1455816704 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002822868605817452, + "loss": 1.3732, + "theoretical_loss": 3.55384115703764, + "tokens_seen": 1455882240 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028227683049147445, + "loss": 1.3546, + "theoretical_loss": 3.5538270852732268, + "tokens_seen": 1455947776 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028226680040120363, + "loss": 1.4896, + "theoretical_loss": 3.553813014319551, + "tokens_seen": 1456013312 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002822567703109328, + "loss": 1.3444, + "theoretical_loss": 3.55379894417653, + "tokens_seen": 1456078848 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282246740220662, + "loss": 1.3325, + "theoretical_loss": 3.5537848748440792, + "tokens_seen": 1456144384 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028223671013039123, + "loss": 1.3085, + "theoretical_loss": 3.5537708063221167, + "tokens_seen": 1456209920 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028222668004012036, + "loss": 1.3228, + "theoretical_loss": 3.5537567386105584, + "tokens_seen": 1456275456 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002822166499498496, + "loss": 1.3621, + "theoretical_loss": 3.5537426717093217, + "tokens_seen": 1456340992 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002822066198595787, + "loss": 1.2851, + "theoretical_loss": 3.553728605618323, + "tokens_seen": 1456406528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028219658976930795, + "loss": 1.3539, + "theoretical_loss": 3.5537145403374795, + "tokens_seen": 1456472064 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028218655967903714, + "loss": 1.4285, + "theoretical_loss": 3.553700475866708, + "tokens_seen": 1456537600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821765295887663, + "loss": 1.3472, + "theoretical_loss": 3.5536864122059257, + "tokens_seen": 1456603136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821664994984955, + "loss": 1.3118, + "theoretical_loss": 3.5536723493550495, + "tokens_seen": 1456668672 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821564694082247, + "loss": 1.3396, + "theoretical_loss": 3.5536582873139952, + "tokens_seen": 1456734208 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028214643931795386, + "loss": 1.4612, + "theoretical_loss": 3.5536442260826813, + "tokens_seen": 1456799744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821364092276831, + "loss": 1.5368, + "theoretical_loss": 3.5536301656610236, + "tokens_seen": 1456865280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821263791374122, + "loss": 1.405, + "theoretical_loss": 3.55361610604894, + "tokens_seen": 1456930816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028211634904714146, + "loss": 1.4824, + "theoretical_loss": 3.553602047246347, + "tokens_seen": 1456996352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002821063189568706, + "loss": 1.3417, + "theoretical_loss": 3.5535879892531614, + "tokens_seen": 1457061888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820962888665998, + "loss": 1.2817, + "theoretical_loss": 3.5535739320693, + "tokens_seen": 1457127424 + }, + { + "epoch": 0.44, + "learning_rate": 0.000282086258776329, + "loss": 1.5056, + "theoretical_loss": 3.5535598756946807, + "tokens_seen": 1457192960 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820762286860582, + "loss": 1.4414, + "theoretical_loss": 3.55354582012922, + "tokens_seen": 1457258496 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028206619859578736, + "loss": 1.4454, + "theoretical_loss": 3.553531765372835, + "tokens_seen": 1457324032 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820561685055166, + "loss": 1.3225, + "theoretical_loss": 3.5535177114254424, + "tokens_seen": 1457389568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820461384152457, + "loss": 1.3004, + "theoretical_loss": 3.5535036582869592, + "tokens_seen": 1457455104 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028203610832497496, + "loss": 1.432, + "theoretical_loss": 3.5534896059573033, + "tokens_seen": 1457520640 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820260782347041, + "loss": 1.3509, + "theoretical_loss": 3.5534755544363916, + "tokens_seen": 1457586176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820160481444333, + "loss": 1.429, + "theoretical_loss": 3.55346150372414, + "tokens_seen": 1457651712 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002820060180541625, + "loss": 1.432, + "theoretical_loss": 3.5534474538204672, + "tokens_seen": 1457717248 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002819959879638917, + "loss": 1.5124, + "theoretical_loss": 3.55343340472529, + "tokens_seen": 1457782784 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028198595787362087, + "loss": 1.4999, + "theoretical_loss": 3.553419356438524, + "tokens_seen": 1457848320 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028197592778335005, + "loss": 1.4467, + "theoretical_loss": 3.5534053089600883, + "tokens_seen": 1457913856 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028196589769307923, + "loss": 1.4105, + "theoretical_loss": 3.553391262289899, + "tokens_seen": 1457979392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028195586760280846, + "loss": 1.3212, + "theoretical_loss": 3.5533772164278736, + "tokens_seen": 1458044928 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002819458375125376, + "loss": 1.4199, + "theoretical_loss": 3.5533631713739293, + "tokens_seen": 1458110464 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 532886, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3000106811523438, + "objective/train/theoretical_loss": 3.553349127127983, + "objective/train/tokens_used": 1478636000, + "theoretical_loss": 3.553349127127983, + "tokens_seen": 1458176000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002819358074222668, + "loss": 1.3754, + "theoretical_loss": 3.553349127127983, + "tokens_seen": 1458176000 + }, + { + "epoch": 0.44, + "learning_rate": 0.000281925777331996, + "loss": 1.3984, + "theoretical_loss": 3.5533350836899524, + "tokens_seen": 1458241536 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002819157472417252, + "loss": 1.2484, + "theoretical_loss": 3.553321041059754, + "tokens_seen": 1458307072 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028190571715145437, + "loss": 1.4007, + "theoretical_loss": 3.553306999237306, + "tokens_seen": 1458372608 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028189568706118355, + "loss": 1.4947, + "theoretical_loss": 3.5532929582225243, + "tokens_seen": 1458438144 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028188565697091273, + "loss": 1.468, + "theoretical_loss": 3.5532789180153275, + "tokens_seen": 1458503680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028187562688064197, + "loss": 1.3136, + "theoretical_loss": 3.5532648786156322, + "tokens_seen": 1458569216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002818655967903711, + "loss": 1.4319, + "theoretical_loss": 3.553250840023356, + "tokens_seen": 1458634752 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028185556670010033, + "loss": 1.3075, + "theoretical_loss": 3.553236802238416, + "tokens_seen": 1458700288 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028184553660982946, + "loss": 1.436, + "theoretical_loss": 3.5532227652607293, + "tokens_seen": 1458765824 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002818355065195587, + "loss": 1.3678, + "theoretical_loss": 3.5532087290902132, + "tokens_seen": 1458831360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002818254764292879, + "loss": 1.3713, + "theoretical_loss": 3.5531946937267858, + "tokens_seen": 1458896896 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028181544633901705, + "loss": 1.3357, + "theoretical_loss": 3.553180659170363, + "tokens_seen": 1458962432 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028180541624874624, + "loss": 1.3877, + "theoretical_loss": 3.5531666254208636, + "tokens_seen": 1459027968 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817953861584754, + "loss": 1.3678, + "theoretical_loss": 3.5531525924782046, + "tokens_seen": 1459093504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817853560682046, + "loss": 1.4429, + "theoretical_loss": 3.553138560342303, + "tokens_seen": 1459159040 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028177532597793383, + "loss": 1.3005, + "theoretical_loss": 3.553124529013076, + "tokens_seen": 1459224576 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028176529588766296, + "loss": 1.2878, + "theoretical_loss": 3.5531104984904416, + "tokens_seen": 1459290112 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817552657973922, + "loss": 1.529, + "theoretical_loss": 3.5530964687743167, + "tokens_seen": 1459355648 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817452357071214, + "loss": 1.3321, + "theoretical_loss": 3.5530824398646192, + "tokens_seen": 1459421184 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028173520561685056, + "loss": 1.5106, + "theoretical_loss": 3.5530684117612665, + "tokens_seen": 1459486720 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028172517552657974, + "loss": 1.4672, + "theoretical_loss": 3.5530543844641755, + "tokens_seen": 1459552256 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817151454363089, + "loss": 1.3454, + "theoretical_loss": 3.5530403579732646, + "tokens_seen": 1459617792 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002817051153460381, + "loss": 1.3453, + "theoretical_loss": 3.55302633228845, + "tokens_seen": 1459683328 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028169508525576734, + "loss": 1.469, + "theoretical_loss": 3.5530123074096505, + "tokens_seen": 1459748864 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028168505516549646, + "loss": 1.4307, + "theoretical_loss": 3.552998283336782, + "tokens_seen": 1459814400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002816750250752257, + "loss": 1.3041, + "theoretical_loss": 3.5529842600697643, + "tokens_seen": 1459879936 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002816649949849548, + "loss": 1.4441, + "theoretical_loss": 3.5529702376085126, + "tokens_seen": 1459945472 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028165496489468406, + "loss": 1.2784, + "theoretical_loss": 3.552956215952946, + "tokens_seen": 1460011008 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028164493480441324, + "loss": 1.4396, + "theoretical_loss": 3.5529421951029816, + "tokens_seen": 1460076544 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002816349047141424, + "loss": 1.4158, + "theoretical_loss": 3.5529281750585366, + "tokens_seen": 1460142080 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002816248746238716, + "loss": 1.2918, + "theoretical_loss": 3.5529141558195287, + "tokens_seen": 1460207616 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002816148445336008, + "loss": 1.4396, + "theoretical_loss": 3.552900137385876, + "tokens_seen": 1460273152 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028160481444332997, + "loss": 1.4175, + "theoretical_loss": 3.5528861197574955, + "tokens_seen": 1460338688 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815947843530592, + "loss": 1.4397, + "theoretical_loss": 3.552872102934305, + "tokens_seen": 1460404224 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028158475426278833, + "loss": 1.3778, + "theoretical_loss": 3.552858086916222, + "tokens_seen": 1460469760 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028157472417251756, + "loss": 1.3223, + "theoretical_loss": 3.552844071703164, + "tokens_seen": 1460535296 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028156469408224674, + "loss": 1.4093, + "theoretical_loss": 3.5528300572950497, + "tokens_seen": 1460600832 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815546639919759, + "loss": 1.469, + "theoretical_loss": 3.5528160436917955, + "tokens_seen": 1460666368 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815446339017051, + "loss": 1.3574, + "theoretical_loss": 3.5528020308933197, + "tokens_seen": 1460731904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815346038114343, + "loss": 1.2978, + "theoretical_loss": 3.5527880188995393, + "tokens_seen": 1460797440 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815245737211635, + "loss": 1.3554, + "theoretical_loss": 3.552774007710373, + "tokens_seen": 1460862976 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815145436308927, + "loss": 1.5222, + "theoretical_loss": 3.5527599973257376, + "tokens_seen": 1460928512 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002815045135406219, + "loss": 1.4396, + "theoretical_loss": 3.552745987745552, + "tokens_seen": 1460994048 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028149448345035107, + "loss": 1.4659, + "theoretical_loss": 3.5527319789697325, + "tokens_seen": 1461059584 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028148445336008025, + "loss": 1.4396, + "theoretical_loss": 3.5527179709981973, + "tokens_seen": 1461125120 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028147442326980943, + "loss": 1.3531, + "theoretical_loss": 3.5527039638308646, + "tokens_seen": 1461190656 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028146439317953866, + "loss": 1.3067, + "theoretical_loss": 3.552689957467652, + "tokens_seen": 1461256192 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002814543630892678, + "loss": 1.4551, + "theoretical_loss": 3.552675951908477, + "tokens_seen": 1461321728 + }, + { + "epoch": 0.44, + "learning_rate": 0.000281444332998997, + "loss": 1.3322, + "theoretical_loss": 3.5526619471532577, + "tokens_seen": 1461387264 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 534190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3692913055419922, + "objective/train/theoretical_loss": 3.552647943201912, + "objective/train/tokens_used": 1481912800, + "theoretical_loss": 3.552647943201912, + "tokens_seen": 1461452800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002814343029087262, + "loss": 1.4072, + "theoretical_loss": 3.552647943201912, + "tokens_seen": 1461452800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002814242728184554, + "loss": 1.4012, + "theoretical_loss": 3.552633940054357, + "tokens_seen": 1461518336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028141424272818457, + "loss": 1.3868, + "theoretical_loss": 3.5526199377105114, + "tokens_seen": 1461583872 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028140421263791375, + "loss": 1.343, + "theoretical_loss": 3.552605936170292, + "tokens_seen": 1461649408 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028139418254764293, + "loss": 1.3348, + "theoretical_loss": 3.552591935433618, + "tokens_seen": 1461714944 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028138415245737217, + "loss": 1.3489, + "theoretical_loss": 3.5525779355004063, + "tokens_seen": 1461780480 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002813741223671013, + "loss": 1.3529, + "theoretical_loss": 3.552563936370575, + "tokens_seen": 1461846016 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028136409227683053, + "loss": 1.3748, + "theoretical_loss": 3.5525499380440424, + "tokens_seen": 1461911552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028135406218655966, + "loss": 1.411, + "theoretical_loss": 3.5525359405207255, + "tokens_seen": 1461977088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002813440320962889, + "loss": 1.4239, + "theoretical_loss": 3.552521943800543, + "tokens_seen": 1462042624 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002813340020060181, + "loss": 1.4184, + "theoretical_loss": 3.552507947883412, + "tokens_seen": 1462108160 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028132397191574725, + "loss": 1.4201, + "theoretical_loss": 3.552493952769252, + "tokens_seen": 1462173696 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028131394182547644, + "loss": 1.4345, + "theoretical_loss": 3.552479958457979, + "tokens_seen": 1462239232 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002813039117352056, + "loss": 1.3587, + "theoretical_loss": 3.552465964949512, + "tokens_seen": 1462304768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002812938816449348, + "loss": 1.3554, + "theoretical_loss": 3.5524519722437695, + "tokens_seen": 1462370304 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028128385155466403, + "loss": 1.4843, + "theoretical_loss": 3.5524379803406685, + "tokens_seen": 1462435840 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028127382146439316, + "loss": 1.4421, + "theoretical_loss": 3.552423989240127, + "tokens_seen": 1462501376 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002812637913741224, + "loss": 1.2601, + "theoretical_loss": 3.552409998942063, + "tokens_seen": 1462566912 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002812537612838516, + "loss": 1.3821, + "theoretical_loss": 3.5523960094463956, + "tokens_seen": 1462632448 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028124373119358076, + "loss": 1.4592, + "theoretical_loss": 3.5523820207530417, + "tokens_seen": 1462697984 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028123370110330994, + "loss": 1.3785, + "theoretical_loss": 3.55236803286192, + "tokens_seen": 1462763520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002812236710130391, + "loss": 1.4743, + "theoretical_loss": 3.552354045772948, + "tokens_seen": 1462829056 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002812136409227683, + "loss": 1.3675, + "theoretical_loss": 3.5523400594860437, + "tokens_seen": 1462894592 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028120361083249754, + "loss": 1.3703, + "theoretical_loss": 3.552326074001126, + "tokens_seen": 1462960128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028119358074222666, + "loss": 1.3344, + "theoretical_loss": 3.5523120893181126, + "tokens_seen": 1463025664 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002811835506519559, + "loss": 1.4404, + "theoretical_loss": 3.5522981054369214, + "tokens_seen": 1463091200 + }, + { + "epoch": 0.44, + "learning_rate": 0.000281173520561685, + "loss": 1.4731, + "theoretical_loss": 3.5522841223574706, + "tokens_seen": 1463156736 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028116349047141426, + "loss": 1.3944, + "theoretical_loss": 3.552270140079678, + "tokens_seen": 1463222272 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028115346038114344, + "loss": 1.394, + "theoretical_loss": 3.5522561586034627, + "tokens_seen": 1463287808 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002811434302908726, + "loss": 1.3552, + "theoretical_loss": 3.5522421779287416, + "tokens_seen": 1463353344 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002811334002006018, + "loss": 1.327, + "theoretical_loss": 3.5522281980554338, + "tokens_seen": 1463418880 + }, + { + "epoch": 0.44, + "learning_rate": 0.000281123370110331, + "loss": 1.4043, + "theoretical_loss": 3.5522142189834574, + "tokens_seen": 1463484416 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028111334002006017, + "loss": 1.4215, + "theoretical_loss": 3.5522002407127298, + "tokens_seen": 1463549952 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002811033099297894, + "loss": 1.3508, + "theoretical_loss": 3.55218626324317, + "tokens_seen": 1463615488 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028109327983951853, + "loss": 1.3405, + "theoretical_loss": 3.552172286574696, + "tokens_seen": 1463681024 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028108324974924776, + "loss": 1.3658, + "theoretical_loss": 3.5521583107072265, + "tokens_seen": 1463746560 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028107321965897695, + "loss": 1.3126, + "theoretical_loss": 3.5521443356406786, + "tokens_seen": 1463812096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002810631895687061, + "loss": 1.3301, + "theoretical_loss": 3.552130361374971, + "tokens_seen": 1463877632 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002810531594784353, + "loss": 1.39, + "theoretical_loss": 3.5521163879100226, + "tokens_seen": 1463943168 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002810431293881645, + "loss": 1.3825, + "theoretical_loss": 3.552102415245751, + "tokens_seen": 1464008704 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028103309929789367, + "loss": 1.3607, + "theoretical_loss": 3.5520884433820745, + "tokens_seen": 1464074240 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002810230692076229, + "loss": 1.4425, + "theoretical_loss": 3.552074472318912, + "tokens_seen": 1464139776 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028101303911735203, + "loss": 1.2893, + "theoretical_loss": 3.552060502056181, + "tokens_seen": 1464205312 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028100300902708127, + "loss": 1.2726, + "theoretical_loss": 3.5520465325938004, + "tokens_seen": 1464270848 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002809929789368104, + "loss": 1.3908, + "theoretical_loss": 3.5520325639316885, + "tokens_seen": 1464336384 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028098294884653963, + "loss": 1.3887, + "theoretical_loss": 3.552018596069763, + "tokens_seen": 1464401920 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002809729187562688, + "loss": 1.4283, + "theoretical_loss": 3.552004629007943, + "tokens_seen": 1464467456 + }, + { + "epoch": 0.44, + "learning_rate": 0.000280962888665998, + "loss": 1.3797, + "theoretical_loss": 3.551990662746147, + "tokens_seen": 1464532992 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028095285857572717, + "loss": 1.4472, + "theoretical_loss": 3.5519766972842923, + "tokens_seen": 1464598528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028094282848545635, + "loss": 1.4386, + "theoretical_loss": 3.5519627326222976, + "tokens_seen": 1464664064 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 534938, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.364405870437622, + "objective/train/theoretical_loss": 3.5519487687600826, + "objective/train/tokens_used": 1485189600, + "theoretical_loss": 3.5519487687600826, + "tokens_seen": 1464729600 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028093279839518553, + "loss": 1.3556, + "theoretical_loss": 3.5519487687600826, + "tokens_seen": 1464729600 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028092276830491477, + "loss": 1.4147, + "theoretical_loss": 3.5519348056975644, + "tokens_seen": 1464795136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002809127382146439, + "loss": 1.3803, + "theoretical_loss": 3.5519208434346616, + "tokens_seen": 1464860672 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028090270812437313, + "loss": 1.3161, + "theoretical_loss": 3.551906881971293, + "tokens_seen": 1464926208 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002808926780341023, + "loss": 1.4292, + "theoretical_loss": 3.551892921307377, + "tokens_seen": 1464991744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002808826479438315, + "loss": 1.3944, + "theoretical_loss": 3.551878961442832, + "tokens_seen": 1465057280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002808726178535607, + "loss": 1.4035, + "theoretical_loss": 3.5518650023775766, + "tokens_seen": 1465122816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028086258776328986, + "loss": 1.2914, + "theoretical_loss": 3.5518510441115287, + "tokens_seen": 1465188352 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028085255767301904, + "loss": 1.4313, + "theoretical_loss": 3.551837086644607, + "tokens_seen": 1465253888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002808425275827483, + "loss": 1.4601, + "theoretical_loss": 3.5518231299767304, + "tokens_seen": 1465319424 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002808324974924774, + "loss": 1.3136, + "theoretical_loss": 3.5518091741078175, + "tokens_seen": 1465384960 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028082246740220664, + "loss": 1.4632, + "theoretical_loss": 3.5517952190377864, + "tokens_seen": 1465450496 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028081243731193576, + "loss": 1.4498, + "theoretical_loss": 3.5517812647665563, + "tokens_seen": 1465516032 + }, + { + "epoch": 0.44, + "learning_rate": 0.000280802407221665, + "loss": 1.4919, + "theoretical_loss": 3.5517673112940447, + "tokens_seen": 1465581568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002807923771313942, + "loss": 1.3353, + "theoretical_loss": 3.5517533586201706, + "tokens_seen": 1465647104 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028078234704112336, + "loss": 1.3386, + "theoretical_loss": 3.5517394067448533, + "tokens_seen": 1465712640 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002807723169508526, + "loss": 1.4291, + "theoretical_loss": 3.5517254556680102, + "tokens_seen": 1465778176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002807622868605818, + "loss": 1.3094, + "theoretical_loss": 3.5517115053895614, + "tokens_seen": 1465843712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028075225677031096, + "loss": 1.4816, + "theoretical_loss": 3.5516975559094237, + "tokens_seen": 1465909248 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028074222668004014, + "loss": 1.477, + "theoretical_loss": 3.551683607227517, + "tokens_seen": 1465974784 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002807321965897693, + "loss": 1.3411, + "theoretical_loss": 3.55166965934376, + "tokens_seen": 1466040320 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002807221664994985, + "loss": 1.4281, + "theoretical_loss": 3.5516557122580705, + "tokens_seen": 1466105856 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028071213640922774, + "loss": 1.3102, + "theoretical_loss": 3.551641765970368, + "tokens_seen": 1466171392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028070210631895686, + "loss": 1.2672, + "theoretical_loss": 3.5516278204805705, + "tokens_seen": 1466236928 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002806920762286861, + "loss": 1.3942, + "theoretical_loss": 3.551613875788597, + "tokens_seen": 1466302464 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002806820461384152, + "loss": 1.368, + "theoretical_loss": 3.5515999318943665, + "tokens_seen": 1466368000 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028067201604814446, + "loss": 1.3194, + "theoretical_loss": 3.551585988797797, + "tokens_seen": 1466433536 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028066198595787364, + "loss": 1.4132, + "theoretical_loss": 3.5515720464988076, + "tokens_seen": 1466499072 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002806519558676028, + "loss": 1.4235, + "theoretical_loss": 3.551558104997318, + "tokens_seen": 1466564608 + }, + { + "epoch": 0.44, + "learning_rate": 0.000280641925777332, + "loss": 1.3026, + "theoretical_loss": 3.551544164293245, + "tokens_seen": 1466630144 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002806318956870612, + "loss": 1.4148, + "theoretical_loss": 3.551530224386509, + "tokens_seen": 1466695680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028062186559679037, + "loss": 1.3122, + "theoretical_loss": 3.551516285277028, + "tokens_seen": 1466761216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002806118355065196, + "loss": 1.4905, + "theoretical_loss": 3.5515023469647207, + "tokens_seen": 1466826752 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028060180541624873, + "loss": 1.4689, + "theoretical_loss": 3.5514884094495063, + "tokens_seen": 1466892288 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028059177532597796, + "loss": 1.3021, + "theoretical_loss": 3.5514744727313032, + "tokens_seen": 1466957824 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028058174523570715, + "loss": 1.3076, + "theoretical_loss": 3.5514605368100307, + "tokens_seen": 1467023360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002805717151454363, + "loss": 1.4089, + "theoretical_loss": 3.5514466016856074, + "tokens_seen": 1467088896 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002805616850551655, + "loss": 1.3093, + "theoretical_loss": 3.551432667357952, + "tokens_seen": 1467154432 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002805516549648947, + "loss": 1.3702, + "theoretical_loss": 3.551418733826984, + "tokens_seen": 1467219968 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028054162487462387, + "loss": 1.3537, + "theoretical_loss": 3.5514048010926214, + "tokens_seen": 1467285504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002805315947843531, + "loss": 1.3643, + "theoretical_loss": 3.5513908691547833, + "tokens_seen": 1467351040 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028052156469408223, + "loss": 1.3968, + "theoretical_loss": 3.551376938013389, + "tokens_seen": 1467416576 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028051153460381147, + "loss": 1.3494, + "theoretical_loss": 3.551363007668357, + "tokens_seen": 1467482112 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002805015045135406, + "loss": 1.3514, + "theoretical_loss": 3.551349078119606, + "tokens_seen": 1467547648 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028049147442326983, + "loss": 1.3204, + "theoretical_loss": 3.5513351493670555, + "tokens_seen": 1467613184 + }, + { + "epoch": 0.44, + "learning_rate": 0.000280481444332999, + "loss": 1.3219, + "theoretical_loss": 3.5513212214106247, + "tokens_seen": 1467678720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002804714142427282, + "loss": 1.4121, + "theoretical_loss": 3.5513072942502317, + "tokens_seen": 1467744256 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028046138415245737, + "loss": 1.4907, + "theoretical_loss": 3.551293367885796, + "tokens_seen": 1467809792 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028045135406218655, + "loss": 1.4316, + "theoretical_loss": 3.5512794423172362, + "tokens_seen": 1467875328 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028044132397191574, + "loss": 1.3664, + "theoretical_loss": 3.551265517544471, + "tokens_seen": 1467940864 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 535989, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3401269912719727, + "objective/train/theoretical_loss": 3.5512515935674207, + "objective/train/tokens_used": 1488466400, + "theoretical_loss": 3.5512515935674207, + "tokens_seen": 1468006400 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028043129388164497, + "loss": 1.391, + "theoretical_loss": 3.5512515935674207, + "tokens_seen": 1468006400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002804212637913741, + "loss": 1.3731, + "theoretical_loss": 3.551237670386003, + "tokens_seen": 1468071936 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028041123370110333, + "loss": 1.3407, + "theoretical_loss": 3.5512237480001376, + "tokens_seen": 1468137472 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002804012036108325, + "loss": 1.3447, + "theoretical_loss": 3.551209826409744, + "tokens_seen": 1468203008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002803911735205617, + "loss": 1.3988, + "theoretical_loss": 3.5511959056147395, + "tokens_seen": 1468268544 + }, + { + "epoch": 0.44, + "learning_rate": 0.0002803811434302909, + "loss": 1.3435, + "theoretical_loss": 3.5511819856150453, + "tokens_seen": 1468334080 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028037111334002006, + "loss": 1.4166, + "theoretical_loss": 3.5511680664105785, + "tokens_seen": 1468399616 + }, + { + "epoch": 0.44, + "learning_rate": 0.00028036108324974924, + "loss": 1.3246, + "theoretical_loss": 3.5511541480012596, + "tokens_seen": 1468465152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002803510531594785, + "loss": 1.3009, + "theoretical_loss": 3.551140230387007, + "tokens_seen": 1468530688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002803410230692076, + "loss": 1.3928, + "theoretical_loss": 3.55112631356774, + "tokens_seen": 1468596224 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028033099297893684, + "loss": 1.3631, + "theoretical_loss": 3.551112397543378, + "tokens_seen": 1468661760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028032096288866596, + "loss": 1.3957, + "theoretical_loss": 3.5510984823138396, + "tokens_seen": 1468727296 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002803109327983952, + "loss": 1.3706, + "theoretical_loss": 3.5510845678790446, + "tokens_seen": 1468792832 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002803009027081244, + "loss": 1.3337, + "theoretical_loss": 3.551070654238911, + "tokens_seen": 1468858368 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028029087261785356, + "loss": 1.4758, + "theoretical_loss": 3.551056741393359, + "tokens_seen": 1468923904 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028028084252758274, + "loss": 1.4486, + "theoretical_loss": 3.551042829342308, + "tokens_seen": 1468989440 + }, + { + "epoch": 0.45, + "learning_rate": 0.000280270812437312, + "loss": 1.2698, + "theoretical_loss": 3.5510289180856764, + "tokens_seen": 1469054976 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002802607823470411, + "loss": 1.4471, + "theoretical_loss": 3.5510150076233833, + "tokens_seen": 1469120512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028025075225677034, + "loss": 1.4252, + "theoretical_loss": 3.551001097955348, + "tokens_seen": 1469186048 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028024072216649947, + "loss": 1.3233, + "theoretical_loss": 3.550987189081491, + "tokens_seen": 1469251584 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002802306920762287, + "loss": 1.2996, + "theoretical_loss": 3.5509732810017303, + "tokens_seen": 1469317120 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002802206619859579, + "loss": 1.3122, + "theoretical_loss": 3.550959373715985, + "tokens_seen": 1469382656 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028021063189568706, + "loss": 1.4625, + "theoretical_loss": 3.5509454672241745, + "tokens_seen": 1469448192 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028020060180541624, + "loss": 1.3744, + "theoretical_loss": 3.5509315615262187, + "tokens_seen": 1469513728 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002801905717151454, + "loss": 1.3377, + "theoretical_loss": 3.550917656622037, + "tokens_seen": 1469579264 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002801805416248746, + "loss": 1.2265, + "theoretical_loss": 3.5509037525115477, + "tokens_seen": 1469644800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028017051153460384, + "loss": 1.4729, + "theoretical_loss": 3.5508898491946703, + "tokens_seen": 1469710336 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028016048144433297, + "loss": 1.2462, + "theoretical_loss": 3.550875946671325, + "tokens_seen": 1469775872 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002801504513540622, + "loss": 1.368, + "theoretical_loss": 3.55086204494143, + "tokens_seen": 1469841408 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028014042126379133, + "loss": 1.3032, + "theoretical_loss": 3.5508481440049056, + "tokens_seen": 1469906944 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028013039117352057, + "loss": 1.3721, + "theoretical_loss": 3.5508342438616705, + "tokens_seen": 1469972480 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028012036108324975, + "loss": 1.379, + "theoretical_loss": 3.5508203445116444, + "tokens_seen": 1470038016 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028011033099297893, + "loss": 1.2762, + "theoretical_loss": 3.5508064459547466, + "tokens_seen": 1470103552 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002801003009027081, + "loss": 1.5019, + "theoretical_loss": 3.5507925481908966, + "tokens_seen": 1470169088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028009027081243735, + "loss": 1.447, + "theoretical_loss": 3.5507786512200132, + "tokens_seen": 1470234624 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028008024072216647, + "loss": 1.3789, + "theoretical_loss": 3.5507647550420165, + "tokens_seen": 1470300160 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002800702106318957, + "loss": 1.3283, + "theoretical_loss": 3.550750859656826, + "tokens_seen": 1470365696 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028006018054162483, + "loss": 1.2904, + "theoretical_loss": 3.5507369650643605, + "tokens_seen": 1470431232 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028005015045135407, + "loss": 1.3814, + "theoretical_loss": 3.5507230712645397, + "tokens_seen": 1470496768 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028004012036108325, + "loss": 1.3111, + "theoretical_loss": 3.550709178257284, + "tokens_seen": 1470562304 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028003009027081243, + "loss": 1.3119, + "theoretical_loss": 3.5506952860425107, + "tokens_seen": 1470627840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028002006018054167, + "loss": 1.3439, + "theoretical_loss": 3.5506813946201414, + "tokens_seen": 1470693376 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002800100300902708, + "loss": 1.3888, + "theoretical_loss": 3.550667503990095, + "tokens_seen": 1470758912 + }, + { + "epoch": 0.45, + "learning_rate": 0.00028000000000000003, + "loss": 1.4073, + "theoretical_loss": 3.5506536141522904, + "tokens_seen": 1470824448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002799899699097292, + "loss": 1.4275, + "theoretical_loss": 3.5506397251066475, + "tokens_seen": 1470889984 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002799799398194584, + "loss": 1.4397, + "theoretical_loss": 3.5506258368530856, + "tokens_seen": 1470955520 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002799699097291876, + "loss": 1.3766, + "theoretical_loss": 3.5506119493915254, + "tokens_seen": 1471021056 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027995987963891675, + "loss": 1.4148, + "theoretical_loss": 3.5505980627218845, + "tokens_seen": 1471086592 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027994984954864594, + "loss": 1.3152, + "theoretical_loss": 3.550584176844084, + "tokens_seen": 1471152128 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027993981945837517, + "loss": 1.3022, + "theoretical_loss": 3.550570291758043, + "tokens_seen": 1471217664 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 537210, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3277535438537598, + "objective/train/theoretical_loss": 3.550556407463681, + "objective/train/tokens_used": 1491743200, + "theoretical_loss": 3.550556407463681, + "tokens_seen": 1471283200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002799297893681043, + "loss": 1.3525, + "theoretical_loss": 3.550556407463681, + "tokens_seen": 1471283200 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027991975927783353, + "loss": 1.3641, + "theoretical_loss": 3.5505425239609174, + "tokens_seen": 1471348736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002799097291875627, + "loss": 1.3978, + "theoretical_loss": 3.550528641249673, + "tokens_seen": 1471414272 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798996990972919, + "loss": 1.463, + "theoretical_loss": 3.5505147593298654, + "tokens_seen": 1471479808 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798896690070211, + "loss": 1.3572, + "theoretical_loss": 3.550500878201416, + "tokens_seen": 1471545344 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027987963891675026, + "loss": 1.4417, + "theoretical_loss": 3.5504869978642435, + "tokens_seen": 1471610880 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027986960882647944, + "loss": 1.3019, + "theoretical_loss": 3.550473118318268, + "tokens_seen": 1471676416 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798595787362087, + "loss": 1.3816, + "theoretical_loss": 3.550459239563409, + "tokens_seen": 1471741952 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798495486459378, + "loss": 1.3061, + "theoretical_loss": 3.550445361599586, + "tokens_seen": 1471807488 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027983951855566704, + "loss": 1.4234, + "theoretical_loss": 3.550431484426719, + "tokens_seen": 1471873024 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027982948846539616, + "loss": 1.2924, + "theoretical_loss": 3.550417608044728, + "tokens_seen": 1471938560 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798194583751254, + "loss": 1.412, + "theoretical_loss": 3.550403732453532, + "tokens_seen": 1472004096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002798094282848546, + "loss": 1.3805, + "theoretical_loss": 3.550389857653051, + "tokens_seen": 1472069632 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027979939819458376, + "loss": 1.3431, + "theoretical_loss": 3.550375983643205, + "tokens_seen": 1472135168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027978936810431294, + "loss": 1.3147, + "theoretical_loss": 3.5503621104239134, + "tokens_seen": 1472200704 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002797793380140422, + "loss": 1.3953, + "theoretical_loss": 3.5503482379950957, + "tokens_seen": 1472266240 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002797693079237713, + "loss": 1.4357, + "theoretical_loss": 3.5503343663566724, + "tokens_seen": 1472331776 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027975927783350054, + "loss": 1.4323, + "theoretical_loss": 3.550320495508563, + "tokens_seen": 1472397312 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027974924774322967, + "loss": 1.5003, + "theoretical_loss": 3.5503066254506876, + "tokens_seen": 1472462848 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002797392176529589, + "loss": 1.3384, + "theoretical_loss": 3.550292756182965, + "tokens_seen": 1472528384 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002797291875626881, + "loss": 1.4254, + "theoretical_loss": 3.550278887705316, + "tokens_seen": 1472593920 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027971915747241726, + "loss": 1.3302, + "theoretical_loss": 3.5502650200176604, + "tokens_seen": 1472659456 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027970912738214644, + "loss": 1.3341, + "theoretical_loss": 3.5502511531199175, + "tokens_seen": 1472724992 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002796990972918756, + "loss": 1.4311, + "theoretical_loss": 3.5502372870120076, + "tokens_seen": 1472790528 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002796890672016048, + "loss": 1.4484, + "theoretical_loss": 3.5502234216938504, + "tokens_seen": 1472856064 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027967903711133404, + "loss": 1.4903, + "theoretical_loss": 3.550209557165366, + "tokens_seen": 1472921600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027966900702106317, + "loss": 1.419, + "theoretical_loss": 3.550195693426473, + "tokens_seen": 1472987136 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002796589769307924, + "loss": 1.4075, + "theoretical_loss": 3.5501818304770936, + "tokens_seen": 1473052672 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027964894684052153, + "loss": 1.3416, + "theoretical_loss": 3.550167968317146, + "tokens_seen": 1473118208 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027963891675025077, + "loss": 1.3635, + "theoretical_loss": 3.550154106946551, + "tokens_seen": 1473183744 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027962888665997995, + "loss": 1.3336, + "theoretical_loss": 3.5501402463652276, + "tokens_seen": 1473249280 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027961885656970913, + "loss": 1.4277, + "theoretical_loss": 3.5501263865730968, + "tokens_seen": 1473314816 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002796088264794383, + "loss": 1.3385, + "theoretical_loss": 3.5501125275700782, + "tokens_seen": 1473380352 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027959879638916755, + "loss": 1.5329, + "theoretical_loss": 3.5500986693560916, + "tokens_seen": 1473445888 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027958876629889667, + "loss": 1.406, + "theoretical_loss": 3.5500848119310566, + "tokens_seen": 1473511424 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002795787362086259, + "loss": 1.4359, + "theoretical_loss": 3.5500709552948937, + "tokens_seen": 1473576960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027956870611835503, + "loss": 1.3034, + "theoretical_loss": 3.550057099447523, + "tokens_seen": 1473642496 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027955867602808427, + "loss": 1.3502, + "theoretical_loss": 3.5500432443888643, + "tokens_seen": 1473708032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027954864593781345, + "loss": 1.3359, + "theoretical_loss": 3.550029390118838, + "tokens_seen": 1473773568 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027953861584754263, + "loss": 1.4139, + "theoretical_loss": 3.550015536637364, + "tokens_seen": 1473839104 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002795285857572718, + "loss": 1.3674, + "theoretical_loss": 3.550001683944362, + "tokens_seen": 1473904640 + }, + { + "epoch": 0.45, + "learning_rate": 0.000279518555667001, + "loss": 1.3726, + "theoretical_loss": 3.5499878320397524, + "tokens_seen": 1473970176 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002795085255767302, + "loss": 1.3994, + "theoretical_loss": 3.549973980923455, + "tokens_seen": 1474035712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794984954864594, + "loss": 1.35, + "theoretical_loss": 3.54996013059539, + "tokens_seen": 1474101248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027948846539618854, + "loss": 1.3914, + "theoretical_loss": 3.549946281055478, + "tokens_seen": 1474166784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794784353059178, + "loss": 1.3518, + "theoretical_loss": 3.5499324323036374, + "tokens_seen": 1474232320 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794684052156469, + "loss": 1.3589, + "theoretical_loss": 3.5499185843397907, + "tokens_seen": 1474297856 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027945837512537614, + "loss": 1.3576, + "theoretical_loss": 3.549904737163857, + "tokens_seen": 1474363392 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794483450351053, + "loss": 1.4407, + "theoretical_loss": 3.549890890775756, + "tokens_seen": 1474428928 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794383149448345, + "loss": 1.4889, + "theoretical_loss": 3.5498770451754083, + "tokens_seen": 1474494464 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 538469, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3359849452972412, + "objective/train/theoretical_loss": 3.549863200362734, + "objective/train/tokens_used": 1495020000, + "theoretical_loss": 3.549863200362734, + "tokens_seen": 1474560000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794282848545637, + "loss": 1.3152, + "theoretical_loss": 3.549863200362734, + "tokens_seen": 1474560000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002794182547642929, + "loss": 1.2569, + "theoretical_loss": 3.5498493563376536, + "tokens_seen": 1474625536 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027940822467402204, + "loss": 1.3671, + "theoretical_loss": 3.5498355131000867, + "tokens_seen": 1474691072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793981945837513, + "loss": 1.267, + "theoretical_loss": 3.549821670649954, + "tokens_seen": 1474756608 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793881644934804, + "loss": 1.4353, + "theoretical_loss": 3.5498078289871753, + "tokens_seen": 1474822144 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027937813440320964, + "loss": 1.395, + "theoretical_loss": 3.549793988111671, + "tokens_seen": 1474887680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793681043129388, + "loss": 1.4243, + "theoretical_loss": 3.5497801480233617, + "tokens_seen": 1474953216 + }, + { + "epoch": 0.45, + "learning_rate": 0.000279358074222668, + "loss": 1.3308, + "theoretical_loss": 3.5497663087221674, + "tokens_seen": 1475018752 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793480441323972, + "loss": 1.4155, + "theoretical_loss": 3.5497524702080083, + "tokens_seen": 1475084288 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027933801404212636, + "loss": 1.435, + "theoretical_loss": 3.5497386324808042, + "tokens_seen": 1475149824 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027932798395185554, + "loss": 1.3185, + "theoretical_loss": 3.5497247955404765, + "tokens_seen": 1475215360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793179538615848, + "loss": 1.3235, + "theoretical_loss": 3.5497109593869447, + "tokens_seen": 1475280896 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002793079237713139, + "loss": 1.4348, + "theoretical_loss": 3.5496971240201294, + "tokens_seen": 1475346432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027929789368104314, + "loss": 1.4983, + "theoretical_loss": 3.5496832894399506, + "tokens_seen": 1475411968 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027928786359077227, + "loss": 1.4131, + "theoretical_loss": 3.5496694556463293, + "tokens_seen": 1475477504 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002792778335005015, + "loss": 1.3093, + "theoretical_loss": 3.549655622639185, + "tokens_seen": 1475543040 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027926780341023074, + "loss": 1.3082, + "theoretical_loss": 3.5496417904184385, + "tokens_seen": 1475608576 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027925777331995987, + "loss": 1.4003, + "theoretical_loss": 3.54962795898401, + "tokens_seen": 1475674112 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002792477432296891, + "loss": 1.3876, + "theoretical_loss": 3.5496141283358202, + "tokens_seen": 1475739648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002792377131394183, + "loss": 1.2348, + "theoretical_loss": 3.5496002984737896, + "tokens_seen": 1475805184 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027922768304914746, + "loss": 1.2044, + "theoretical_loss": 3.5495864693978376, + "tokens_seen": 1475870720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027921765295887664, + "loss": 1.4571, + "theoretical_loss": 3.5495726411078863, + "tokens_seen": 1475936256 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002792076228686058, + "loss": 1.3275, + "theoretical_loss": 3.5495588136038547, + "tokens_seen": 1476001792 + }, + { + "epoch": 0.45, + "learning_rate": 0.000279197592778335, + "loss": 1.4801, + "theoretical_loss": 3.549544986885664, + "tokens_seen": 1476067328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027918756268806424, + "loss": 1.3731, + "theoretical_loss": 3.549531160953234, + "tokens_seen": 1476132864 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027917753259779337, + "loss": 1.408, + "theoretical_loss": 3.5495173358064855, + "tokens_seen": 1476198400 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002791675025075226, + "loss": 1.3686, + "theoretical_loss": 3.549503511445339, + "tokens_seen": 1476263936 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027915747241725173, + "loss": 1.3491, + "theoretical_loss": 3.5494896878697153, + "tokens_seen": 1476329472 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027914744232698097, + "loss": 1.3437, + "theoretical_loss": 3.549475865079534, + "tokens_seen": 1476395008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027913741223671015, + "loss": 1.3415, + "theoretical_loss": 3.5494620430747164, + "tokens_seen": 1476460544 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027912738214643933, + "loss": 1.467, + "theoretical_loss": 3.549448221855183, + "tokens_seen": 1476526080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002791173520561685, + "loss": 1.4397, + "theoretical_loss": 3.549434401420854, + "tokens_seen": 1476591616 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027910732196589775, + "loss": 1.5024, + "theoretical_loss": 3.5494205817716504, + "tokens_seen": 1476657152 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027909729187562687, + "loss": 1.4637, + "theoretical_loss": 3.549406762907492, + "tokens_seen": 1476722688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002790872617853561, + "loss": 1.3547, + "theoretical_loss": 3.5493929448283, + "tokens_seen": 1476788224 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027907723169508523, + "loss": 1.4241, + "theoretical_loss": 3.549379127533995, + "tokens_seen": 1476853760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027906720160481447, + "loss": 1.3534, + "theoretical_loss": 3.5493653110244976, + "tokens_seen": 1476919296 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027905717151454365, + "loss": 1.3544, + "theoretical_loss": 3.549351495299728, + "tokens_seen": 1476984832 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027904714142427283, + "loss": 1.2975, + "theoretical_loss": 3.5493376803596064, + "tokens_seen": 1477050368 + }, + { + "epoch": 0.45, + "learning_rate": 0.000279037111334002, + "loss": 1.3633, + "theoretical_loss": 3.5493238662040545, + "tokens_seen": 1477115904 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002790270812437312, + "loss": 1.4207, + "theoretical_loss": 3.5493100528329924, + "tokens_seen": 1477181440 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002790170511534604, + "loss": 1.4503, + "theoretical_loss": 3.549296240246341, + "tokens_seen": 1477246976 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002790070210631896, + "loss": 1.3858, + "theoretical_loss": 3.5492824284440205, + "tokens_seen": 1477312512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027899699097291874, + "loss": 1.4471, + "theoretical_loss": 3.5492686174259522, + "tokens_seen": 1477378048 + }, + { + "epoch": 0.45, + "learning_rate": 0.000278986960882648, + "loss": 1.3778, + "theoretical_loss": 3.549254807192056, + "tokens_seen": 1477443584 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789769307923771, + "loss": 1.3184, + "theoretical_loss": 3.5492409977422534, + "tokens_seen": 1477509120 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027896690070210634, + "loss": 1.3812, + "theoretical_loss": 3.549227189076465, + "tokens_seen": 1477574656 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789568706118355, + "loss": 1.3485, + "theoretical_loss": 3.54921338119461, + "tokens_seen": 1477640192 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789468405215647, + "loss": 1.2726, + "theoretical_loss": 3.549199574096612, + "tokens_seen": 1477705728 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789368104312939, + "loss": 1.3778, + "theoretical_loss": 3.549185767782389, + "tokens_seen": 1477771264 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 539614, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3246023654937744, + "objective/train/theoretical_loss": 3.549171962251863, + "objective/train/tokens_used": 1498296800, + "theoretical_loss": 3.549171962251863, + "tokens_seen": 1477836800 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789267803410231, + "loss": 1.3013, + "theoretical_loss": 3.549171962251863, + "tokens_seen": 1477836800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027891675025075224, + "loss": 1.3504, + "theoretical_loss": 3.5491581575049547, + "tokens_seen": 1477902336 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002789067201604815, + "loss": 1.392, + "theoretical_loss": 3.549144353541585, + "tokens_seen": 1477967872 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002788966900702106, + "loss": 1.4326, + "theoretical_loss": 3.5491305503616744, + "tokens_seen": 1478033408 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027888665997993984, + "loss": 1.4127, + "theoretical_loss": 3.5491167479651438, + "tokens_seen": 1478098944 + }, + { + "epoch": 0.45, + "learning_rate": 0.000278876629889669, + "loss": 1.4074, + "theoretical_loss": 3.5491029463519137, + "tokens_seen": 1478164480 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002788665997993982, + "loss": 1.3783, + "theoretical_loss": 3.5490891455219056, + "tokens_seen": 1478230016 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002788565697091274, + "loss": 1.2934, + "theoretical_loss": 3.54907534547504, + "tokens_seen": 1478295552 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027884653961885656, + "loss": 1.3408, + "theoretical_loss": 3.549061546211237, + "tokens_seen": 1478361088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027883650952858574, + "loss": 1.3826, + "theoretical_loss": 3.5490477477304188, + "tokens_seen": 1478426624 + }, + { + "epoch": 0.45, + "learning_rate": 0.000278826479438315, + "loss": 1.3816, + "theoretical_loss": 3.5490339500325057, + "tokens_seen": 1478492160 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002788164493480441, + "loss": 1.3954, + "theoretical_loss": 3.5490201531174184, + "tokens_seen": 1478557696 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027880641925777334, + "loss": 1.3251, + "theoretical_loss": 3.549006356985078, + "tokens_seen": 1478623232 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027879638916750247, + "loss": 1.4048, + "theoretical_loss": 3.548992561635405, + "tokens_seen": 1478688768 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002787863590772317, + "loss": 1.3708, + "theoretical_loss": 3.548978767068321, + "tokens_seen": 1478754304 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002787763289869609, + "loss": 1.2956, + "theoretical_loss": 3.548964973283746, + "tokens_seen": 1478819840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027876629889669007, + "loss": 1.464, + "theoretical_loss": 3.548951180281602, + "tokens_seen": 1478885376 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027875626880641925, + "loss": 1.3987, + "theoretical_loss": 3.548937388061809, + "tokens_seen": 1478950912 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002787462387161485, + "loss": 1.521, + "theoretical_loss": 3.548923596624288, + "tokens_seen": 1479016448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002787362086258776, + "loss": 1.3396, + "theoretical_loss": 3.548909805968961, + "tokens_seen": 1479081984 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027872617853560685, + "loss": 1.3992, + "theoretical_loss": 3.548896016095748, + "tokens_seen": 1479147520 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027871614844533597, + "loss": 1.4818, + "theoretical_loss": 3.5488822270045706, + "tokens_seen": 1479213056 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002787061183550652, + "loss": 1.4708, + "theoretical_loss": 3.5488684386953495, + "tokens_seen": 1479278592 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002786960882647944, + "loss": 1.3479, + "theoretical_loss": 3.5488546511680057, + "tokens_seen": 1479344128 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027868605817452357, + "loss": 1.3917, + "theoretical_loss": 3.54884086442246, + "tokens_seen": 1479409664 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027867602808425275, + "loss": 1.3505, + "theoretical_loss": 3.548827078458634, + "tokens_seen": 1479475200 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027866599799398193, + "loss": 1.4058, + "theoretical_loss": 3.5488132932764485, + "tokens_seen": 1479540736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002786559679037111, + "loss": 1.2513, + "theoretical_loss": 3.5487995088758244, + "tokens_seen": 1479606272 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027864593781344035, + "loss": 1.4063, + "theoretical_loss": 3.548785725256683, + "tokens_seen": 1479671808 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002786359077231695, + "loss": 1.4111, + "theoretical_loss": 3.548771942418945, + "tokens_seen": 1479737344 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002786258776328987, + "loss": 1.4901, + "theoretical_loss": 3.5487581603625324, + "tokens_seen": 1479802880 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002786158475426279, + "loss": 1.429, + "theoretical_loss": 3.548744379087365, + "tokens_seen": 1479868416 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027860581745235707, + "loss": 1.4125, + "theoretical_loss": 3.548730598593365, + "tokens_seen": 1479933952 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027859578736208625, + "loss": 1.3463, + "theoretical_loss": 3.548716818880453, + "tokens_seen": 1479999488 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027858575727181544, + "loss": 1.4231, + "theoretical_loss": 3.5487030399485504, + "tokens_seen": 1480065024 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002785757271815446, + "loss": 1.4982, + "theoretical_loss": 3.5486892617975783, + "tokens_seen": 1480130560 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027856569709127385, + "loss": 1.4022, + "theoretical_loss": 3.5486754844274575, + "tokens_seen": 1480196096 + }, + { + "epoch": 0.45, + "learning_rate": 0.000278555667001003, + "loss": 1.3487, + "theoretical_loss": 3.54866170783811, + "tokens_seen": 1480261632 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002785456369107322, + "loss": 1.3652, + "theoretical_loss": 3.548647932029456, + "tokens_seen": 1480327168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027853560682046134, + "loss": 1.4103, + "theoretical_loss": 3.548634157001417, + "tokens_seen": 1480392704 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002785255767301906, + "loss": 1.4004, + "theoretical_loss": 3.548620382753915, + "tokens_seen": 1480458240 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002785155466399198, + "loss": 1.535, + "theoretical_loss": 3.54860660928687, + "tokens_seen": 1480523776 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027850551654964894, + "loss": 1.3491, + "theoretical_loss": 3.5485928366002044, + "tokens_seen": 1480589312 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784954864593782, + "loss": 1.4136, + "theoretical_loss": 3.5485790646938384, + "tokens_seen": 1480654848 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784854563691073, + "loss": 1.4075, + "theoretical_loss": 3.548565293567694, + "tokens_seen": 1480720384 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027847542627883654, + "loss": 1.3906, + "theoretical_loss": 3.5485515232216924, + "tokens_seen": 1480785920 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784653961885657, + "loss": 1.4444, + "theoretical_loss": 3.5485377536557543, + "tokens_seen": 1480851456 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784553660982949, + "loss": 1.3388, + "theoretical_loss": 3.5485239848698016, + "tokens_seen": 1480916992 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784453360080241, + "loss": 1.4629, + "theoretical_loss": 3.5485102168637557, + "tokens_seen": 1480982528 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784353059177533, + "loss": 1.4121, + "theoretical_loss": 3.548496449637537, + "tokens_seen": 1481048064 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 540822, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4360809326171875, + "objective/train/theoretical_loss": 3.548482683191067, + "objective/train/tokens_used": 1501573600, + "theoretical_loss": 3.548482683191067, + "tokens_seen": 1481113600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027842527582748244, + "loss": 1.465, + "theoretical_loss": 3.548482683191067, + "tokens_seen": 1481113600 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784152457372117, + "loss": 1.3296, + "theoretical_loss": 3.5484689175242687, + "tokens_seen": 1481179136 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002784052156469408, + "loss": 1.4978, + "theoretical_loss": 3.548455152637062, + "tokens_seen": 1481244672 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027839518555667004, + "loss": 1.3365, + "theoretical_loss": 3.5484413885293673, + "tokens_seen": 1481310208 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002783851554663992, + "loss": 1.4136, + "theoretical_loss": 3.548427625201108, + "tokens_seen": 1481375744 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002783751253761284, + "loss": 1.4335, + "theoretical_loss": 3.548413862652205, + "tokens_seen": 1481441280 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002783650952858576, + "loss": 1.3928, + "theoretical_loss": 3.5484001008825787, + "tokens_seen": 1481506816 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027835506519558676, + "loss": 1.3756, + "theoretical_loss": 3.5483863398921516, + "tokens_seen": 1481572352 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027834503510531594, + "loss": 1.3295, + "theoretical_loss": 3.5483725796808443, + "tokens_seen": 1481637888 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002783350050150452, + "loss": 1.4404, + "theoretical_loss": 3.5483588202485787, + "tokens_seen": 1481703424 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002783249749247743, + "loss": 1.3606, + "theoretical_loss": 3.548345061595276, + "tokens_seen": 1481768960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027831494483450354, + "loss": 1.4163, + "theoretical_loss": 3.5483313037208575, + "tokens_seen": 1481834496 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027830491474423267, + "loss": 1.3106, + "theoretical_loss": 3.5483175466252455, + "tokens_seen": 1481900032 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782948846539619, + "loss": 1.4808, + "theoretical_loss": 3.5483037903083607, + "tokens_seen": 1481965568 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782848545636911, + "loss": 1.4361, + "theoretical_loss": 3.5482900347701247, + "tokens_seen": 1482031104 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027827482447342027, + "loss": 1.313, + "theoretical_loss": 3.5482762800104592, + "tokens_seen": 1482096640 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027826479438314945, + "loss": 1.4507, + "theoretical_loss": 3.5482625260292853, + "tokens_seen": 1482162176 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782547642928787, + "loss": 1.3632, + "theoretical_loss": 3.5482487728265255, + "tokens_seen": 1482227712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782447342026078, + "loss": 1.4128, + "theoretical_loss": 3.5482350204020996, + "tokens_seen": 1482293248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027823470411233705, + "loss": 1.4342, + "theoretical_loss": 3.548221268755931, + "tokens_seen": 1482358784 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027822467402206617, + "loss": 1.324, + "theoretical_loss": 3.54820751788794, + "tokens_seen": 1482424320 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782146439317954, + "loss": 1.3442, + "theoretical_loss": 3.548193767798049, + "tokens_seen": 1482489856 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002782046138415246, + "loss": 1.4645, + "theoretical_loss": 3.548180018486179, + "tokens_seen": 1482555392 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027819458375125377, + "loss": 1.3229, + "theoretical_loss": 3.548166269952252, + "tokens_seen": 1482620928 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027818455366098295, + "loss": 1.4107, + "theoretical_loss": 3.548152522196189, + "tokens_seen": 1482686464 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027817452357071213, + "loss": 1.3382, + "theoretical_loss": 3.548138775217912, + "tokens_seen": 1482752000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002781644934804413, + "loss": 1.3667, + "theoretical_loss": 3.548125029017343, + "tokens_seen": 1482817536 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027815446339017055, + "loss": 1.3557, + "theoretical_loss": 3.5481112835944026, + "tokens_seen": 1482883072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002781444332998997, + "loss": 1.4257, + "theoretical_loss": 3.5480975389490137, + "tokens_seen": 1482948608 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002781344032096289, + "loss": 1.3978, + "theoretical_loss": 3.548083795081097, + "tokens_seen": 1483014144 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002781243731193581, + "loss": 1.4004, + "theoretical_loss": 3.5480700519905746, + "tokens_seen": 1483079680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002781143430290873, + "loss": 1.4256, + "theoretical_loss": 3.548056309677368, + "tokens_seen": 1483145216 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027810431293881645, + "loss": 1.4701, + "theoretical_loss": 3.548042568141399, + "tokens_seen": 1483210752 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027809428284854564, + "loss": 1.3726, + "theoretical_loss": 3.5480288273825895, + "tokens_seen": 1483276288 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002780842527582748, + "loss": 1.5676, + "theoretical_loss": 3.5480150874008602, + "tokens_seen": 1483341824 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027807422266800405, + "loss": 1.4229, + "theoretical_loss": 3.5480013481961343, + "tokens_seen": 1483407360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002780641925777332, + "loss": 1.4467, + "theoretical_loss": 3.5479876097683327, + "tokens_seen": 1483472896 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002780541624874624, + "loss": 1.4347, + "theoretical_loss": 3.5479738721173772, + "tokens_seen": 1483538432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027804413239719154, + "loss": 1.3748, + "theoretical_loss": 3.5479601352431898, + "tokens_seen": 1483603968 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002780341023069208, + "loss": 1.4644, + "theoretical_loss": 3.547946399145692, + "tokens_seen": 1483669504 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027802407221664996, + "loss": 1.3982, + "theoretical_loss": 3.547932663824806, + "tokens_seen": 1483735040 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027801404212637914, + "loss": 1.4821, + "theoretical_loss": 3.547918929280453, + "tokens_seen": 1483800576 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002780040120361083, + "loss": 1.5916, + "theoretical_loss": 3.5479051955125556, + "tokens_seen": 1483866112 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779939819458375, + "loss": 1.4295, + "theoretical_loss": 3.5478914625210343, + "tokens_seen": 1483931648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779839518555667, + "loss": 1.3545, + "theoretical_loss": 3.547877730305812, + "tokens_seen": 1483997184 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779739217652959, + "loss": 1.3987, + "theoretical_loss": 3.547863998866811, + "tokens_seen": 1484062720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027796389167502504, + "loss": 1.2766, + "theoretical_loss": 3.5478502682039514, + "tokens_seen": 1484128256 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779538615847543, + "loss": 1.3433, + "theoretical_loss": 3.5478365383171564, + "tokens_seen": 1484193792 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027794383149448346, + "loss": 1.3948, + "theoretical_loss": 3.547822809206348, + "tokens_seen": 1484259328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027793380140421264, + "loss": 1.4523, + "theoretical_loss": 3.5478090808714473, + "tokens_seen": 1484324864 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 542078, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.388979196548462, + "objective/train/theoretical_loss": 3.5477953533123765, + "objective/train/tokens_used": 1504850400, + "theoretical_loss": 3.5477953533123765, + "tokens_seen": 1484390400 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779237713139418, + "loss": 1.4015, + "theoretical_loss": 3.5477953533123765, + "tokens_seen": 1484390400 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277913741223671, + "loss": 1.3635, + "theoretical_loss": 3.5477816265290576, + "tokens_seen": 1484455936 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002779037111334002, + "loss": 1.3477, + "theoretical_loss": 3.5477679005214124, + "tokens_seen": 1484521472 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778936810431294, + "loss": 1.3963, + "theoretical_loss": 3.547754175289363, + "tokens_seen": 1484587008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027788365095285855, + "loss": 1.3898, + "theoretical_loss": 3.5477404508328307, + "tokens_seen": 1484652544 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778736208625878, + "loss": 1.3889, + "theoretical_loss": 3.5477267271517388, + "tokens_seen": 1484718080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778635907723169, + "loss": 1.448, + "theoretical_loss": 3.547713004246008, + "tokens_seen": 1484783616 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027785356068204614, + "loss": 1.4797, + "theoretical_loss": 3.547699282115561, + "tokens_seen": 1484849152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778435305917753, + "loss": 1.3453, + "theoretical_loss": 3.547685560760319, + "tokens_seen": 1484914688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778335005015045, + "loss": 1.4002, + "theoretical_loss": 3.5476718401802048, + "tokens_seen": 1484980224 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002778234704112337, + "loss": 1.4938, + "theoretical_loss": 3.54765812037514, + "tokens_seen": 1485045760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027781344032096287, + "loss": 1.4783, + "theoretical_loss": 3.547644401345047, + "tokens_seen": 1485111296 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027780341023069205, + "loss": 1.4465, + "theoretical_loss": 3.5476306830898467, + "tokens_seen": 1485176832 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002777933801404213, + "loss": 1.4418, + "theoretical_loss": 3.547616965609463, + "tokens_seen": 1485242368 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002777833500501504, + "loss": 1.3741, + "theoretical_loss": 3.5476032489038163, + "tokens_seen": 1485307904 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027777331995987965, + "loss": 1.2916, + "theoretical_loss": 3.5475895329728298, + "tokens_seen": 1485373440 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002777632898696089, + "loss": 1.3347, + "theoretical_loss": 3.5475758178164245, + "tokens_seen": 1485438976 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277753259779338, + "loss": 1.3984, + "theoretical_loss": 3.5475621034345233, + "tokens_seen": 1485504512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027774322968906725, + "loss": 1.4175, + "theoretical_loss": 3.547548389827048, + "tokens_seen": 1485570048 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027773319959879637, + "loss": 1.4077, + "theoretical_loss": 3.547534676993921, + "tokens_seen": 1485635584 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002777231695085256, + "loss": 1.4479, + "theoretical_loss": 3.547520964935064, + "tokens_seen": 1485701120 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002777131394182548, + "loss": 1.3723, + "theoretical_loss": 3.5475072536503998, + "tokens_seen": 1485766656 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027770310932798397, + "loss": 1.2761, + "theoretical_loss": 3.547493543139849, + "tokens_seen": 1485832192 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027769307923771315, + "loss": 1.431, + "theoretical_loss": 3.5474798334033357, + "tokens_seen": 1485897728 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027768304914744233, + "loss": 1.4112, + "theoretical_loss": 3.547466124440781, + "tokens_seen": 1485963264 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002776730190571715, + "loss": 1.3861, + "theoretical_loss": 3.5474524162521073, + "tokens_seen": 1486028800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027766298896690075, + "loss": 1.3784, + "theoretical_loss": 3.5474387088372366, + "tokens_seen": 1486094336 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002776529588766299, + "loss": 1.3959, + "theoretical_loss": 3.547425002196091, + "tokens_seen": 1486159872 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002776429287863591, + "loss": 1.413, + "theoretical_loss": 3.547411296328593, + "tokens_seen": 1486225408 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002776328986960883, + "loss": 1.3445, + "theoretical_loss": 3.5473975912346654, + "tokens_seen": 1486290944 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002776228686058175, + "loss": 1.4628, + "theoretical_loss": 3.547383886914229, + "tokens_seen": 1486356480 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027761283851554665, + "loss": 1.3179, + "theoretical_loss": 3.5473701833672076, + "tokens_seen": 1486422016 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027760280842527584, + "loss": 1.397, + "theoretical_loss": 3.547356480593522, + "tokens_seen": 1486487552 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277592778335005, + "loss": 1.3803, + "theoretical_loss": 3.5473427785930958, + "tokens_seen": 1486553088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027758274824473425, + "loss": 1.3694, + "theoretical_loss": 3.5473290773658506, + "tokens_seen": 1486618624 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002775727181544634, + "loss": 1.4116, + "theoretical_loss": 3.5473153769117083, + "tokens_seen": 1486684160 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002775626880641926, + "loss": 1.4383, + "theoretical_loss": 3.5473016772305916, + "tokens_seen": 1486749696 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027755265797392174, + "loss": 1.3053, + "theoretical_loss": 3.5472879783224234, + "tokens_seen": 1486815232 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277542627883651, + "loss": 1.3025, + "theoretical_loss": 3.547274280187125, + "tokens_seen": 1486880768 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027753259779338016, + "loss": 1.3772, + "theoretical_loss": 3.5472605828246193, + "tokens_seen": 1486946304 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027752256770310934, + "loss": 1.4197, + "theoretical_loss": 3.547246886234829, + "tokens_seen": 1487011840 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002775125376128385, + "loss": 1.4749, + "theoretical_loss": 3.5472331904176757, + "tokens_seen": 1487077376 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002775025075225677, + "loss": 1.3764, + "theoretical_loss": 3.547219495373082, + "tokens_seen": 1487142912 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774924774322969, + "loss": 1.3792, + "theoretical_loss": 3.5472058011009704, + "tokens_seen": 1487208448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774824473420261, + "loss": 1.3777, + "theoretical_loss": 3.547192107601263, + "tokens_seen": 1487273984 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027747241725175524, + "loss": 1.2332, + "theoretical_loss": 3.5471784148738825, + "tokens_seen": 1487339520 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774623871614845, + "loss": 1.3543, + "theoretical_loss": 3.5471647229187515, + "tokens_seen": 1487405056 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027745235707121366, + "loss": 1.3543, + "theoretical_loss": 3.547151031735792, + "tokens_seen": 1487470592 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027744232698094284, + "loss": 1.3908, + "theoretical_loss": 3.5471373413249268, + "tokens_seen": 1487536128 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277432296890672, + "loss": 1.2856, + "theoretical_loss": 3.547123651686078, + "tokens_seen": 1487601664 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 543335, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3396685123443604, + "objective/train/theoretical_loss": 3.5471099628191682, + "objective/train/tokens_used": 1508127200, + "theoretical_loss": 3.5471099628191682, + "tokens_seen": 1487667200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774222668004012, + "loss": 1.3652, + "theoretical_loss": 3.5471099628191682, + "tokens_seen": 1487667200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774122367101304, + "loss": 1.3781, + "theoretical_loss": 3.5470962747241197, + "tokens_seen": 1487732736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002774022066198596, + "loss": 1.3272, + "theoretical_loss": 3.5470825874008556, + "tokens_seen": 1487798272 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027739217652958875, + "loss": 1.4572, + "theoretical_loss": 3.5470689008492977, + "tokens_seen": 1487863808 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277382146439318, + "loss": 1.5253, + "theoretical_loss": 3.547055215069369, + "tokens_seen": 1487929344 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002773721163490471, + "loss": 1.3786, + "theoretical_loss": 3.5470415300609917, + "tokens_seen": 1487994880 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027736208625877634, + "loss": 1.489, + "theoretical_loss": 3.547027845824088, + "tokens_seen": 1488060416 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002773520561685055, + "loss": 1.3993, + "theoretical_loss": 3.5470141623585807, + "tokens_seen": 1488125952 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002773420260782347, + "loss": 1.347, + "theoretical_loss": 3.5470004796643932, + "tokens_seen": 1488191488 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002773319959879639, + "loss": 1.4637, + "theoretical_loss": 3.5469867977414467, + "tokens_seen": 1488257024 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027732196589769307, + "loss": 1.4041, + "theoretical_loss": 3.546973116589665, + "tokens_seen": 1488322560 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027731193580742225, + "loss": 1.3895, + "theoretical_loss": 3.5469594362089696, + "tokens_seen": 1488388096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002773019057171515, + "loss": 1.4121, + "theoretical_loss": 3.5469457565992837, + "tokens_seen": 1488453632 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002772918756268806, + "loss": 1.3655, + "theoretical_loss": 3.54693207776053, + "tokens_seen": 1488519168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027728184553660985, + "loss": 1.3234, + "theoretical_loss": 3.546918399692631, + "tokens_seen": 1488584704 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027727181544633903, + "loss": 1.396, + "theoretical_loss": 3.5469047223955084, + "tokens_seen": 1488650240 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002772617853560682, + "loss": 1.3899, + "theoretical_loss": 3.5468910458690863, + "tokens_seen": 1488715776 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002772517552657974, + "loss": 1.3601, + "theoretical_loss": 3.5468773701132865, + "tokens_seen": 1488781312 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027724172517552657, + "loss": 1.3803, + "theoretical_loss": 3.5468636951280317, + "tokens_seen": 1488846848 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027723169508525575, + "loss": 1.4038, + "theoretical_loss": 3.5468500209132454, + "tokens_seen": 1488912384 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277221664994985, + "loss": 1.4465, + "theoretical_loss": 3.546836347468849, + "tokens_seen": 1488977920 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002772116349047141, + "loss": 1.3699, + "theoretical_loss": 3.546822674794766, + "tokens_seen": 1489043456 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027720160481444335, + "loss": 1.37, + "theoretical_loss": 3.5468090028909187, + "tokens_seen": 1489108992 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771915747241725, + "loss": 1.3284, + "theoretical_loss": 3.54679533175723, + "tokens_seen": 1489174528 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771815446339017, + "loss": 1.4068, + "theoretical_loss": 3.5467816613936227, + "tokens_seen": 1489240064 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771715145436309, + "loss": 1.31, + "theoretical_loss": 3.54676799180002, + "tokens_seen": 1489305600 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771614844533601, + "loss": 1.3519, + "theoretical_loss": 3.5467543229763434, + "tokens_seen": 1489371136 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027715145436308926, + "loss": 1.39, + "theoretical_loss": 3.546740654922517, + "tokens_seen": 1489436672 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771414242728185, + "loss": 1.331, + "theoretical_loss": 3.546726987638462, + "tokens_seen": 1489502208 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771313941825476, + "loss": 1.42, + "theoretical_loss": 3.5467133211241033, + "tokens_seen": 1489567744 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027712136409227685, + "loss": 1.3686, + "theoretical_loss": 3.5466996553793617, + "tokens_seen": 1489633280 + }, + { + "epoch": 0.45, + "learning_rate": 0.000277111334002006, + "loss": 1.4624, + "theoretical_loss": 3.5466859904041614, + "tokens_seen": 1489698816 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002771013039117352, + "loss": 1.3193, + "theoretical_loss": 3.5466723261984243, + "tokens_seen": 1489764352 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770912738214644, + "loss": 1.4114, + "theoretical_loss": 3.5466586627620735, + "tokens_seen": 1489829888 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770812437311936, + "loss": 1.4423, + "theoretical_loss": 3.5466450000950323, + "tokens_seen": 1489895424 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027707121364092276, + "loss": 1.4102, + "theoretical_loss": 3.5466313381972228, + "tokens_seen": 1489960960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027706118355065194, + "loss": 1.3457, + "theoretical_loss": 3.5466176770685682, + "tokens_seen": 1490026496 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770511534603811, + "loss": 1.4072, + "theoretical_loss": 3.546604016708992, + "tokens_seen": 1490092032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027704112337011036, + "loss": 1.358, + "theoretical_loss": 3.546590357118416, + "tokens_seen": 1490157568 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770310932798395, + "loss": 1.2611, + "theoretical_loss": 3.546576698296763, + "tokens_seen": 1490223104 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770210631895687, + "loss": 1.2879, + "theoretical_loss": 3.546563040243957, + "tokens_seen": 1490288640 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770110330992979, + "loss": 1.3231, + "theoretical_loss": 3.5465493829599204, + "tokens_seen": 1490354176 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002770010030090271, + "loss": 1.2933, + "theoretical_loss": 3.546535726444576, + "tokens_seen": 1490419712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769909729187563, + "loss": 1.3213, + "theoretical_loss": 3.546522070697847, + "tokens_seen": 1490485248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027698094282848544, + "loss": 1.343, + "theoretical_loss": 3.546508415719656, + "tokens_seen": 1490550784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769709127382147, + "loss": 1.4443, + "theoretical_loss": 3.5464947615099263, + "tokens_seen": 1490616320 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027696088264794386, + "loss": 1.2792, + "theoretical_loss": 3.5464811080685807, + "tokens_seen": 1490681856 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027695085255767304, + "loss": 1.3734, + "theoretical_loss": 3.5464674553955415, + "tokens_seen": 1490747392 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769408224674022, + "loss": 1.4598, + "theoretical_loss": 3.546453803490733, + "tokens_seen": 1490812928 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769307923771314, + "loss": 1.3659, + "theoretical_loss": 3.5464401523540783, + "tokens_seen": 1490878464 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 544526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4075956344604492, + "objective/train/theoretical_loss": 3.546426501985499, + "objective/train/tokens_used": 1511404000, + "theoretical_loss": 3.546426501985499, + "tokens_seen": 1490944000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769207622868606, + "loss": 1.3794, + "theoretical_loss": 3.546426501985499, + "tokens_seen": 1490944000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002769107321965898, + "loss": 1.3936, + "theoretical_loss": 3.5464128523849183, + "tokens_seen": 1491009536 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027690070210631895, + "loss": 1.3468, + "theoretical_loss": 3.5463992035522605, + "tokens_seen": 1491075072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768906720160482, + "loss": 1.3363, + "theoretical_loss": 3.546385555487448, + "tokens_seen": 1491140608 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768806419257773, + "loss": 1.4067, + "theoretical_loss": 3.5463719081904035, + "tokens_seen": 1491206144 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027687061183550655, + "loss": 1.3175, + "theoretical_loss": 3.54635826166105, + "tokens_seen": 1491271680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768605817452357, + "loss": 1.3905, + "theoretical_loss": 3.5463446158993115, + "tokens_seen": 1491337216 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768505516549649, + "loss": 1.2461, + "theoretical_loss": 3.5463309709051103, + "tokens_seen": 1491402752 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768405215646941, + "loss": 1.4029, + "theoretical_loss": 3.54631732667837, + "tokens_seen": 1491468288 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027683049147442327, + "loss": 1.3927, + "theoretical_loss": 3.546303683219013, + "tokens_seen": 1491533824 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027682046138415245, + "loss": 1.3894, + "theoretical_loss": 3.546290040526963, + "tokens_seen": 1491599360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768104312938817, + "loss": 1.3025, + "theoretical_loss": 3.546276398602143, + "tokens_seen": 1491664896 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002768004012036108, + "loss": 1.3431, + "theoretical_loss": 3.546262757444476, + "tokens_seen": 1491730432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027679037111334005, + "loss": 1.415, + "theoretical_loss": 3.5462491170538852, + "tokens_seen": 1491795968 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027678034102306923, + "loss": 1.4381, + "theoretical_loss": 3.5462354774302947, + "tokens_seen": 1491861504 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002767703109327984, + "loss": 1.3447, + "theoretical_loss": 3.546221838573626, + "tokens_seen": 1491927040 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002767602808425276, + "loss": 1.4713, + "theoretical_loss": 3.546208200483804, + "tokens_seen": 1491992576 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027675025075225677, + "loss": 1.4602, + "theoretical_loss": 3.54619456316075, + "tokens_seen": 1492058112 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027674022066198595, + "loss": 1.3759, + "theoretical_loss": 3.546180926604389, + "tokens_seen": 1492123648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002767301905717152, + "loss": 1.4062, + "theoretical_loss": 3.546167290814643, + "tokens_seen": 1492189184 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002767201604814443, + "loss": 1.3187, + "theoretical_loss": 3.5461536557914357, + "tokens_seen": 1492254720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027671013039117355, + "loss": 1.3185, + "theoretical_loss": 3.546140021534691, + "tokens_seen": 1492320256 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002767001003009027, + "loss": 1.3574, + "theoretical_loss": 3.546126388044331, + "tokens_seen": 1492385792 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766900702106319, + "loss": 1.535, + "theoretical_loss": 3.54611275532028, + "tokens_seen": 1492451328 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766800401203611, + "loss": 1.3897, + "theoretical_loss": 3.5460991233624606, + "tokens_seen": 1492516864 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766700100300903, + "loss": 1.2992, + "theoretical_loss": 3.5460854921707963, + "tokens_seen": 1492582400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027665997993981946, + "loss": 1.4888, + "theoretical_loss": 3.54607186174521, + "tokens_seen": 1492647936 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766499498495487, + "loss": 1.2489, + "theoretical_loss": 3.5460582320856258, + "tokens_seen": 1492713472 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766399197592778, + "loss": 1.3595, + "theoretical_loss": 3.5460446031919663, + "tokens_seen": 1492779008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027662988966900705, + "loss": 1.415, + "theoretical_loss": 3.546030975064155, + "tokens_seen": 1492844544 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766198595787362, + "loss": 1.3075, + "theoretical_loss": 3.546017347702116, + "tokens_seen": 1492910080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002766098294884654, + "loss": 1.4068, + "theoretical_loss": 3.5460037211057718, + "tokens_seen": 1492975616 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765997993981946, + "loss": 1.3708, + "theoretical_loss": 3.545990095275046, + "tokens_seen": 1493041152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765897693079238, + "loss": 1.4101, + "theoretical_loss": 3.5459764702098617, + "tokens_seen": 1493106688 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027657973921765296, + "loss": 1.4148, + "theoretical_loss": 3.545962845910143, + "tokens_seen": 1493172224 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027656970912738214, + "loss": 1.2575, + "theoretical_loss": 3.5459492223758127, + "tokens_seen": 1493237760 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765596790371113, + "loss": 1.395, + "theoretical_loss": 3.5459355996067945, + "tokens_seen": 1493303296 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027654964894684056, + "loss": 1.3894, + "theoretical_loss": 3.5459219776030118, + "tokens_seen": 1493368832 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765396188565697, + "loss": 1.4144, + "theoretical_loss": 3.5459083563643876, + "tokens_seen": 1493434368 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765295887662989, + "loss": 1.3601, + "theoretical_loss": 3.5458947358908466, + "tokens_seen": 1493499904 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027651955867602805, + "loss": 1.3477, + "theoretical_loss": 3.5458811161823105, + "tokens_seen": 1493565440 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002765095285857573, + "loss": 1.2467, + "theoretical_loss": 3.545867497238704, + "tokens_seen": 1493630976 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027649949849548646, + "loss": 1.2854, + "theoretical_loss": 3.5458538790599503, + "tokens_seen": 1493696512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027648946840521564, + "loss": 1.3366, + "theoretical_loss": 3.5458402616459725, + "tokens_seen": 1493762048 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002764794383149448, + "loss": 1.368, + "theoretical_loss": 3.545826644996694, + "tokens_seen": 1493827584 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027646940822467406, + "loss": 1.3034, + "theoretical_loss": 3.5458130291120398, + "tokens_seen": 1493893120 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002764593781344032, + "loss": 1.3938, + "theoretical_loss": 3.545799413991932, + "tokens_seen": 1493958656 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002764493480441324, + "loss": 1.4022, + "theoretical_loss": 3.545785799636294, + "tokens_seen": 1494024192 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027643931795386155, + "loss": 1.3011, + "theoretical_loss": 3.54577218604505, + "tokens_seen": 1494089728 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002764292878635908, + "loss": 1.3715, + "theoretical_loss": 3.545758573218124, + "tokens_seen": 1494155264 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 545793, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2787797451019287, + "objective/train/theoretical_loss": 3.545744961155438, + "objective/train/tokens_used": 1514680800, + "theoretical_loss": 3.545744961155438, + "tokens_seen": 1494220800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027641925777331997, + "loss": 1.2673, + "theoretical_loss": 3.545744961155438, + "tokens_seen": 1494220800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027640922768304915, + "loss": 1.4417, + "theoretical_loss": 3.545731349856917, + "tokens_seen": 1494286336 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027639919759277833, + "loss": 1.3465, + "theoretical_loss": 3.545717739322484, + "tokens_seen": 1494351872 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002763891675025075, + "loss": 1.3224, + "theoretical_loss": 3.545704129552063, + "tokens_seen": 1494417408 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002763791374122367, + "loss": 1.323, + "theoretical_loss": 3.545690520545577, + "tokens_seen": 1494482944 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002763691073219659, + "loss": 1.3796, + "theoretical_loss": 3.5456769123029503, + "tokens_seen": 1494548480 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027635907723169505, + "loss": 1.4144, + "theoretical_loss": 3.5456633048241057, + "tokens_seen": 1494614016 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002763490471414243, + "loss": 1.4109, + "theoretical_loss": 3.545649698108968, + "tokens_seen": 1494679552 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002763390170511534, + "loss": 1.3063, + "theoretical_loss": 3.5456360921574595, + "tokens_seen": 1494745088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027632898696088265, + "loss": 1.3597, + "theoretical_loss": 3.5456224869695045, + "tokens_seen": 1494810624 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027631895687061183, + "loss": 1.4095, + "theoretical_loss": 3.5456088825450274, + "tokens_seen": 1494876160 + }, + { + "epoch": 0.45, + "learning_rate": 0.000276308926780341, + "loss": 1.3824, + "theoretical_loss": 3.5455952788839507, + "tokens_seen": 1494941696 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002762988966900702, + "loss": 1.3349, + "theoretical_loss": 3.545581675986199, + "tokens_seen": 1495007232 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027628886659979943, + "loss": 1.2441, + "theoretical_loss": 3.5455680738516953, + "tokens_seen": 1495072768 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027627883650952856, + "loss": 1.3737, + "theoretical_loss": 3.545554472480364, + "tokens_seen": 1495138304 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002762688064192578, + "loss": 1.398, + "theoretical_loss": 3.545540871872128, + "tokens_seen": 1495203840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027625877632898697, + "loss": 1.3791, + "theoretical_loss": 3.5455272720269115, + "tokens_seen": 1495269376 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027624874623871615, + "loss": 1.3343, + "theoretical_loss": 3.5455136729446384, + "tokens_seen": 1495334912 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002762387161484454, + "loss": 1.3872, + "theoretical_loss": 3.5455000746252328, + "tokens_seen": 1495400448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002762286860581745, + "loss": 1.4108, + "theoretical_loss": 3.5454864770686174, + "tokens_seen": 1495465984 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027621865596790375, + "loss": 1.4305, + "theoretical_loss": 3.5454728802747164, + "tokens_seen": 1495531520 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002762086258776329, + "loss": 1.3525, + "theoretical_loss": 3.5454592842434547, + "tokens_seen": 1495597056 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761985957873621, + "loss": 1.4357, + "theoretical_loss": 3.5454456889747545, + "tokens_seen": 1495662592 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761885656970913, + "loss": 1.4706, + "theoretical_loss": 3.5454320944685405, + "tokens_seen": 1495728128 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761785356068205, + "loss": 1.3128, + "theoretical_loss": 3.545418500724736, + "tokens_seen": 1495793664 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027616850551654966, + "loss": 1.3939, + "theoretical_loss": 3.545404907743266, + "tokens_seen": 1495859200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761584754262789, + "loss": 1.3592, + "theoretical_loss": 3.545391315524053, + "tokens_seen": 1495924736 + }, + { + "epoch": 0.45, + "learning_rate": 0.000276148445336008, + "loss": 1.4537, + "theoretical_loss": 3.5453777240670217, + "tokens_seen": 1495990272 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027613841524573725, + "loss": 1.3842, + "theoretical_loss": 3.5453641333720953, + "tokens_seen": 1496055808 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761283851554664, + "loss": 1.2835, + "theoretical_loss": 3.5453505434391985, + "tokens_seen": 1496121344 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761183550651956, + "loss": 1.3815, + "theoretical_loss": 3.5453369542682545, + "tokens_seen": 1496186880 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002761083249749248, + "loss": 1.3815, + "theoretical_loss": 3.545323365859187, + "tokens_seen": 1496252416 + }, + { + "epoch": 0.45, + "learning_rate": 0.000276098294884654, + "loss": 1.3432, + "theoretical_loss": 3.545309778211921, + "tokens_seen": 1496317952 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027608826479438316, + "loss": 1.4209, + "theoretical_loss": 3.54529619132638, + "tokens_seen": 1496383488 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027607823470411234, + "loss": 1.4881, + "theoretical_loss": 3.5452826052024875, + "tokens_seen": 1496449024 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002760682046138415, + "loss": 1.344, + "theoretical_loss": 3.545269019840167, + "tokens_seen": 1496514560 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027605817452357076, + "loss": 1.5791, + "theoretical_loss": 3.545255435239344, + "tokens_seen": 1496580096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002760481444332999, + "loss": 1.4461, + "theoretical_loss": 3.5452418513999415, + "tokens_seen": 1496645632 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002760381143430291, + "loss": 1.2943, + "theoretical_loss": 3.5452282683218836, + "tokens_seen": 1496711168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027602808425275825, + "loss": 1.3375, + "theoretical_loss": 3.5452146860050946, + "tokens_seen": 1496776704 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002760180541624875, + "loss": 1.3817, + "theoretical_loss": 3.5452011044494975, + "tokens_seen": 1496842240 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027600802407221666, + "loss": 1.4648, + "theoretical_loss": 3.5451875236550174, + "tokens_seen": 1496907776 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027599799398194584, + "loss": 1.3718, + "theoretical_loss": 3.5451739436215783, + "tokens_seen": 1496973312 + }, + { + "epoch": 0.45, + "learning_rate": 0.000275987963891675, + "loss": 1.3474, + "theoretical_loss": 3.545160364349104, + "tokens_seen": 1497038848 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027597793380140426, + "loss": 1.2076, + "theoretical_loss": 3.5451467858375176, + "tokens_seen": 1497104384 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002759679037111334, + "loss": 1.3982, + "theoretical_loss": 3.5451332080867446, + "tokens_seen": 1497169920 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002759578736208626, + "loss": 1.454, + "theoretical_loss": 3.5451196310967084, + "tokens_seen": 1497235456 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027594784353059175, + "loss": 1.3902, + "theoretical_loss": 3.5451060548673334, + "tokens_seen": 1497300992 + }, + { + "epoch": 0.45, + "learning_rate": 0.000275937813440321, + "loss": 1.5294, + "theoretical_loss": 3.545092479398543, + "tokens_seen": 1497366528 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027592778335005017, + "loss": 1.4185, + "theoretical_loss": 3.545078904690262, + "tokens_seen": 1497432064 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 547043, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2956106662750244, + "objective/train/theoretical_loss": 3.545065330742414, + "objective/train/tokens_used": 1517957600, + "theoretical_loss": 3.545065330742414, + "tokens_seen": 1497497600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027591775325977935, + "loss": 1.3344, + "theoretical_loss": 3.545065330742414, + "tokens_seen": 1497497600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027590772316950853, + "loss": 1.3617, + "theoretical_loss": 3.545051757554924, + "tokens_seen": 1497563136 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758976930792377, + "loss": 1.4369, + "theoretical_loss": 3.545038185127715, + "tokens_seen": 1497628672 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758876629889669, + "loss": 1.4066, + "theoretical_loss": 3.545024613460712, + "tokens_seen": 1497694208 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758776328986961, + "loss": 1.4659, + "theoretical_loss": 3.545011042553839, + "tokens_seen": 1497759744 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027586760280842525, + "loss": 1.4123, + "theoretical_loss": 3.5449974724070197, + "tokens_seen": 1497825280 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758575727181545, + "loss": 1.4238, + "theoretical_loss": 3.544983903020178, + "tokens_seen": 1497890816 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758475426278836, + "loss": 1.2713, + "theoretical_loss": 3.5449703343932395, + "tokens_seen": 1497956352 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027583751253761285, + "loss": 1.3892, + "theoretical_loss": 3.5449567665261275, + "tokens_seen": 1498021888 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027582748244734203, + "loss": 1.3631, + "theoretical_loss": 3.544943199418766, + "tokens_seen": 1498087424 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758174523570712, + "loss": 1.4066, + "theoretical_loss": 3.5449296330710798, + "tokens_seen": 1498152960 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002758074222668004, + "loss": 1.4574, + "theoretical_loss": 3.5449160674829927, + "tokens_seen": 1498218496 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027579739217652963, + "loss": 1.5302, + "theoretical_loss": 3.544902502654429, + "tokens_seen": 1498284032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027578736208625876, + "loss": 1.3898, + "theoretical_loss": 3.5448889385853137, + "tokens_seen": 1498349568 + }, + { + "epoch": 0.45, + "learning_rate": 0.000275777331995988, + "loss": 1.3802, + "theoretical_loss": 3.5448753752755695, + "tokens_seen": 1498415104 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002757673019057171, + "loss": 1.354, + "theoretical_loss": 3.544861812725122, + "tokens_seen": 1498480640 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027575727181544635, + "loss": 1.3785, + "theoretical_loss": 3.544848250933895, + "tokens_seen": 1498546176 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027574724172517554, + "loss": 1.3851, + "theoretical_loss": 3.5448346899018133, + "tokens_seen": 1498611712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002757372116349047, + "loss": 1.3043, + "theoretical_loss": 3.5448211296288004, + "tokens_seen": 1498677248 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002757271815446339, + "loss": 1.3016, + "theoretical_loss": 3.5448075701147808, + "tokens_seen": 1498742784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002757171514543631, + "loss": 1.4576, + "theoretical_loss": 3.54479401135968, + "tokens_seen": 1498808320 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027570712136409226, + "loss": 1.4766, + "theoretical_loss": 3.5447804533634204, + "tokens_seen": 1498873856 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756970912738215, + "loss": 1.3472, + "theoretical_loss": 3.5447668961259278, + "tokens_seen": 1498939392 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756870611835506, + "loss": 1.3488, + "theoretical_loss": 3.5447533396471256, + "tokens_seen": 1499004928 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027567703109327986, + "loss": 1.3711, + "theoretical_loss": 3.5447397839269392, + "tokens_seen": 1499070464 + }, + { + "epoch": 0.45, + "learning_rate": 0.000275667001003009, + "loss": 1.4599, + "theoretical_loss": 3.5447262289652928, + "tokens_seen": 1499136000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756569709127382, + "loss": 1.3719, + "theoretical_loss": 3.54471267476211, + "tokens_seen": 1499201536 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756469408224674, + "loss": 1.4605, + "theoretical_loss": 3.5446991213173153, + "tokens_seen": 1499267072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756369107321966, + "loss": 1.3335, + "theoretical_loss": 3.544685568630834, + "tokens_seen": 1499332608 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027562688064192576, + "loss": 1.3962, + "theoretical_loss": 3.54467201670259, + "tokens_seen": 1499398144 + }, + { + "epoch": 0.45, + "learning_rate": 0.000275616850551655, + "loss": 1.2977, + "theoretical_loss": 3.5446584655325073, + "tokens_seen": 1499463680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002756068204613841, + "loss": 1.4151, + "theoretical_loss": 3.5446449151205113, + "tokens_seen": 1499529216 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027559679037111336, + "loss": 1.397, + "theoretical_loss": 3.544631365466526, + "tokens_seen": 1499594752 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002755867602808425, + "loss": 1.4244, + "theoretical_loss": 3.5446178165704754, + "tokens_seen": 1499660288 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002755767301905717, + "loss": 1.3453, + "theoretical_loss": 3.544604268432285, + "tokens_seen": 1499725824 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002755667001003009, + "loss": 1.4227, + "theoretical_loss": 3.544590721051878, + "tokens_seen": 1499791360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002755566700100301, + "loss": 1.3803, + "theoretical_loss": 3.5445771744291803, + "tokens_seen": 1499856896 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027554663991975927, + "loss": 1.2971, + "theoretical_loss": 3.5445636285641156, + "tokens_seen": 1499922432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027553660982948845, + "loss": 1.3892, + "theoretical_loss": 3.544550083456608, + "tokens_seen": 1499987968 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027552657973921763, + "loss": 1.5163, + "theoretical_loss": 3.5445365391065833, + "tokens_seen": 1500053504 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027551654964894686, + "loss": 1.3759, + "theoretical_loss": 3.544522995513965, + "tokens_seen": 1500119040 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027550651955867604, + "loss": 1.4587, + "theoretical_loss": 3.5445094526786782, + "tokens_seen": 1500184576 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002754964894684052, + "loss": 1.422, + "theoretical_loss": 3.5444959106006473, + "tokens_seen": 1500250112 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027548645937813446, + "loss": 1.5075, + "theoretical_loss": 3.5444823692797973, + "tokens_seen": 1500315648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002754764292878636, + "loss": 1.3501, + "theoretical_loss": 3.5444688287160515, + "tokens_seen": 1500381184 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002754663991975928, + "loss": 1.3362, + "theoretical_loss": 3.5444552889093357, + "tokens_seen": 1500446720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027545636910732195, + "loss": 1.4954, + "theoretical_loss": 3.5444417498595744, + "tokens_seen": 1500512256 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002754463390170512, + "loss": 1.4043, + "theoretical_loss": 3.5444282115666916, + "tokens_seen": 1500577792 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027543630892678037, + "loss": 1.372, + "theoretical_loss": 3.5444146740306124, + "tokens_seen": 1500643328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027542627883650955, + "loss": 1.2798, + "theoretical_loss": 3.5444011372512616, + "tokens_seen": 1500708864 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 548178, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3959131240844727, + "objective/train/theoretical_loss": 3.5443876012285638, + "objective/train/tokens_used": 1521234400, + "theoretical_loss": 3.5443876012285638, + "tokens_seen": 1500774400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027541624874623873, + "loss": 1.3021, + "theoretical_loss": 3.5443876012285638, + "tokens_seen": 1500774400 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002754062186559679, + "loss": 1.3722, + "theoretical_loss": 3.544374065962443, + "tokens_seen": 1500839936 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753961885656971, + "loss": 1.345, + "theoretical_loss": 3.544360531452825, + "tokens_seen": 1500905472 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753861584754263, + "loss": 1.2905, + "theoretical_loss": 3.544346997699633, + "tokens_seen": 1500971008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027537612838515545, + "loss": 1.4093, + "theoretical_loss": 3.5443334647027935, + "tokens_seen": 1501036544 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753660982948847, + "loss": 1.4218, + "theoretical_loss": 3.54431993246223, + "tokens_seen": 1501102080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753560682046138, + "loss": 1.4702, + "theoretical_loss": 3.5443064009778675, + "tokens_seen": 1501167616 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027534603811434305, + "loss": 1.3484, + "theoretical_loss": 3.54429287024963, + "tokens_seen": 1501233152 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027533600802407223, + "loss": 1.3689, + "theoretical_loss": 3.544279340277444, + "tokens_seen": 1501298688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753259779338014, + "loss": 1.3845, + "theoretical_loss": 3.5442658110612326, + "tokens_seen": 1501364224 + }, + { + "epoch": 0.45, + "learning_rate": 0.0002753159478435306, + "loss": 1.3778, + "theoretical_loss": 3.5442522826009215, + "tokens_seen": 1501429760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00027530591775325983, + "loss": 1.3133, + "theoretical_loss": 3.544238754896435, + "tokens_seen": 1501495296 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027529588766298896, + "loss": 1.5106, + "theoretical_loss": 3.544225227947698, + "tokens_seen": 1501560832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752858575727182, + "loss": 1.3955, + "theoretical_loss": 3.544211701754636, + "tokens_seen": 1501626368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752758274824473, + "loss": 1.4884, + "theoretical_loss": 3.5441981763171726, + "tokens_seen": 1501691904 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027526579739217655, + "loss": 1.448, + "theoretical_loss": 3.5441846516352333, + "tokens_seen": 1501757440 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027525576730190574, + "loss": 1.3269, + "theoretical_loss": 3.5441711277087427, + "tokens_seen": 1501822976 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752457372116349, + "loss": 1.4915, + "theoretical_loss": 3.5441576045376255, + "tokens_seen": 1501888512 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752357071213641, + "loss": 1.3489, + "theoretical_loss": 3.5441440821218073, + "tokens_seen": 1501954048 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752256770310933, + "loss": 1.3476, + "theoretical_loss": 3.544130560461212, + "tokens_seen": 1502019584 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027521564694082246, + "loss": 1.3885, + "theoretical_loss": 3.5441170395557653, + "tokens_seen": 1502085120 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002752056168505517, + "loss": 1.3753, + "theoretical_loss": 3.5441035194053914, + "tokens_seen": 1502150656 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751955867602808, + "loss": 1.361, + "theoretical_loss": 3.5440900000100157, + "tokens_seen": 1502216192 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027518555667001006, + "loss": 1.3474, + "theoretical_loss": 3.5440764813695633, + "tokens_seen": 1502281728 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751755265797392, + "loss": 1.3706, + "theoretical_loss": 3.5440629634839578, + "tokens_seen": 1502347264 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751654964894684, + "loss": 1.353, + "theoretical_loss": 3.544049446353126, + "tokens_seen": 1502412800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751554663991976, + "loss": 1.3649, + "theoretical_loss": 3.5440359299769915, + "tokens_seen": 1502478336 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751454363089268, + "loss": 1.5394, + "theoretical_loss": 3.5440224143554793, + "tokens_seen": 1502543872 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027513540621865596, + "loss": 1.3914, + "theoretical_loss": 3.544008899488515, + "tokens_seen": 1502609408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751253761283852, + "loss": 1.2987, + "theoretical_loss": 3.5439953853760233, + "tokens_seen": 1502674944 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002751153460381143, + "loss": 1.3235, + "theoretical_loss": 3.543981872017929, + "tokens_seen": 1502740480 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027510531594784356, + "loss": 1.4037, + "theoretical_loss": 3.5439683594141576, + "tokens_seen": 1502806016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750952858575727, + "loss": 1.4331, + "theoretical_loss": 3.543954847564633, + "tokens_seen": 1502871552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750852557673019, + "loss": 1.3644, + "theoretical_loss": 3.543941336469282, + "tokens_seen": 1502937088 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750752256770311, + "loss": 1.4173, + "theoretical_loss": 3.5439278261280274, + "tokens_seen": 1503002624 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750651955867603, + "loss": 1.438, + "theoretical_loss": 3.5439143165407962, + "tokens_seen": 1503068160 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027505516549648947, + "loss": 1.394, + "theoretical_loss": 3.5439008077075123, + "tokens_seen": 1503133696 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027504513540621865, + "loss": 1.3826, + "theoretical_loss": 3.5438872996281012, + "tokens_seen": 1503199232 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027503510531594783, + "loss": 1.4012, + "theoretical_loss": 3.543873792302488, + "tokens_seen": 1503264768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027502507522567706, + "loss": 1.2731, + "theoretical_loss": 3.5438602857305974, + "tokens_seen": 1503330304 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750150451354062, + "loss": 1.3728, + "theoretical_loss": 3.5438467799123545, + "tokens_seen": 1503395840 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002750050150451354, + "loss": 1.3053, + "theoretical_loss": 3.5438332748476853, + "tokens_seen": 1503461376 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027499498495486455, + "loss": 1.4146, + "theoretical_loss": 3.5438197705365138, + "tokens_seen": 1503526912 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002749849548645938, + "loss": 1.3764, + "theoretical_loss": 3.5438062669787653, + "tokens_seen": 1503592448 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027497492477432297, + "loss": 1.2693, + "theoretical_loss": 3.5437927641743654, + "tokens_seen": 1503657984 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027496489468405215, + "loss": 1.4158, + "theoretical_loss": 3.5437792621232393, + "tokens_seen": 1503723520 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027495486459378133, + "loss": 1.3889, + "theoretical_loss": 3.5437657608253117, + "tokens_seen": 1503789056 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027494483450351057, + "loss": 1.2773, + "theoretical_loss": 3.543752260280508, + "tokens_seen": 1503854592 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002749348044132397, + "loss": 1.4831, + "theoretical_loss": 3.543738760488753, + "tokens_seen": 1503920128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027492477432296893, + "loss": 1.4596, + "theoretical_loss": 3.543725261449972, + "tokens_seen": 1503985664 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 549351, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3666679859161377, + "objective/train/theoretical_loss": 3.543711763164091, + "objective/train/tokens_used": 1524511200, + "theoretical_loss": 3.543711763164091, + "tokens_seen": 1504051200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027491474423269806, + "loss": 1.3786, + "theoretical_loss": 3.543711763164091, + "tokens_seen": 1504051200 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002749047141424273, + "loss": 1.3075, + "theoretical_loss": 3.5436982656310345, + "tokens_seen": 1504116736 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027489468405215647, + "loss": 1.3553, + "theoretical_loss": 3.5436847688507274, + "tokens_seen": 1504182272 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027488465396188565, + "loss": 1.307, + "theoretical_loss": 3.5436712728230955, + "tokens_seen": 1504247808 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027487462387161483, + "loss": 1.3503, + "theoretical_loss": 3.543657777548064, + "tokens_seen": 1504313344 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274864593781344, + "loss": 1.277, + "theoretical_loss": 3.543644283025558, + "tokens_seen": 1504378880 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002748545636910732, + "loss": 1.3945, + "theoretical_loss": 3.5436307892555026, + "tokens_seen": 1504444416 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027484453360080243, + "loss": 1.2893, + "theoretical_loss": 3.543617296237823, + "tokens_seen": 1504509952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027483450351053156, + "loss": 1.3969, + "theoretical_loss": 3.543603803972445, + "tokens_seen": 1504575488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002748244734202608, + "loss": 1.3456, + "theoretical_loss": 3.543590312459294, + "tokens_seen": 1504641024 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027481444332999, + "loss": 1.3335, + "theoretical_loss": 3.543576821698294, + "tokens_seen": 1504706560 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027480441323971916, + "loss": 1.3143, + "theoretical_loss": 3.543563331689372, + "tokens_seen": 1504772096 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027479438314944834, + "loss": 1.3858, + "theoretical_loss": 3.5435498424324523, + "tokens_seen": 1504837632 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002747843530591775, + "loss": 1.3798, + "theoretical_loss": 3.5435363539274602, + "tokens_seen": 1504903168 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027477432296890675, + "loss": 1.4158, + "theoretical_loss": 3.543522866174322, + "tokens_seen": 1504968704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027476429287863594, + "loss": 1.4324, + "theoretical_loss": 3.5435093791729617, + "tokens_seen": 1505034240 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002747542627883651, + "loss": 1.4812, + "theoretical_loss": 3.543495892923305, + "tokens_seen": 1505099776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002747442326980943, + "loss": 1.3202, + "theoretical_loss": 3.5434824074252784, + "tokens_seen": 1505165312 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002747342026078235, + "loss": 1.498, + "theoretical_loss": 3.543468922678806, + "tokens_seen": 1505230848 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027472417251755266, + "loss": 1.3038, + "theoretical_loss": 3.543455438683814, + "tokens_seen": 1505296384 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002747141424272819, + "loss": 1.2409, + "theoretical_loss": 3.5434419554402274, + "tokens_seen": 1505361920 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274704112337011, + "loss": 1.3978, + "theoretical_loss": 3.543428472947972, + "tokens_seen": 1505427456 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027469408224674026, + "loss": 1.3674, + "theoretical_loss": 3.5434149912069723, + "tokens_seen": 1505492992 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746840521564694, + "loss": 1.263, + "theoretical_loss": 3.543401510217155, + "tokens_seen": 1505558528 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746740220661986, + "loss": 1.3977, + "theoretical_loss": 3.5433880299784444, + "tokens_seen": 1505624064 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746639919759278, + "loss": 1.418, + "theoretical_loss": 3.5433745504907668, + "tokens_seen": 1505689600 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274653961885657, + "loss": 1.3888, + "theoretical_loss": 3.543361071754047, + "tokens_seen": 1505755136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027464393179538616, + "loss": 1.3661, + "theoretical_loss": 3.5433475937682113, + "tokens_seen": 1505820672 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746339017051154, + "loss": 1.3358, + "theoretical_loss": 3.543334116533184, + "tokens_seen": 1505886208 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746238716148445, + "loss": 1.4236, + "theoretical_loss": 3.543320640048892, + "tokens_seen": 1505951744 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027461384152457376, + "loss": 1.3862, + "theoretical_loss": 3.54330716431526, + "tokens_seen": 1506017280 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002746038114343029, + "loss": 1.2152, + "theoretical_loss": 3.543293689332214, + "tokens_seen": 1506082816 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002745937813440321, + "loss": 1.4797, + "theoretical_loss": 3.5432802150996787, + "tokens_seen": 1506148352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002745837512537613, + "loss": 1.2793, + "theoretical_loss": 3.5432667416175803, + "tokens_seen": 1506213888 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002745737211634905, + "loss": 1.3605, + "theoretical_loss": 3.5432532688858442, + "tokens_seen": 1506279424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027456369107321967, + "loss": 1.259, + "theoretical_loss": 3.543239796904396, + "tokens_seen": 1506344960 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027455366098294885, + "loss": 1.4105, + "theoretical_loss": 3.543226325673161, + "tokens_seen": 1506410496 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027454363089267803, + "loss": 1.1963, + "theoretical_loss": 3.543212855192065, + "tokens_seen": 1506476032 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027453360080240726, + "loss": 1.4004, + "theoretical_loss": 3.543199385461034, + "tokens_seen": 1506541568 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002745235707121364, + "loss": 1.4258, + "theoretical_loss": 3.5431859164799935, + "tokens_seen": 1506607104 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002745135406218656, + "loss": 1.3915, + "theoretical_loss": 3.5431724482488685, + "tokens_seen": 1506672640 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027450351053159475, + "loss": 1.2991, + "theoretical_loss": 3.543158980767585, + "tokens_seen": 1506738176 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274493480441324, + "loss": 1.4651, + "theoretical_loss": 3.543145514036069, + "tokens_seen": 1506803712 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027448345035105317, + "loss": 1.335, + "theoretical_loss": 3.543132048054245, + "tokens_seen": 1506869248 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027447342026078235, + "loss": 1.4151, + "theoretical_loss": 3.54311858282204, + "tokens_seen": 1506934784 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027446339017051153, + "loss": 1.3556, + "theoretical_loss": 3.543105118339379, + "tokens_seen": 1507000320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027445336008024077, + "loss": 1.3404, + "theoretical_loss": 3.5430916546061875, + "tokens_seen": 1507065856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002744433299899699, + "loss": 1.3721, + "theoretical_loss": 3.543078191622392, + "tokens_seen": 1507131392 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027443329989969913, + "loss": 1.4112, + "theoretical_loss": 3.5430647293879174, + "tokens_seen": 1507196928 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027442326980942826, + "loss": 1.4021, + "theoretical_loss": 3.5430512679026895, + "tokens_seen": 1507262464 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 550589, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4698947668075562, + "objective/train/theoretical_loss": 3.543037807166635, + "objective/train/tokens_used": 1527788000, + "theoretical_loss": 3.543037807166635, + "tokens_seen": 1507328000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002744132397191575, + "loss": 1.3611, + "theoretical_loss": 3.543037807166635, + "tokens_seen": 1507328000 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027440320962888667, + "loss": 1.3485, + "theoretical_loss": 3.5430243471796783, + "tokens_seen": 1507393536 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027439317953861585, + "loss": 1.4494, + "theoretical_loss": 3.543010887941746, + "tokens_seen": 1507459072 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027438314944834503, + "loss": 1.4297, + "theoretical_loss": 3.5429974294527633, + "tokens_seen": 1507524608 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002743731193580742, + "loss": 1.504, + "theoretical_loss": 3.5429839717126566, + "tokens_seen": 1507590144 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002743630892678034, + "loss": 1.3892, + "theoretical_loss": 3.5429705147213513, + "tokens_seen": 1507655680 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027435305917753263, + "loss": 1.4092, + "theoretical_loss": 3.542957058478773, + "tokens_seen": 1507721216 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027434302908726176, + "loss": 1.3204, + "theoretical_loss": 3.5429436029848476, + "tokens_seen": 1507786752 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274332998996991, + "loss": 1.3235, + "theoretical_loss": 3.542930148239501, + "tokens_seen": 1507852288 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002743229689067202, + "loss": 1.3363, + "theoretical_loss": 3.5429166942426598, + "tokens_seen": 1507917824 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027431293881644936, + "loss": 1.3664, + "theoretical_loss": 3.5429032409942485, + "tokens_seen": 1507983360 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027430290872617854, + "loss": 1.3486, + "theoretical_loss": 3.5428897884941932, + "tokens_seen": 1508048896 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742928786359077, + "loss": 1.3942, + "theoretical_loss": 3.5428763367424203, + "tokens_seen": 1508114432 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742828485456369, + "loss": 1.4558, + "theoretical_loss": 3.542862885738856, + "tokens_seen": 1508179968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027427281845536614, + "loss": 1.3854, + "theoretical_loss": 3.542849435483425, + "tokens_seen": 1508245504 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027426278836509526, + "loss": 1.4334, + "theoretical_loss": 3.542835985976054, + "tokens_seen": 1508311040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742527582748245, + "loss": 1.3243, + "theoretical_loss": 3.5428225372166686, + "tokens_seen": 1508376576 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742427281845536, + "loss": 1.4072, + "theoretical_loss": 3.542809089205195, + "tokens_seen": 1508442112 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027423269809428286, + "loss": 1.3484, + "theoretical_loss": 3.5427956419415585, + "tokens_seen": 1508507648 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027422266800401204, + "loss": 1.3372, + "theoretical_loss": 3.542782195425686, + "tokens_seen": 1508573184 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742126379137412, + "loss": 1.362, + "theoretical_loss": 3.542768749657503, + "tokens_seen": 1508638720 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002742026078234704, + "loss": 1.3905, + "theoretical_loss": 3.5427553046369344, + "tokens_seen": 1508704256 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002741925777331996, + "loss": 1.4174, + "theoretical_loss": 3.5427418603639076, + "tokens_seen": 1508769792 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027418254764292877, + "loss": 1.3704, + "theoretical_loss": 3.542728416838348, + "tokens_seen": 1508835328 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274172517552658, + "loss": 1.3461, + "theoretical_loss": 3.542714974060182, + "tokens_seen": 1508900864 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027416248746238713, + "loss": 1.5785, + "theoretical_loss": 3.542701532029335, + "tokens_seen": 1508966400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027415245737211636, + "loss": 1.4279, + "theoretical_loss": 3.5426880907457328, + "tokens_seen": 1509031936 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027414242728184554, + "loss": 1.3518, + "theoretical_loss": 3.5426746502093023, + "tokens_seen": 1509097472 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002741323971915747, + "loss": 1.3108, + "theoretical_loss": 3.542661210419969, + "tokens_seen": 1509163008 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002741223671013039, + "loss": 1.3663, + "theoretical_loss": 3.542647771377659, + "tokens_seen": 1509228544 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002741123370110331, + "loss": 1.4186, + "theoretical_loss": 3.5426343330822982, + "tokens_seen": 1509294080 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027410230692076227, + "loss": 1.3724, + "theoretical_loss": 3.5426208955338128, + "tokens_seen": 1509359616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002740922768304915, + "loss": 1.3669, + "theoretical_loss": 3.542607458732129, + "tokens_seen": 1509425152 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027408224674022063, + "loss": 1.3912, + "theoretical_loss": 3.5425940226771724, + "tokens_seen": 1509490688 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027407221664994987, + "loss": 1.3436, + "theoretical_loss": 3.54258058736887, + "tokens_seen": 1509556224 + }, + { + "epoch": 0.46, + "learning_rate": 0.000274062186559679, + "loss": 1.3622, + "theoretical_loss": 3.5425671528071465, + "tokens_seen": 1509621760 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027405215646940823, + "loss": 1.3458, + "theoretical_loss": 3.542553718991929, + "tokens_seen": 1509687296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002740421263791374, + "loss": 1.4173, + "theoretical_loss": 3.5425402859231436, + "tokens_seen": 1509752832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002740320962888666, + "loss": 1.4302, + "theoretical_loss": 3.542526853600717, + "tokens_seen": 1509818368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002740220661985958, + "loss": 1.4156, + "theoretical_loss": 3.5425134220245735, + "tokens_seen": 1509883904 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027401203610832495, + "loss": 1.3529, + "theoretical_loss": 3.5424999911946404, + "tokens_seen": 1509949440 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002740020060180542, + "loss": 1.3264, + "theoretical_loss": 3.5424865611108443, + "tokens_seen": 1510014976 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027399197592778337, + "loss": 1.5059, + "theoretical_loss": 3.542473131773111, + "tokens_seen": 1510080512 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027398194583751255, + "loss": 1.3882, + "theoretical_loss": 3.542459703181366, + "tokens_seen": 1510146048 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027397191574724173, + "loss": 1.4334, + "theoretical_loss": 3.5424462753355366, + "tokens_seen": 1510211584 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027396188565697097, + "loss": 1.3773, + "theoretical_loss": 3.542432848235548, + "tokens_seen": 1510277120 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002739518555667001, + "loss": 1.4114, + "theoretical_loss": 3.5424194218813265, + "tokens_seen": 1510342656 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027394182547642933, + "loss": 1.3491, + "theoretical_loss": 3.542405996272799, + "tokens_seen": 1510408192 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027393179538615846, + "loss": 1.4344, + "theoretical_loss": 3.5423925714098914, + "tokens_seen": 1510473728 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002739217652958877, + "loss": 1.458, + "theoretical_loss": 3.5423791472925306, + "tokens_seen": 1510539264 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 551665, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3904643058776855, + "objective/train/theoretical_loss": 3.542365723920641, + "objective/train/tokens_used": 1531064800, + "theoretical_loss": 3.542365723920641, + "tokens_seen": 1510604800 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027391173520561687, + "loss": 1.3697, + "theoretical_loss": 3.542365723920641, + "tokens_seen": 1510604800 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027390170511534605, + "loss": 1.3636, + "theoretical_loss": 3.5423523012941507, + "tokens_seen": 1510670336 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027389167502507524, + "loss": 1.3574, + "theoretical_loss": 3.5423388794129855, + "tokens_seen": 1510735872 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002738816449348044, + "loss": 1.3646, + "theoretical_loss": 3.5423254582770713, + "tokens_seen": 1510801408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002738716148445336, + "loss": 1.2322, + "theoretical_loss": 3.542312037886335, + "tokens_seen": 1510866944 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027386158475426283, + "loss": 1.5065, + "theoretical_loss": 3.542298618240702, + "tokens_seen": 1510932480 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027385155466399196, + "loss": 1.448, + "theoretical_loss": 3.542285199340099, + "tokens_seen": 1510998016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002738415245737212, + "loss": 1.4323, + "theoretical_loss": 3.5422717811844526, + "tokens_seen": 1511063552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002738314944834504, + "loss": 1.3749, + "theoretical_loss": 3.5422583637736893, + "tokens_seen": 1511129088 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027382146439317956, + "loss": 1.4345, + "theoretical_loss": 3.5422449471077346, + "tokens_seen": 1511194624 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027381143430290874, + "loss": 1.3751, + "theoretical_loss": 3.542231531186516, + "tokens_seen": 1511260160 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002738014042126379, + "loss": 1.3716, + "theoretical_loss": 3.542218116009959, + "tokens_seen": 1511325696 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737913741223671, + "loss": 1.3701, + "theoretical_loss": 3.5422047015779903, + "tokens_seen": 1511391232 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027378134403209634, + "loss": 1.296, + "theoretical_loss": 3.542191287890536, + "tokens_seen": 1511456768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027377131394182546, + "loss": 1.3398, + "theoretical_loss": 3.542177874947523, + "tokens_seen": 1511522304 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737612838515547, + "loss": 1.3375, + "theoretical_loss": 3.542164462748877, + "tokens_seen": 1511587840 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737512537612838, + "loss": 1.4322, + "theoretical_loss": 3.5421510512945256, + "tokens_seen": 1511653376 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027374122367101306, + "loss": 1.3851, + "theoretical_loss": 3.542137640584394, + "tokens_seen": 1511718912 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027373119358074224, + "loss": 1.4602, + "theoretical_loss": 3.542124230618409, + "tokens_seen": 1511784448 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737211634904714, + "loss": 1.3582, + "theoretical_loss": 3.5421108213964976, + "tokens_seen": 1511849984 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737111334002006, + "loss": 1.4684, + "theoretical_loss": 3.5420974129185856, + "tokens_seen": 1511915520 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002737011033099298, + "loss": 1.3736, + "theoretical_loss": 3.5420840051845994, + "tokens_seen": 1511981056 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027369107321965897, + "loss": 1.4413, + "theoretical_loss": 3.5420705981944662, + "tokens_seen": 1512046592 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002736810431293882, + "loss": 1.2995, + "theoretical_loss": 3.5420571919481123, + "tokens_seen": 1512112128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027367101303911733, + "loss": 1.3698, + "theoretical_loss": 3.542043786445464, + "tokens_seen": 1512177664 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027366098294884656, + "loss": 1.409, + "theoretical_loss": 3.5420303816864473, + "tokens_seen": 1512243200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027365095285857574, + "loss": 1.3647, + "theoretical_loss": 3.5420169776709898, + "tokens_seen": 1512308736 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002736409227683049, + "loss": 1.3315, + "theoretical_loss": 3.542003574399017, + "tokens_seen": 1512374272 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002736308926780341, + "loss": 1.4105, + "theoretical_loss": 3.5419901718704563, + "tokens_seen": 1512439808 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002736208625877633, + "loss": 1.2748, + "theoretical_loss": 3.541976770085234, + "tokens_seen": 1512505344 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027361083249749247, + "loss": 1.3373, + "theoretical_loss": 3.541963369043276, + "tokens_seen": 1512570880 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002736008024072217, + "loss": 1.409, + "theoretical_loss": 3.5419499687445093, + "tokens_seen": 1512636416 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027359077231695083, + "loss": 1.3315, + "theoretical_loss": 3.541936569188861, + "tokens_seen": 1512701952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027358074222668007, + "loss": 1.4508, + "theoretical_loss": 3.5419231703762577, + "tokens_seen": 1512767488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002735707121364092, + "loss": 1.2926, + "theoretical_loss": 3.5419097723066253, + "tokens_seen": 1512833024 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027356068204613843, + "loss": 1.4091, + "theoretical_loss": 3.54189637497989, + "tokens_seen": 1512898560 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002735506519558676, + "loss": 1.4072, + "theoretical_loss": 3.5418829783959804, + "tokens_seen": 1512964096 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002735406218655968, + "loss": 1.3382, + "theoretical_loss": 3.5418695825548214, + "tokens_seen": 1513029632 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027353059177532597, + "loss": 1.3654, + "theoretical_loss": 3.54185618745634, + "tokens_seen": 1513095168 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027352056168505515, + "loss": 1.4565, + "theoretical_loss": 3.541842793100463, + "tokens_seen": 1513160704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027351053159478433, + "loss": 1.382, + "theoretical_loss": 3.541829399487117, + "tokens_seen": 1513226240 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027350050150451357, + "loss": 1.3696, + "theoretical_loss": 3.541816006616229, + "tokens_seen": 1513291776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734904714142427, + "loss": 1.3384, + "theoretical_loss": 3.5418026144877257, + "tokens_seen": 1513357312 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027348044132397193, + "loss": 1.3754, + "theoretical_loss": 3.5417892231015333, + "tokens_seen": 1513422848 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734704112337011, + "loss": 1.4346, + "theoretical_loss": 3.5417758324575788, + "tokens_seen": 1513488384 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734603811434303, + "loss": 1.3871, + "theoretical_loss": 3.5417624425557888, + "tokens_seen": 1513553920 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734503510531595, + "loss": 1.4254, + "theoretical_loss": 3.54174905339609, + "tokens_seen": 1513619456 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027344032096288866, + "loss": 1.4116, + "theoretical_loss": 3.5417356649784097, + "tokens_seen": 1513684992 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027343029087261784, + "loss": 1.4445, + "theoretical_loss": 3.541722277302674, + "tokens_seen": 1513750528 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734202607823471, + "loss": 1.3092, + "theoretical_loss": 3.5417088903688096, + "tokens_seen": 1513816064 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 552897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2360005378723145, + "objective/train/theoretical_loss": 3.541695504176744, + "objective/train/tokens_used": 1534341600, + "theoretical_loss": 3.541695504176744, + "tokens_seen": 1513881600 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002734102306920762, + "loss": 1.3325, + "theoretical_loss": 3.541695504176744, + "tokens_seen": 1513881600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027340020060180544, + "loss": 1.3033, + "theoretical_loss": 3.5416821187264036, + "tokens_seen": 1513947136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027339017051153456, + "loss": 1.3026, + "theoretical_loss": 3.541668734017715, + "tokens_seen": 1514012672 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002733801404212638, + "loss": 1.429, + "theoretical_loss": 3.541655350050605, + "tokens_seen": 1514078208 + }, + { + "epoch": 0.46, + "learning_rate": 0.000273370110330993, + "loss": 1.3443, + "theoretical_loss": 3.5416419668250008, + "tokens_seen": 1514143744 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027336008024072216, + "loss": 1.4605, + "theoretical_loss": 3.5416285843408293, + "tokens_seen": 1514209280 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027335005015045134, + "loss": 1.351, + "theoretical_loss": 3.5416152025980163, + "tokens_seen": 1514274816 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002733400200601806, + "loss": 1.316, + "theoretical_loss": 3.54160182159649, + "tokens_seen": 1514340352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002733299899699097, + "loss": 1.3554, + "theoretical_loss": 3.5415884413361765, + "tokens_seen": 1514405888 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027331995987963894, + "loss": 1.4067, + "theoretical_loss": 3.541575061817003, + "tokens_seen": 1514471424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027330992978936807, + "loss": 1.3508, + "theoretical_loss": 3.5415616830388954, + "tokens_seen": 1514536960 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002732998996990973, + "loss": 1.3703, + "theoretical_loss": 3.541548305001782, + "tokens_seen": 1514602496 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002732898696088265, + "loss": 1.4142, + "theoretical_loss": 3.541534927705589, + "tokens_seen": 1514668032 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027327983951855566, + "loss": 1.4286, + "theoretical_loss": 3.5415215511502436, + "tokens_seen": 1514733568 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002732698094282849, + "loss": 1.2686, + "theoretical_loss": 3.5415081753356725, + "tokens_seen": 1514799104 + }, + { + "epoch": 0.46, + "learning_rate": 0.000273259779338014, + "loss": 1.4029, + "theoretical_loss": 3.5414948002618023, + "tokens_seen": 1514864640 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027324974924774326, + "loss": 1.3953, + "theoretical_loss": 3.54148142592856, + "tokens_seen": 1514930176 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027323971915747244, + "loss": 1.3858, + "theoretical_loss": 3.5414680523358735, + "tokens_seen": 1514995712 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002732296890672016, + "loss": 1.4609, + "theoretical_loss": 3.541454679483669, + "tokens_seen": 1515061248 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002732196589769308, + "loss": 1.3615, + "theoretical_loss": 3.5414413073718736, + "tokens_seen": 1515126784 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027320962888666, + "loss": 1.3356, + "theoretical_loss": 3.5414279360004137, + "tokens_seen": 1515192320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027319959879638917, + "loss": 1.3391, + "theoretical_loss": 3.5414145653692177, + "tokens_seen": 1515257856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002731895687061184, + "loss": 1.4681, + "theoretical_loss": 3.541401195478211, + "tokens_seen": 1515323392 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027317953861584753, + "loss": 1.241, + "theoretical_loss": 3.5413878263273215, + "tokens_seen": 1515388928 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027316950852557676, + "loss": 1.4188, + "theoretical_loss": 3.5413744579164765, + "tokens_seen": 1515454464 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027315947843530594, + "loss": 1.3952, + "theoretical_loss": 3.5413610902456023, + "tokens_seen": 1515520000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002731494483450351, + "loss": 1.366, + "theoretical_loss": 3.5413477233146264, + "tokens_seen": 1515585536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002731394182547643, + "loss": 1.4265, + "theoretical_loss": 3.5413343571234757, + "tokens_seen": 1515651072 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002731293881644935, + "loss": 1.48, + "theoretical_loss": 3.5413209916720776, + "tokens_seen": 1515716608 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027311935807422267, + "loss": 1.3788, + "theoretical_loss": 3.5413076269603585, + "tokens_seen": 1515782144 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002731093279839519, + "loss": 1.2871, + "theoretical_loss": 3.541294262988246, + "tokens_seen": 1515847680 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027309929789368103, + "loss": 1.3093, + "theoretical_loss": 3.541280899755667, + "tokens_seen": 1515913216 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027308926780341027, + "loss": 1.3885, + "theoretical_loss": 3.5412675372625486, + "tokens_seen": 1515978752 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002730792377131394, + "loss": 1.4624, + "theoretical_loss": 3.5412541755088176, + "tokens_seen": 1516044288 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027306920762286863, + "loss": 1.3568, + "theoretical_loss": 3.541240814494402, + "tokens_seen": 1516109824 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002730591775325978, + "loss": 1.3047, + "theoretical_loss": 3.5412274542192286, + "tokens_seen": 1516175360 + }, + { + "epoch": 0.46, + "learning_rate": 0.000273049147442327, + "loss": 1.3402, + "theoretical_loss": 3.541214094683224, + "tokens_seen": 1516240896 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027303911735205617, + "loss": 1.3441, + "theoretical_loss": 3.541200735886316, + "tokens_seen": 1516306432 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027302908726178535, + "loss": 1.3931, + "theoretical_loss": 3.541187377828431, + "tokens_seen": 1516371968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027301905717151453, + "loss": 1.3994, + "theoretical_loss": 3.541174020509497, + "tokens_seen": 1516437504 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027300902708124377, + "loss": 1.3709, + "theoretical_loss": 3.541160663929441, + "tokens_seen": 1516503040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729989969909729, + "loss": 1.3069, + "theoretical_loss": 3.54114730808819, + "tokens_seen": 1516568576 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027298896690070213, + "loss": 1.4731, + "theoretical_loss": 3.5411339529856716, + "tokens_seen": 1516634112 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729789368104313, + "loss": 1.2039, + "theoretical_loss": 3.5411205986218124, + "tokens_seen": 1516699648 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729689067201605, + "loss": 1.3193, + "theoretical_loss": 3.5411072449965397, + "tokens_seen": 1516765184 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729588766298897, + "loss": 1.3185, + "theoretical_loss": 3.5410938921097816, + "tokens_seen": 1516830720 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027294884653961886, + "loss": 1.4339, + "theoretical_loss": 3.5410805399614644, + "tokens_seen": 1516896256 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027293881644934804, + "loss": 1.4284, + "theoretical_loss": 3.5410671885515157, + "tokens_seen": 1516961792 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729287863590773, + "loss": 1.2942, + "theoretical_loss": 3.541053837879863, + "tokens_seen": 1517027328 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002729187562688064, + "loss": 1.3473, + "theoretical_loss": 3.5410404879464332, + "tokens_seen": 1517092864 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 554075, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.420208215713501, + "objective/train/theoretical_loss": 3.5410271387511534, + "objective/train/tokens_used": 1537618400, + "theoretical_loss": 3.5410271387511534, + "tokens_seen": 1517158400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027290872617853564, + "loss": 1.4077, + "theoretical_loss": 3.5410271387511534, + "tokens_seen": 1517158400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027289869608826476, + "loss": 1.3956, + "theoretical_loss": 3.5410137902939516, + "tokens_seen": 1517223936 + }, + { + "epoch": 0.46, + "learning_rate": 0.000272888665997994, + "loss": 1.411, + "theoretical_loss": 3.541000442574755, + "tokens_seen": 1517289472 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002728786359077232, + "loss": 1.404, + "theoretical_loss": 3.54098709559349, + "tokens_seen": 1517355008 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027286860581745236, + "loss": 1.4087, + "theoretical_loss": 3.5409737493500852, + "tokens_seen": 1517420544 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027285857572718154, + "loss": 1.3617, + "theoretical_loss": 3.5409604038444673, + "tokens_seen": 1517486080 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002728485456369108, + "loss": 1.3953, + "theoretical_loss": 3.5409470590765633, + "tokens_seen": 1517551616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002728385155466399, + "loss": 1.4457, + "theoretical_loss": 3.540933715046301, + "tokens_seen": 1517617152 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027282848545636914, + "loss": 1.3716, + "theoretical_loss": 3.540920371753608, + "tokens_seen": 1517682688 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027281845536609827, + "loss": 1.3179, + "theoretical_loss": 3.540907029198412, + "tokens_seen": 1517748224 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002728084252758275, + "loss": 1.31, + "theoretical_loss": 3.540893687380639, + "tokens_seen": 1517813760 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002727983951855567, + "loss": 1.4569, + "theoretical_loss": 3.540880346300218, + "tokens_seen": 1517879296 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027278836509528586, + "loss": 1.3271, + "theoretical_loss": 3.5408670059570753, + "tokens_seen": 1517944832 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027277833500501504, + "loss": 1.3684, + "theoretical_loss": 3.540853666351138, + "tokens_seen": 1518010368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002727683049147442, + "loss": 1.3603, + "theoretical_loss": 3.5408403274823352, + "tokens_seen": 1518075904 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002727582748244734, + "loss": 1.4201, + "theoretical_loss": 3.540826989350593, + "tokens_seen": 1518141440 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027274824473420264, + "loss": 1.3116, + "theoretical_loss": 3.5408136519558395, + "tokens_seen": 1518206976 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027273821464393177, + "loss": 1.2977, + "theoretical_loss": 3.5408003152980014, + "tokens_seen": 1518272512 + }, + { + "epoch": 0.46, + "learning_rate": 0.000272728184553661, + "loss": 1.2816, + "theoretical_loss": 3.540786979377007, + "tokens_seen": 1518338048 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027271815446339013, + "loss": 1.4804, + "theoretical_loss": 3.5407736441927833, + "tokens_seen": 1518403584 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027270812437311937, + "loss": 1.3875, + "theoretical_loss": 3.540760309745258, + "tokens_seen": 1518469120 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027269809428284855, + "loss": 1.4031, + "theoretical_loss": 3.5407469760343586, + "tokens_seen": 1518534656 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027268806419257773, + "loss": 1.3708, + "theoretical_loss": 3.5407336430600127, + "tokens_seen": 1518600192 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002726780341023069, + "loss": 1.3085, + "theoretical_loss": 3.5407203108221474, + "tokens_seen": 1518665728 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027266800401203614, + "loss": 1.3943, + "theoretical_loss": 3.540706979320691, + "tokens_seen": 1518731264 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027265797392176527, + "loss": 1.4857, + "theoretical_loss": 3.5406936485555702, + "tokens_seen": 1518796800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002726479438314945, + "loss": 1.4069, + "theoretical_loss": 3.5406803185267135, + "tokens_seen": 1518862336 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027263791374122363, + "loss": 1.4107, + "theoretical_loss": 3.5406669892340474, + "tokens_seen": 1518927872 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027262788365095287, + "loss": 1.3566, + "theoretical_loss": 3.5406536606775, + "tokens_seen": 1518993408 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027261785356068205, + "loss": 1.4637, + "theoretical_loss": 3.540640332856999, + "tokens_seen": 1519058944 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027260782347041123, + "loss": 1.3556, + "theoretical_loss": 3.540627005772472, + "tokens_seen": 1519124480 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002725977933801404, + "loss": 1.332, + "theoretical_loss": 3.5406136794238465, + "tokens_seen": 1519190016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002725877632898696, + "loss": 1.4239, + "theoretical_loss": 3.5406003538110506, + "tokens_seen": 1519255552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002725777331995988, + "loss": 1.3979, + "theoretical_loss": 3.540587028934011, + "tokens_seen": 1519321088 + }, + { + "epoch": 0.46, + "learning_rate": 0.000272567703109328, + "loss": 1.3508, + "theoretical_loss": 3.5405737047926555, + "tokens_seen": 1519386624 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027255767301905714, + "loss": 1.3551, + "theoretical_loss": 3.540560381386913, + "tokens_seen": 1519452160 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027254764292878637, + "loss": 1.2451, + "theoretical_loss": 3.5405470587167094, + "tokens_seen": 1519517696 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002725376128385155, + "loss": 1.3289, + "theoretical_loss": 3.540533736781973, + "tokens_seen": 1519583232 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027252758274824473, + "loss": 1.4141, + "theoretical_loss": 3.5405204155826326, + "tokens_seen": 1519648768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027251755265797397, + "loss": 1.3835, + "theoretical_loss": 3.5405070951186146, + "tokens_seen": 1519714304 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002725075225677031, + "loss": 1.4335, + "theoretical_loss": 3.540493775389847, + "tokens_seen": 1519779840 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027249749247743233, + "loss": 1.2718, + "theoretical_loss": 3.5404804563962573, + "tokens_seen": 1519845376 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002724874623871615, + "loss": 1.3725, + "theoretical_loss": 3.5404671381377737, + "tokens_seen": 1519910912 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002724774322968907, + "loss": 1.468, + "theoretical_loss": 3.5404538206143243, + "tokens_seen": 1519976448 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002724674022066199, + "loss": 1.3587, + "theoretical_loss": 3.5404405038258355, + "tokens_seen": 1520041984 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027245737211634906, + "loss": 1.3735, + "theoretical_loss": 3.540427187772236, + "tokens_seen": 1520107520 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027244734202607824, + "loss": 1.3904, + "theoretical_loss": 3.540413872453454, + "tokens_seen": 1520173056 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002724373119358075, + "loss": 1.1642, + "theoretical_loss": 3.5404005578694164, + "tokens_seen": 1520238592 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002724272818455366, + "loss": 1.3962, + "theoretical_loss": 3.5403872440200512, + "tokens_seen": 1520304128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027241725175526584, + "loss": 1.3128, + "theoretical_loss": 3.540373930905286, + "tokens_seen": 1520369664 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 555119, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4170680046081543, + "objective/train/theoretical_loss": 3.5403606185250496, + "objective/train/tokens_used": 1540895200, + "theoretical_loss": 3.5403606185250496, + "tokens_seen": 1520435200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027240722166499496, + "loss": 1.455, + "theoretical_loss": 3.5403606185250496, + "tokens_seen": 1520435200 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002723971915747242, + "loss": 1.3579, + "theoretical_loss": 3.5403473068792684, + "tokens_seen": 1520500736 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002723871614844534, + "loss": 1.3617, + "theoretical_loss": 3.5403339959678712, + "tokens_seen": 1520566272 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027237713139418256, + "loss": 1.3153, + "theoretical_loss": 3.5403206857907854, + "tokens_seen": 1520631808 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027236710130391174, + "loss": 1.2979, + "theoretical_loss": 3.540307376347939, + "tokens_seen": 1520697344 + }, + { + "epoch": 0.46, + "learning_rate": 0.000272357071213641, + "loss": 1.443, + "theoretical_loss": 3.54029406763926, + "tokens_seen": 1520762880 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002723470411233701, + "loss": 1.496, + "theoretical_loss": 3.540280759664676, + "tokens_seen": 1520828416 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027233701103309934, + "loss": 1.4257, + "theoretical_loss": 3.540267452424115, + "tokens_seen": 1520893952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027232698094282847, + "loss": 1.3632, + "theoretical_loss": 3.540254145917505, + "tokens_seen": 1520959488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002723169508525577, + "loss": 1.3898, + "theoretical_loss": 3.540240840144774, + "tokens_seen": 1521025024 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002723069207622869, + "loss": 1.3309, + "theoretical_loss": 3.540227535105849, + "tokens_seen": 1521090560 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027229689067201606, + "loss": 1.2889, + "theoretical_loss": 3.5402142308006592, + "tokens_seen": 1521156096 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027228686058174524, + "loss": 1.3641, + "theoretical_loss": 3.540200927229132, + "tokens_seen": 1521221632 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002722768304914744, + "loss": 1.4197, + "theoretical_loss": 3.540187624391195, + "tokens_seen": 1521287168 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002722668004012036, + "loss": 1.3888, + "theoretical_loss": 3.5401743222867763, + "tokens_seen": 1521352704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027225677031093284, + "loss": 1.4595, + "theoretical_loss": 3.540161020915804, + "tokens_seen": 1521418240 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027224674022066197, + "loss": 1.3028, + "theoretical_loss": 3.540147720278206, + "tokens_seen": 1521483776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002722367101303912, + "loss": 1.3357, + "theoretical_loss": 3.540134420373911, + "tokens_seen": 1521549312 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027222668004012033, + "loss": 1.4419, + "theoretical_loss": 3.5401211212028456, + "tokens_seen": 1521614848 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027221664994984957, + "loss": 1.3025, + "theoretical_loss": 3.5401078227649387, + "tokens_seen": 1521680384 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027220661985957875, + "loss": 1.3479, + "theoretical_loss": 3.5400945250601183, + "tokens_seen": 1521745920 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027219658976930793, + "loss": 1.2729, + "theoretical_loss": 3.540081228088312, + "tokens_seen": 1521811456 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002721865596790371, + "loss": 1.3751, + "theoretical_loss": 3.540067931849448, + "tokens_seen": 1521876992 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027217652958876635, + "loss": 1.3737, + "theoretical_loss": 3.540054636343455, + "tokens_seen": 1521942528 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027216649949849547, + "loss": 1.5241, + "theoretical_loss": 3.5400413415702596, + "tokens_seen": 1522008064 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002721564694082247, + "loss": 1.4775, + "theoretical_loss": 3.5400280475297916, + "tokens_seen": 1522073600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027214643931795383, + "loss": 1.4144, + "theoretical_loss": 3.5400147542219775, + "tokens_seen": 1522139136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027213640922768307, + "loss": 1.3592, + "theoretical_loss": 3.5400014616467463, + "tokens_seen": 1522204672 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027212637913741225, + "loss": 1.4338, + "theoretical_loss": 3.539988169804026, + "tokens_seen": 1522270208 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027211634904714143, + "loss": 1.4762, + "theoretical_loss": 3.5399748786937444, + "tokens_seen": 1522335744 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002721063189568706, + "loss": 1.427, + "theoretical_loss": 3.5399615883158297, + "tokens_seen": 1522401280 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720962888665998, + "loss": 1.401, + "theoretical_loss": 3.53994829867021, + "tokens_seen": 1522466816 + }, + { + "epoch": 0.46, + "learning_rate": 0.000272086258776329, + "loss": 1.3608, + "theoretical_loss": 3.5399350097568143, + "tokens_seen": 1522532352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720762286860582, + "loss": 1.4803, + "theoretical_loss": 3.5399217215755696, + "tokens_seen": 1522597888 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027206619859578734, + "loss": 1.2536, + "theoretical_loss": 3.5399084341264038, + "tokens_seen": 1522663424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027205616850551657, + "loss": 1.1865, + "theoretical_loss": 3.539895147409246, + "tokens_seen": 1522728960 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720461384152457, + "loss": 1.4442, + "theoretical_loss": 3.539881861424025, + "tokens_seen": 1522794496 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027203610832497493, + "loss": 1.3645, + "theoretical_loss": 3.539868576170667, + "tokens_seen": 1522860032 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720260782347041, + "loss": 1.3314, + "theoretical_loss": 3.5398552916491015, + "tokens_seen": 1522925568 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720160481444333, + "loss": 1.3884, + "theoretical_loss": 3.539842007859256, + "tokens_seen": 1522991104 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002720060180541625, + "loss": 1.3615, + "theoretical_loss": 3.5398287248010596, + "tokens_seen": 1523056640 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002719959879638917, + "loss": 1.2394, + "theoretical_loss": 3.53981544247444, + "tokens_seen": 1523122176 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027198595787362084, + "loss": 1.3537, + "theoretical_loss": 3.5398021608793258, + "tokens_seen": 1523187712 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002719759277833501, + "loss": 1.3262, + "theoretical_loss": 3.539788880015644, + "tokens_seen": 1523253248 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002719658976930792, + "loss": 1.3677, + "theoretical_loss": 3.539775599883325, + "tokens_seen": 1523318784 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027195586760280844, + "loss": 1.3278, + "theoretical_loss": 3.5397623204822946, + "tokens_seen": 1523384320 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002719458375125376, + "loss": 1.2486, + "theoretical_loss": 3.5397490418124833, + "tokens_seen": 1523449856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002719358074222668, + "loss": 1.4879, + "theoretical_loss": 3.539735763873818, + "tokens_seen": 1523515392 + }, + { + "epoch": 0.46, + "learning_rate": 0.000271925777331996, + "loss": 1.4767, + "theoretical_loss": 3.5397224866662276, + "tokens_seen": 1523580928 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027191574724172516, + "loss": 1.4224, + "theoretical_loss": 3.53970921018964, + "tokens_seen": 1523646464 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 556367, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4135661125183105, + "objective/train/theoretical_loss": 3.5396959344439836, + "objective/train/tokens_used": 1544172000, + "theoretical_loss": 3.5396959344439836, + "tokens_seen": 1523712000 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027190571715145434, + "loss": 1.4855, + "theoretical_loss": 3.5396959344439836, + "tokens_seen": 1523712000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718956870611836, + "loss": 1.4936, + "theoretical_loss": 3.5396826594291873, + "tokens_seen": 1523777536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718856569709127, + "loss": 1.3576, + "theoretical_loss": 3.539669385145179, + "tokens_seen": 1523843072 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027187562688064194, + "loss": 1.3426, + "theoretical_loss": 3.5396561115918863, + "tokens_seen": 1523908608 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027186559679037107, + "loss": 1.4442, + "theoretical_loss": 3.5396428387692387, + "tokens_seen": 1523974144 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718555667001003, + "loss": 1.3692, + "theoretical_loss": 3.5396295666771644, + "tokens_seen": 1524039680 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718455366098295, + "loss": 1.3441, + "theoretical_loss": 3.539616295315591, + "tokens_seen": 1524105216 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027183550651955867, + "loss": 1.3867, + "theoretical_loss": 3.539603024684448, + "tokens_seen": 1524170752 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027182547642928785, + "loss": 1.441, + "theoretical_loss": 3.5395897547836626, + "tokens_seen": 1524236288 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718154463390171, + "loss": 1.2751, + "theoretical_loss": 3.539576485613164, + "tokens_seen": 1524301824 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002718054162487462, + "loss": 1.3464, + "theoretical_loss": 3.539563217172881, + "tokens_seen": 1524367360 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027179538615847544, + "loss": 1.2873, + "theoretical_loss": 3.5395499494627405, + "tokens_seen": 1524432896 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027178535606820457, + "loss": 1.4055, + "theoretical_loss": 3.539536682482672, + "tokens_seen": 1524498432 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002717753259779338, + "loss": 1.4032, + "theoretical_loss": 3.539523416232605, + "tokens_seen": 1524563968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027176529588766304, + "loss": 1.3639, + "theoretical_loss": 3.5395101507124656, + "tokens_seen": 1524629504 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027175526579739217, + "loss": 1.3693, + "theoretical_loss": 3.5394968859221834, + "tokens_seen": 1524695040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002717452357071214, + "loss": 1.5035, + "theoretical_loss": 3.539483621861687, + "tokens_seen": 1524760576 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027173520561685053, + "loss": 1.4031, + "theoretical_loss": 3.539470358530905, + "tokens_seen": 1524826112 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027172517552657977, + "loss": 1.503, + "theoretical_loss": 3.539457095929766, + "tokens_seen": 1524891648 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027171514543630895, + "loss": 1.3835, + "theoretical_loss": 3.539443834058198, + "tokens_seen": 1524957184 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027170511534603813, + "loss": 1.3481, + "theoretical_loss": 3.539430572916129, + "tokens_seen": 1525022720 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002716950852557673, + "loss": 1.3287, + "theoretical_loss": 3.539417312503489, + "tokens_seen": 1525088256 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027168505516549655, + "loss": 1.2626, + "theoretical_loss": 3.5394040528202058, + "tokens_seen": 1525153792 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027167502507522567, + "loss": 1.264, + "theoretical_loss": 3.5393907938662075, + "tokens_seen": 1525219328 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002716649949849549, + "loss": 1.3724, + "theoretical_loss": 3.5393775356414237, + "tokens_seen": 1525284864 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027165496489468403, + "loss": 1.4875, + "theoretical_loss": 3.5393642781457815, + "tokens_seen": 1525350400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027164493480441327, + "loss": 1.2826, + "theoretical_loss": 3.5393510213792108, + "tokens_seen": 1525415936 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027163490471414245, + "loss": 1.4136, + "theoretical_loss": 3.5393377653416396, + "tokens_seen": 1525481472 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027162487462387163, + "loss": 1.4218, + "theoretical_loss": 3.539324510032997, + "tokens_seen": 1525547008 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002716148445336008, + "loss": 1.3304, + "theoretical_loss": 3.5393112554532102, + "tokens_seen": 1525612544 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027160481444333, + "loss": 1.3394, + "theoretical_loss": 3.5392980016022095, + "tokens_seen": 1525678080 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715947843530592, + "loss": 1.2901, + "theoretical_loss": 3.539284748479923, + "tokens_seen": 1525743616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715847542627884, + "loss": 1.4374, + "theoretical_loss": 3.539271496086279, + "tokens_seen": 1525809152 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027157472417251754, + "loss": 1.3447, + "theoretical_loss": 3.539258244421206, + "tokens_seen": 1525874688 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715646940822468, + "loss": 1.2625, + "theoretical_loss": 3.5392449934846333, + "tokens_seen": 1525940224 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715546639919759, + "loss": 1.4136, + "theoretical_loss": 3.5392317432764893, + "tokens_seen": 1526005760 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027154463390170514, + "loss": 1.406, + "theoretical_loss": 3.539218493796702, + "tokens_seen": 1526071296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715346038114343, + "loss": 1.4446, + "theoretical_loss": 3.539205245045202, + "tokens_seen": 1526136832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715245737211635, + "loss": 1.3782, + "theoretical_loss": 3.5391919970219154, + "tokens_seen": 1526202368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715145436308927, + "loss": 1.4171, + "theoretical_loss": 3.539178749726773, + "tokens_seen": 1526267904 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002715045135406219, + "loss": 1.3746, + "theoretical_loss": 3.5391655031597025, + "tokens_seen": 1526333440 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027149448345035104, + "loss": 1.427, + "theoretical_loss": 3.539152257320633, + "tokens_seen": 1526398976 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002714844533600803, + "loss": 1.2298, + "theoretical_loss": 3.5391390122094926, + "tokens_seen": 1526464512 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002714744232698094, + "loss": 1.3462, + "theoretical_loss": 3.539125767826211, + "tokens_seen": 1526530048 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027146439317953864, + "loss": 1.4627, + "theoretical_loss": 3.5391125241707164, + "tokens_seen": 1526595584 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002714543630892678, + "loss": 1.3288, + "theoretical_loss": 3.5390992812429376, + "tokens_seen": 1526661120 + }, + { + "epoch": 0.46, + "learning_rate": 0.000271444332998997, + "loss": 1.2909, + "theoretical_loss": 3.539086039042804, + "tokens_seen": 1526726656 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002714343029087262, + "loss": 1.4193, + "theoretical_loss": 3.5390727975702427, + "tokens_seen": 1526792192 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027142427281845536, + "loss": 1.4278, + "theoretical_loss": 3.539059556825184, + "tokens_seen": 1526857728 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027141424272818454, + "loss": 1.3757, + "theoretical_loss": 3.539046316807557, + "tokens_seen": 1526923264 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 557485, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.30222749710083, + "objective/train/theoretical_loss": 3.5390330775172893, + "objective/train/tokens_used": 1547448800, + "theoretical_loss": 3.5390330775172893, + "tokens_seen": 1526988800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002714042126379138, + "loss": 1.3614, + "theoretical_loss": 3.5390330775172893, + "tokens_seen": 1526988800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002713941825476429, + "loss": 1.3735, + "theoretical_loss": 3.53901983895431, + "tokens_seen": 1527054336 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027138415245737214, + "loss": 1.3244, + "theoretical_loss": 3.5390066011185493, + "tokens_seen": 1527119872 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027137412236710127, + "loss": 1.2474, + "theoretical_loss": 3.538993364009934, + "tokens_seen": 1527185408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002713640922768305, + "loss": 1.3113, + "theoretical_loss": 3.538980127628394, + "tokens_seen": 1527250944 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002713540621865597, + "loss": 1.3395, + "theoretical_loss": 3.5389668919738586, + "tokens_seen": 1527316480 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027134403209628887, + "loss": 1.3879, + "theoretical_loss": 3.5389536570462554, + "tokens_seen": 1527382016 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027133400200601805, + "loss": 1.3557, + "theoretical_loss": 3.5389404228455144, + "tokens_seen": 1527447552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002713239719157473, + "loss": 1.3863, + "theoretical_loss": 3.538927189371564, + "tokens_seen": 1527513088 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002713139418254764, + "loss": 1.3447, + "theoretical_loss": 3.5389139566243335, + "tokens_seen": 1527578624 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027130391173520564, + "loss": 1.4119, + "theoretical_loss": 3.538900724603752, + "tokens_seen": 1527644160 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027129388164493477, + "loss": 1.423, + "theoretical_loss": 3.538887493309747, + "tokens_seen": 1527709696 + }, + { + "epoch": 0.46, + "learning_rate": 0.000271283851554664, + "loss": 1.3045, + "theoretical_loss": 3.538874262742249, + "tokens_seen": 1527775232 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002712738214643932, + "loss": 1.49, + "theoretical_loss": 3.5388610329011865, + "tokens_seen": 1527840768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027126379137412237, + "loss": 1.2947, + "theoretical_loss": 3.5388478037864877, + "tokens_seen": 1527906304 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027125376128385155, + "loss": 1.3304, + "theoretical_loss": 3.5388345753980825, + "tokens_seen": 1527971840 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027124373119358073, + "loss": 1.3542, + "theoretical_loss": 3.5388213477359, + "tokens_seen": 1528037376 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002712337011033099, + "loss": 1.3541, + "theoretical_loss": 3.5388081207998683, + "tokens_seen": 1528102912 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027122367101303915, + "loss": 1.2906, + "theoretical_loss": 3.538794894589917, + "tokens_seen": 1528168448 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002712136409227683, + "loss": 1.4798, + "theoretical_loss": 3.5387816691059752, + "tokens_seen": 1528233984 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002712036108324975, + "loss": 1.267, + "theoretical_loss": 3.5387684443479714, + "tokens_seen": 1528299520 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002711935807422267, + "loss": 1.422, + "theoretical_loss": 3.538755220315835, + "tokens_seen": 1528365056 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027118355065195587, + "loss": 1.4585, + "theoretical_loss": 3.538741997009495, + "tokens_seen": 1528430592 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027117352056168505, + "loss": 1.3864, + "theoretical_loss": 3.53872877442888, + "tokens_seen": 1528496128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027116349047141423, + "loss": 1.3667, + "theoretical_loss": 3.53871555257392, + "tokens_seen": 1528561664 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002711534603811434, + "loss": 1.3795, + "theoretical_loss": 3.5387023314445436, + "tokens_seen": 1528627200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027114343029087265, + "loss": 1.426, + "theoretical_loss": 3.538689111040679, + "tokens_seen": 1528692736 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002711334002006018, + "loss": 1.3814, + "theoretical_loss": 3.5386758913622565, + "tokens_seen": 1528758272 + }, + { + "epoch": 0.46, + "learning_rate": 0.000271123370110331, + "loss": 1.3154, + "theoretical_loss": 3.538662672409205, + "tokens_seen": 1528823808 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027111334002006014, + "loss": 1.2462, + "theoretical_loss": 3.5386494541814533, + "tokens_seen": 1528889344 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002711033099297894, + "loss": 1.4015, + "theoretical_loss": 3.5386362366789306, + "tokens_seen": 1528954880 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027109327983951856, + "loss": 1.5651, + "theoretical_loss": 3.538623019901566, + "tokens_seen": 1529020416 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027108324974924774, + "loss": 1.4164, + "theoretical_loss": 3.5386098038492886, + "tokens_seen": 1529085952 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710732196589769, + "loss": 1.3698, + "theoretical_loss": 3.5385965885220276, + "tokens_seen": 1529151488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710631895687061, + "loss": 1.3157, + "theoretical_loss": 3.538583373919712, + "tokens_seen": 1529217024 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710531594784353, + "loss": 1.3865, + "theoretical_loss": 3.5385701600422714, + "tokens_seen": 1529282560 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710431293881645, + "loss": 1.4842, + "theoretical_loss": 3.538556946889635, + "tokens_seen": 1529348096 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027103309929789364, + "loss": 1.408, + "theoretical_loss": 3.538543734461731, + "tokens_seen": 1529413632 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710230692076229, + "loss": 1.4361, + "theoretical_loss": 3.5385305227584896, + "tokens_seen": 1529479168 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002710130391173521, + "loss": 1.3168, + "theoretical_loss": 3.53851731177984, + "tokens_seen": 1529544704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027100300902708124, + "loss": 1.3287, + "theoretical_loss": 3.538504101525711, + "tokens_seen": 1529610240 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002709929789368105, + "loss": 1.3301, + "theoretical_loss": 3.5384908919960316, + "tokens_seen": 1529675776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002709829488465396, + "loss": 1.358, + "theoretical_loss": 3.538477683190732, + "tokens_seen": 1529741312 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027097291875626884, + "loss": 1.4695, + "theoretical_loss": 3.5384644751097403, + "tokens_seen": 1529806848 + }, + { + "epoch": 0.46, + "learning_rate": 0.000270962888665998, + "loss": 1.2418, + "theoretical_loss": 3.5384512677529862, + "tokens_seen": 1529872384 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002709528585757272, + "loss": 1.4568, + "theoretical_loss": 3.5384380611203987, + "tokens_seen": 1529937920 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002709428284854564, + "loss": 1.4371, + "theoretical_loss": 3.5384248552119084, + "tokens_seen": 1530003456 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027093279839518556, + "loss": 1.4223, + "theoretical_loss": 3.538411650027443, + "tokens_seen": 1530068992 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027092276830491474, + "loss": 1.3573, + "theoretical_loss": 3.5383984455669326, + "tokens_seen": 1530134528 + }, + { + "epoch": 0.46, + "learning_rate": 0.000270912738214644, + "loss": 1.3621, + "theoretical_loss": 3.538385241830306, + "tokens_seen": 1530200064 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 558616, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2839939594268799, + "objective/train/theoretical_loss": 3.5383720388174935, + "objective/train/tokens_used": 1550725600, + "theoretical_loss": 3.5383720388174935, + "tokens_seen": 1530265600 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002709027081243731, + "loss": 1.4013, + "theoretical_loss": 3.5383720388174935, + "tokens_seen": 1530265600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027089267803410234, + "loss": 1.42, + "theoretical_loss": 3.538358836528423, + "tokens_seen": 1530331136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027088264794383147, + "loss": 1.4443, + "theoretical_loss": 3.538345634963025, + "tokens_seen": 1530396672 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002708726178535607, + "loss": 1.2994, + "theoretical_loss": 3.5383324341212283, + "tokens_seen": 1530462208 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002708625877632899, + "loss": 1.4253, + "theoretical_loss": 3.5383192340029623, + "tokens_seen": 1530527744 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027085255767301907, + "loss": 1.4517, + "theoretical_loss": 3.5383060346081567, + "tokens_seen": 1530593280 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027084252758274825, + "loss": 1.4215, + "theoretical_loss": 3.5382928359367405, + "tokens_seen": 1530658816 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002708324974924775, + "loss": 1.4132, + "theoretical_loss": 3.5382796379886434, + "tokens_seen": 1530724352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002708224674022066, + "loss": 1.2537, + "theoretical_loss": 3.5382664407637945, + "tokens_seen": 1530789888 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027081243731193584, + "loss": 1.4155, + "theoretical_loss": 3.538253244262123, + "tokens_seen": 1530855424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027080240722166497, + "loss": 1.4008, + "theoretical_loss": 3.538240048483559, + "tokens_seen": 1530920960 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707923771313942, + "loss": 1.3769, + "theoretical_loss": 3.5382268534280317, + "tokens_seen": 1530986496 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707823470411234, + "loss": 1.3004, + "theoretical_loss": 3.53821365909547, + "tokens_seen": 1531052032 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027077231695085257, + "loss": 1.3336, + "theoretical_loss": 3.5382004654858035, + "tokens_seen": 1531117568 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027076228686058175, + "loss": 1.378, + "theoretical_loss": 3.5381872725989627, + "tokens_seen": 1531183104 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027075225677031093, + "loss": 1.3034, + "theoretical_loss": 3.5381740804348754, + "tokens_seen": 1531248640 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707422266800401, + "loss": 1.4509, + "theoretical_loss": 3.538160888993472, + "tokens_seen": 1531314176 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027073219658976935, + "loss": 1.4563, + "theoretical_loss": 3.538147698274683, + "tokens_seen": 1531379712 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707221664994985, + "loss": 1.4102, + "theoretical_loss": 3.5381345082784357, + "tokens_seen": 1531445248 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707121364092277, + "loss": 1.2847, + "theoretical_loss": 3.538121319004661, + "tokens_seen": 1531510784 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002707021063189569, + "loss": 1.4119, + "theoretical_loss": 3.538108130453288, + "tokens_seen": 1531576320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027069207622868607, + "loss": 1.5373, + "theoretical_loss": 3.5380949426242463, + "tokens_seen": 1531641856 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027068204613841525, + "loss": 1.3287, + "theoretical_loss": 3.538081755517466, + "tokens_seen": 1531707392 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027067201604814443, + "loss": 1.4078, + "theoretical_loss": 3.5380685691328755, + "tokens_seen": 1531772928 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002706619859578736, + "loss": 1.3966, + "theoretical_loss": 3.5380553834704056, + "tokens_seen": 1531838464 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027065195586760285, + "loss": 1.4213, + "theoretical_loss": 3.5380421985299844, + "tokens_seen": 1531904000 + }, + { + "epoch": 0.46, + "learning_rate": 0.000270641925777332, + "loss": 1.4211, + "theoretical_loss": 3.538029014311543, + "tokens_seen": 1531969536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002706318956870612, + "loss": 1.3311, + "theoretical_loss": 3.5380158308150103, + "tokens_seen": 1532035072 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027062186559679034, + "loss": 1.4427, + "theoretical_loss": 3.5380026480403153, + "tokens_seen": 1532100608 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002706118355065196, + "loss": 1.3897, + "theoretical_loss": 3.5379894659873887, + "tokens_seen": 1532166144 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027060180541624876, + "loss": 1.2887, + "theoretical_loss": 3.5379762846561595, + "tokens_seen": 1532231680 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027059177532597794, + "loss": 1.2862, + "theoretical_loss": 3.5379631040465576, + "tokens_seen": 1532297216 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705817452357071, + "loss": 1.3121, + "theoretical_loss": 3.5379499241585117, + "tokens_seen": 1532362752 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705717151454363, + "loss": 1.3865, + "theoretical_loss": 3.5379367449919528, + "tokens_seen": 1532428288 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705616850551655, + "loss": 1.3057, + "theoretical_loss": 3.53792356654681, + "tokens_seen": 1532493824 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705516549648947, + "loss": 1.4203, + "theoretical_loss": 3.537910388823013, + "tokens_seen": 1532559360 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027054162487462384, + "loss": 1.3448, + "theoretical_loss": 3.537897211820491, + "tokens_seen": 1532624896 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705315947843531, + "loss": 1.3823, + "theoretical_loss": 3.5378840355391743, + "tokens_seen": 1532690432 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027052156469408226, + "loss": 1.3333, + "theoretical_loss": 3.537870859978992, + "tokens_seen": 1532755968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027051153460381144, + "loss": 1.4025, + "theoretical_loss": 3.5378576851398744, + "tokens_seen": 1532821504 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002705015045135406, + "loss": 1.6197, + "theoretical_loss": 3.537844511021751, + "tokens_seen": 1532887040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704914744232698, + "loss": 1.3875, + "theoretical_loss": 3.5378313376245516, + "tokens_seen": 1532952576 + }, + { + "epoch": 0.46, + "learning_rate": 0.000270481444332999, + "loss": 1.3164, + "theoretical_loss": 3.537818164948206, + "tokens_seen": 1533018112 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704714142427282, + "loss": 1.3557, + "theoretical_loss": 3.537804992992643, + "tokens_seen": 1533083648 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027046138415245735, + "loss": 1.3599, + "theoretical_loss": 3.5377918217577937, + "tokens_seen": 1533149184 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704513540621866, + "loss": 1.425, + "theoretical_loss": 3.5377786512435874, + "tokens_seen": 1533214720 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704413239719157, + "loss": 1.4605, + "theoretical_loss": 3.537765481449954, + "tokens_seen": 1533280256 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027043129388164494, + "loss": 1.365, + "theoretical_loss": 3.5377523123768224, + "tokens_seen": 1533345792 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704212637913741, + "loss": 1.3696, + "theoretical_loss": 3.5377391440241226, + "tokens_seen": 1533411328 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704112337011033, + "loss": 1.4984, + "theoretical_loss": 3.5377259763917857, + "tokens_seen": 1533476864 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 559697, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4352564811706543, + "objective/train/theoretical_loss": 3.53771280947974, + "objective/train/tokens_used": 1554002400, + "theoretical_loss": 3.53771280947974, + "tokens_seen": 1533542400 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002704012036108325, + "loss": 1.54, + "theoretical_loss": 3.53771280947974, + "tokens_seen": 1533542400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027039117352056167, + "loss": 1.4577, + "theoretical_loss": 3.5376996432879166, + "tokens_seen": 1533607936 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027038114343029085, + "loss": 1.3408, + "theoretical_loss": 3.5376864778162442, + "tokens_seen": 1533673472 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002703711133400201, + "loss": 1.3627, + "theoretical_loss": 3.537673313064653, + "tokens_seen": 1533739008 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002703610832497492, + "loss": 1.3837, + "theoretical_loss": 3.5376601490330737, + "tokens_seen": 1533804544 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027035105315947845, + "loss": 1.2871, + "theoretical_loss": 3.5376469857214348, + "tokens_seen": 1533870080 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027034102306920763, + "loss": 1.3909, + "theoretical_loss": 3.537633823129667, + "tokens_seen": 1533935616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002703309929789368, + "loss": 1.4487, + "theoretical_loss": 3.5376206612576997, + "tokens_seen": 1534001152 + }, + { + "epoch": 0.46, + "learning_rate": 0.000270320962888666, + "loss": 1.455, + "theoretical_loss": 3.537607500105463, + "tokens_seen": 1534066688 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027031093279839517, + "loss": 1.4063, + "theoretical_loss": 3.5375943396728875, + "tokens_seen": 1534132224 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027030090270812435, + "loss": 1.4234, + "theoretical_loss": 3.5375811799599015, + "tokens_seen": 1534197760 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002702908726178536, + "loss": 1.4466, + "theoretical_loss": 3.537568020966437, + "tokens_seen": 1534263296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0002702808425275827, + "loss": 1.3706, + "theoretical_loss": 3.537554862692422, + "tokens_seen": 1534328832 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027027081243731195, + "loss": 1.5796, + "theoretical_loss": 3.5375417051377873, + "tokens_seen": 1534394368 + }, + { + "epoch": 0.46, + "learning_rate": 0.00027026078234704113, + "loss": 1.4282, + "theoretical_loss": 3.537528548302463, + "tokens_seen": 1534459904 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002702507522567703, + "loss": 1.4389, + "theoretical_loss": 3.5375153921863793, + "tokens_seen": 1534525440 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027024072216649955, + "loss": 1.4119, + "theoretical_loss": 3.5375022367894653, + "tokens_seen": 1534590976 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002702306920762287, + "loss": 1.4107, + "theoretical_loss": 3.537489082111651, + "tokens_seen": 1534656512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002702206619859579, + "loss": 1.4202, + "theoretical_loss": 3.537475928152867, + "tokens_seen": 1534722048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002702106318956871, + "loss": 1.4675, + "theoretical_loss": 3.5374627749130436, + "tokens_seen": 1534787584 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027020060180541627, + "loss": 1.3895, + "theoretical_loss": 3.53744962239211, + "tokens_seen": 1534853120 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027019057171514545, + "loss": 1.4608, + "theoretical_loss": 3.537436470589996, + "tokens_seen": 1534918656 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027018054162487463, + "loss": 1.3421, + "theoretical_loss": 3.5374233195066327, + "tokens_seen": 1534984192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002701705115346038, + "loss": 1.4196, + "theoretical_loss": 3.5374101691419497, + "tokens_seen": 1535049728 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027016048144433305, + "loss": 1.379, + "theoretical_loss": 3.5373970194958764, + "tokens_seen": 1535115264 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002701504513540622, + "loss": 1.3873, + "theoretical_loss": 3.5373838705683434, + "tokens_seen": 1535180800 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002701404212637914, + "loss": 1.3381, + "theoretical_loss": 3.537370722359281, + "tokens_seen": 1535246336 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027013039117352054, + "loss": 1.3359, + "theoretical_loss": 3.5373575748686186, + "tokens_seen": 1535311872 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002701203610832498, + "loss": 1.3946, + "theoretical_loss": 3.5373444280962874, + "tokens_seen": 1535377408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027011033099297896, + "loss": 1.5047, + "theoretical_loss": 3.5373312820422163, + "tokens_seen": 1535442944 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027010030090270814, + "loss": 1.3292, + "theoretical_loss": 3.537318136706336, + "tokens_seen": 1535508480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700902708124373, + "loss": 1.4115, + "theoretical_loss": 3.5373049920885764, + "tokens_seen": 1535574016 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700802407221665, + "loss": 1.4544, + "theoretical_loss": 3.5372918481888673, + "tokens_seen": 1535639552 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700702106318957, + "loss": 1.5071, + "theoretical_loss": 3.53727870500714, + "tokens_seen": 1535705088 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700601805416249, + "loss": 1.4065, + "theoretical_loss": 3.5372655625433236, + "tokens_seen": 1535770624 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027005015045135404, + "loss": 1.367, + "theoretical_loss": 3.5372524207973486, + "tokens_seen": 1535836160 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700401203610833, + "loss": 1.4617, + "theoretical_loss": 3.537239279769145, + "tokens_seen": 1535901696 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027003009027081246, + "loss": 1.4168, + "theoretical_loss": 3.5372261394586433, + "tokens_seen": 1535967232 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027002006018054164, + "loss": 1.3385, + "theoretical_loss": 3.5372129998657735, + "tokens_seen": 1536032768 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002700100300902708, + "loss": 1.4404, + "theoretical_loss": 3.537199860990466, + "tokens_seen": 1536098304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00027, + "loss": 1.4351, + "theoretical_loss": 3.53718672283265, + "tokens_seen": 1536163840 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699899699097292, + "loss": 1.546, + "theoretical_loss": 3.5371735853922566, + "tokens_seen": 1536229376 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699799398194584, + "loss": 1.3918, + "theoretical_loss": 3.5371604486692165, + "tokens_seen": 1536294912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026996990972918755, + "loss": 1.5378, + "theoretical_loss": 3.5371473126634587, + "tokens_seen": 1536360448 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699598796389168, + "loss": 1.4495, + "theoretical_loss": 3.5371341773749143, + "tokens_seen": 1536425984 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699498495486459, + "loss": 1.3997, + "theoretical_loss": 3.5371210428035136, + "tokens_seen": 1536491520 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026993981945837514, + "loss": 1.3426, + "theoretical_loss": 3.537107908949186, + "tokens_seen": 1536557056 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699297893681043, + "loss": 1.4542, + "theoretical_loss": 3.537094775811863, + "tokens_seen": 1536622592 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699197592778335, + "loss": 1.3011, + "theoretical_loss": 3.537081643391474, + "tokens_seen": 1536688128 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002699097291875627, + "loss": 1.5247, + "theoretical_loss": 3.5370685116879494, + "tokens_seen": 1536753664 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 561026, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.264318823814392, + "objective/train/theoretical_loss": 3.5370553807012195, + "objective/train/tokens_used": 1557279200, + "theoretical_loss": 3.5370553807012195, + "tokens_seen": 1536819200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026989969909729187, + "loss": 1.4142, + "theoretical_loss": 3.5370553807012195, + "tokens_seen": 1536819200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026988966900702105, + "loss": 1.3441, + "theoretical_loss": 3.537042250431215, + "tokens_seen": 1536884736 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002698796389167503, + "loss": 1.436, + "theoretical_loss": 3.5370291208778655, + "tokens_seen": 1536950272 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002698696088264794, + "loss": 1.3389, + "theoretical_loss": 3.537015992041103, + "tokens_seen": 1537015808 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026985957873620865, + "loss": 1.3629, + "theoretical_loss": 3.537002863920855, + "tokens_seen": 1537081344 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026984954864593783, + "loss": 1.4364, + "theoretical_loss": 3.5369897365170546, + "tokens_seen": 1537146880 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269839518555667, + "loss": 1.4086, + "theoretical_loss": 3.5369766098296305, + "tokens_seen": 1537212416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002698294884653962, + "loss": 1.4612, + "theoretical_loss": 3.536963483858514, + "tokens_seen": 1537277952 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026981945837512537, + "loss": 1.3856, + "theoretical_loss": 3.536950358603635, + "tokens_seen": 1537343488 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026980942828485455, + "loss": 1.5501, + "theoretical_loss": 3.5369372340649234, + "tokens_seen": 1537409024 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697993981945838, + "loss": 1.375, + "theoretical_loss": 3.5369241102423103, + "tokens_seen": 1537474560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697893681043129, + "loss": 1.5029, + "theoretical_loss": 3.5369109871357267, + "tokens_seen": 1537540096 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026977933801404215, + "loss": 1.3498, + "theoretical_loss": 3.536897864745101, + "tokens_seen": 1537605632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697693079237713, + "loss": 1.4259, + "theoretical_loss": 3.536884743070366, + "tokens_seen": 1537671168 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697592778335005, + "loss": 1.3185, + "theoretical_loss": 3.536871622111451, + "tokens_seen": 1537736704 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697492477432297, + "loss": 1.3189, + "theoretical_loss": 3.536858501868286, + "tokens_seen": 1537802240 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697392176529589, + "loss": 1.361, + "theoretical_loss": 3.536845382340802, + "tokens_seen": 1537867776 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026972918756268806, + "loss": 1.471, + "theoretical_loss": 3.5368322635289298, + "tokens_seen": 1537933312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697191574724173, + "loss": 1.4419, + "theoretical_loss": 3.536819145432599, + "tokens_seen": 1537998848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002697091273821464, + "loss": 1.3467, + "theoretical_loss": 3.536806028051741, + "tokens_seen": 1538064384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026969909729187565, + "loss": 1.448, + "theoretical_loss": 3.5367929113862853, + "tokens_seen": 1538129920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002696890672016048, + "loss": 1.3414, + "theoretical_loss": 3.5367797954361633, + "tokens_seen": 1538195456 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269679037111334, + "loss": 1.4771, + "theoretical_loss": 3.536766680201305, + "tokens_seen": 1538260992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002696690070210632, + "loss": 1.4298, + "theoretical_loss": 3.536753565681641, + "tokens_seen": 1538326528 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002696589769307924, + "loss": 1.2534, + "theoretical_loss": 3.5367404518771024, + "tokens_seen": 1538392064 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026964894684052156, + "loss": 1.4602, + "theoretical_loss": 3.5367273387876192, + "tokens_seen": 1538457600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026963891675025074, + "loss": 1.4323, + "theoretical_loss": 3.536714226413122, + "tokens_seen": 1538523136 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002696288866599799, + "loss": 1.416, + "theoretical_loss": 3.5367011147535408, + "tokens_seen": 1538588672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026961885656970916, + "loss": 1.3103, + "theoretical_loss": 3.5366880038088073, + "tokens_seen": 1538654208 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002696088264794383, + "loss": 1.399, + "theoretical_loss": 3.5366748935788515, + "tokens_seen": 1538719744 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002695987963891675, + "loss": 1.4124, + "theoretical_loss": 3.5366617840636043, + "tokens_seen": 1538785280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026958876629889665, + "loss": 1.3624, + "theoretical_loss": 3.5366486752629953, + "tokens_seen": 1538850816 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002695787362086259, + "loss": 1.262, + "theoretical_loss": 3.5366355671769565, + "tokens_seen": 1538916352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026956870611835506, + "loss": 1.3743, + "theoretical_loss": 3.5366224598054177, + "tokens_seen": 1538981888 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026955867602808424, + "loss": 1.4147, + "theoretical_loss": 3.5366093531483096, + "tokens_seen": 1539047424 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002695486459378134, + "loss": 1.4124, + "theoretical_loss": 3.536596247205563, + "tokens_seen": 1539112960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026953861584754266, + "loss": 1.4701, + "theoretical_loss": 3.5365831419771085, + "tokens_seen": 1539178496 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002695285857572718, + "loss": 1.4202, + "theoretical_loss": 3.5365700374628766, + "tokens_seen": 1539244032 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269518555667001, + "loss": 1.2647, + "theoretical_loss": 3.5365569336627987, + "tokens_seen": 1539309568 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002695085255767302, + "loss": 1.5176, + "theoretical_loss": 3.5365438305768038, + "tokens_seen": 1539375104 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694984954864594, + "loss": 1.2631, + "theoretical_loss": 3.5365307282048244, + "tokens_seen": 1539440640 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694884653961886, + "loss": 1.3801, + "theoretical_loss": 3.5365176265467904, + "tokens_seen": 1539506176 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026947843530591775, + "loss": 1.4132, + "theoretical_loss": 3.5365045256026324, + "tokens_seen": 1539571712 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269468405215647, + "loss": 1.4378, + "theoretical_loss": 3.5364914253722812, + "tokens_seen": 1539637248 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694583751253761, + "loss": 1.4643, + "theoretical_loss": 3.5364783258556685, + "tokens_seen": 1539702784 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026944834503510534, + "loss": 1.354, + "theoretical_loss": 3.536465227052723, + "tokens_seen": 1539768320 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694383149448345, + "loss": 1.4356, + "theoretical_loss": 3.5364521289633775, + "tokens_seen": 1539833856 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694282848545637, + "loss": 1.3865, + "theoretical_loss": 3.5364390315875616, + "tokens_seen": 1539899392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002694182547642929, + "loss": 1.4911, + "theoretical_loss": 3.536425934925206, + "tokens_seen": 1539964928 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026940822467402207, + "loss": 1.4865, + "theoretical_loss": 3.5364128389762417, + "tokens_seen": 1540030464 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 562258, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3878109455108643, + "objective/train/theoretical_loss": 3.5363997437406, + "objective/train/tokens_used": 1560556000, + "theoretical_loss": 3.5363997437406, + "tokens_seen": 1540096000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026939819458375125, + "loss": 1.3803, + "theoretical_loss": 3.5363997437406, + "tokens_seen": 1540096000 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002693881644934805, + "loss": 1.4087, + "theoretical_loss": 3.5363866492182114, + "tokens_seen": 1540161536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002693781344032096, + "loss": 1.5096, + "theoretical_loss": 3.536373555409006, + "tokens_seen": 1540227072 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026936810431293885, + "loss": 1.6357, + "theoretical_loss": 3.536360462312916, + "tokens_seen": 1540292608 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026935807422266803, + "loss": 1.4229, + "theoretical_loss": 3.536347369929871, + "tokens_seen": 1540358144 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002693480441323972, + "loss": 1.5013, + "theoretical_loss": 3.536334278259802, + "tokens_seen": 1540423680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002693380140421264, + "loss": 1.3703, + "theoretical_loss": 3.536321187302641, + "tokens_seen": 1540489216 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026932798395185557, + "loss": 1.4263, + "theoretical_loss": 3.536308097058317, + "tokens_seen": 1540554752 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026931795386158475, + "loss": 1.3719, + "theoretical_loss": 3.5362950075267623, + "tokens_seen": 1540620288 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269307923771314, + "loss": 1.3317, + "theoretical_loss": 3.536281918707907, + "tokens_seen": 1540685824 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692978936810431, + "loss": 1.4378, + "theoretical_loss": 3.5362688306016823, + "tokens_seen": 1540751360 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026928786359077235, + "loss": 1.3839, + "theoretical_loss": 3.5362557432080193, + "tokens_seen": 1540816896 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692778335005015, + "loss": 1.3014, + "theoretical_loss": 3.5362426565268485, + "tokens_seen": 1540882432 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692678034102307, + "loss": 1.3182, + "theoretical_loss": 3.536229570558101, + "tokens_seen": 1540947968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692577733199599, + "loss": 1.4661, + "theoretical_loss": 3.5362164853017077, + "tokens_seen": 1541013504 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692477432296891, + "loss": 1.3797, + "theoretical_loss": 3.5362034007575995, + "tokens_seen": 1541079040 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026923771313941826, + "loss": 1.4119, + "theoretical_loss": 3.5361903169257074, + "tokens_seen": 1541144576 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692276830491475, + "loss": 1.4086, + "theoretical_loss": 3.5361772338059625, + "tokens_seen": 1541210112 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002692176529588766, + "loss": 1.4533, + "theoretical_loss": 3.536164151398295, + "tokens_seen": 1541275648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026920762286860585, + "loss": 1.3993, + "theoretical_loss": 3.536151069702637, + "tokens_seen": 1541341184 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269197592778335, + "loss": 1.3909, + "theoretical_loss": 3.5361379887189184, + "tokens_seen": 1541406720 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691875626880642, + "loss": 1.4492, + "theoretical_loss": 3.5361249084470714, + "tokens_seen": 1541472256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691775325977934, + "loss": 1.3889, + "theoretical_loss": 3.5361118288870257, + "tokens_seen": 1541537792 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691675025075226, + "loss": 1.4373, + "theoretical_loss": 3.536098750038713, + "tokens_seen": 1541603328 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026915747241725176, + "loss": 1.4256, + "theoretical_loss": 3.5360856719020646, + "tokens_seen": 1541668864 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026914744232698094, + "loss": 1.3858, + "theoretical_loss": 3.5360725944770106, + "tokens_seen": 1541734400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691374122367101, + "loss": 1.4418, + "theoretical_loss": 3.5360595177634826, + "tokens_seen": 1541799936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026912738214643936, + "loss": 1.2854, + "theoretical_loss": 3.536046441761412, + "tokens_seen": 1541865472 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691173520561685, + "loss": 1.4814, + "theoretical_loss": 3.5360333664707295, + "tokens_seen": 1541931008 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002691073219658977, + "loss": 1.3913, + "theoretical_loss": 3.536020291891366, + "tokens_seen": 1541996544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026909729187562685, + "loss": 1.3766, + "theoretical_loss": 3.5360072180232525, + "tokens_seen": 1542062080 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002690872617853561, + "loss": 1.3931, + "theoretical_loss": 3.5359941448663204, + "tokens_seen": 1542127616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026907723169508526, + "loss": 1.3459, + "theoretical_loss": 3.535981072420501, + "tokens_seen": 1542193152 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026906720160481444, + "loss": 1.4386, + "theoretical_loss": 3.535968000685725, + "tokens_seen": 1542258688 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002690571715145436, + "loss": 1.4064, + "theoretical_loss": 3.5359549296619237, + "tokens_seen": 1542324224 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026904714142427286, + "loss": 1.4344, + "theoretical_loss": 3.5359418593490277, + "tokens_seen": 1542389760 + }, + { + "epoch": 0.47, + "learning_rate": 0.000269037111334002, + "loss": 1.3981, + "theoretical_loss": 3.535928789746969, + "tokens_seen": 1542455296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002690270812437312, + "loss": 1.4505, + "theoretical_loss": 3.5359157208556775, + "tokens_seen": 1542520832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026901705115346035, + "loss": 1.2927, + "theoretical_loss": 3.5359026526750865, + "tokens_seen": 1542586368 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002690070210631896, + "loss": 1.4898, + "theoretical_loss": 3.5358895852051244, + "tokens_seen": 1542651904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026899699097291877, + "loss": 1.4079, + "theoretical_loss": 3.5358765184457246, + "tokens_seen": 1542717440 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026898696088264795, + "loss": 1.3805, + "theoretical_loss": 3.5358634523968178, + "tokens_seen": 1542782976 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026897693079237713, + "loss": 1.3285, + "theoretical_loss": 3.535850387058334, + "tokens_seen": 1542848512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002689669007021063, + "loss": 1.4998, + "theoretical_loss": 3.535837322430206, + "tokens_seen": 1542914048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002689568706118355, + "loss": 1.3408, + "theoretical_loss": 3.5358242585123634, + "tokens_seen": 1542979584 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002689468405215647, + "loss": 1.425, + "theoretical_loss": 3.5358111953047393, + "tokens_seen": 1543045120 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026893681043129385, + "loss": 1.4452, + "theoretical_loss": 3.535798132807263, + "tokens_seen": 1543110656 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002689267803410231, + "loss": 1.4388, + "theoretical_loss": 3.535785071019867, + "tokens_seen": 1543176192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002689167502507522, + "loss": 1.4674, + "theoretical_loss": 3.535772009942482, + "tokens_seen": 1543241728 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026890672016048145, + "loss": 1.353, + "theoretical_loss": 3.53575894957504, + "tokens_seen": 1543307264 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 563510, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3855209350585938, + "objective/train/theoretical_loss": 3.5357458899174716, + "objective/train/tokens_used": 1563832800, + "theoretical_loss": 3.5357458899174716, + "tokens_seen": 1543372800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026889669007021063, + "loss": 1.2866, + "theoretical_loss": 3.5357458899174716, + "tokens_seen": 1543372800 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002688866599799398, + "loss": 1.4539, + "theoretical_loss": 3.5357328309697076, + "tokens_seen": 1543438336 + }, + { + "epoch": 0.47, + "learning_rate": 0.000268876629889669, + "loss": 1.3751, + "theoretical_loss": 3.5357197727316807, + "tokens_seen": 1543503872 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026886659979939823, + "loss": 1.525, + "theoretical_loss": 3.535706715203321, + "tokens_seen": 1543569408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026885656970912736, + "loss": 1.3266, + "theoretical_loss": 3.5356936583845595, + "tokens_seen": 1543634944 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002688465396188566, + "loss": 1.4874, + "theoretical_loss": 3.5356806022753284, + "tokens_seen": 1543700480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002688365095285857, + "loss": 1.4005, + "theoretical_loss": 3.5356675468755596, + "tokens_seen": 1543766016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026882647943831495, + "loss": 1.3951, + "theoretical_loss": 3.5356544921851834, + "tokens_seen": 1543831552 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026881644934804413, + "loss": 1.3629, + "theoretical_loss": 3.535641438204131, + "tokens_seen": 1543897088 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002688064192577733, + "loss": 1.4532, + "theoretical_loss": 3.535628384932334, + "tokens_seen": 1543962624 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687963891675025, + "loss": 1.2924, + "theoretical_loss": 3.5356153323697246, + "tokens_seen": 1544028160 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687863590772317, + "loss": 1.4091, + "theoretical_loss": 3.5356022805162333, + "tokens_seen": 1544093696 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026877632898696086, + "loss": 1.3684, + "theoretical_loss": 3.5355892293717908, + "tokens_seen": 1544159232 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687662988966901, + "loss": 1.4661, + "theoretical_loss": 3.5355761789363305, + "tokens_seen": 1544224768 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687562688064193, + "loss": 1.436, + "theoretical_loss": 3.5355631292097813, + "tokens_seen": 1544290304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026874623871614846, + "loss": 1.5147, + "theoretical_loss": 3.535550080192077, + "tokens_seen": 1544355840 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687362086258777, + "loss": 1.4239, + "theoretical_loss": 3.535537031883148, + "tokens_seen": 1544421376 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687261785356068, + "loss": 1.5512, + "theoretical_loss": 3.535523984282925, + "tokens_seen": 1544486912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026871614844533605, + "loss": 1.3341, + "theoretical_loss": 3.535510937391341, + "tokens_seen": 1544552448 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002687061183550652, + "loss": 1.387, + "theoretical_loss": 3.5354978912083257, + "tokens_seen": 1544617984 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686960882647944, + "loss": 1.2926, + "theoretical_loss": 3.5354848457338117, + "tokens_seen": 1544683520 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686860581745236, + "loss": 1.3288, + "theoretical_loss": 3.5354718009677306, + "tokens_seen": 1544749056 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686760280842528, + "loss": 1.4233, + "theoretical_loss": 3.535458756910013, + "tokens_seen": 1544814592 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026866599799398196, + "loss": 1.2896, + "theoretical_loss": 3.5354457135605912, + "tokens_seen": 1544880128 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026865596790371114, + "loss": 1.3754, + "theoretical_loss": 3.5354326709193957, + "tokens_seen": 1544945664 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686459378134403, + "loss": 1.451, + "theoretical_loss": 3.5354196289863595, + "tokens_seen": 1545011200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026863590772316956, + "loss": 1.4606, + "theoretical_loss": 3.535406587761413, + "tokens_seen": 1545076736 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686258776328987, + "loss": 1.3465, + "theoretical_loss": 3.535393547244488, + "tokens_seen": 1545142272 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002686158475426279, + "loss": 1.4811, + "theoretical_loss": 3.535380507435516, + "tokens_seen": 1545207808 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026860581745235705, + "loss": 1.3538, + "theoretical_loss": 3.535367468334428, + "tokens_seen": 1545273344 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002685957873620863, + "loss": 1.2869, + "theoretical_loss": 3.5353544299411572, + "tokens_seen": 1545338880 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026858575727181546, + "loss": 1.5129, + "theoretical_loss": 3.535341392255633, + "tokens_seen": 1545404416 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026857572718154464, + "loss": 1.3842, + "theoretical_loss": 3.5353283552777888, + "tokens_seen": 1545469952 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002685656970912738, + "loss": 1.4094, + "theoretical_loss": 3.5353153190075552, + "tokens_seen": 1545535488 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026855566700100306, + "loss": 1.3723, + "theoretical_loss": 3.5353022834448637, + "tokens_seen": 1545601024 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002685456369107322, + "loss": 1.4455, + "theoretical_loss": 3.535289248589647, + "tokens_seen": 1545666560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002685356068204614, + "loss": 1.4597, + "theoretical_loss": 3.5352762144418355, + "tokens_seen": 1545732096 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026852557673019055, + "loss": 1.5186, + "theoretical_loss": 3.535263181001361, + "tokens_seen": 1545797632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002685155466399198, + "loss": 1.6088, + "theoretical_loss": 3.5352501482681555, + "tokens_seen": 1545863168 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026850551654964897, + "loss": 1.4248, + "theoretical_loss": 3.535237116242151, + "tokens_seen": 1545928704 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026849548645937815, + "loss": 1.3943, + "theoretical_loss": 3.5352240849232786, + "tokens_seen": 1545994240 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026848545636910733, + "loss": 1.3795, + "theoretical_loss": 3.5352110543114694, + "tokens_seen": 1546059776 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002684754262788365, + "loss": 1.4194, + "theoretical_loss": 3.535198024406656, + "tokens_seen": 1546125312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002684653961885657, + "loss": 1.3203, + "theoretical_loss": 3.5351849952087697, + "tokens_seen": 1546190848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002684553660982949, + "loss": 1.449, + "theoretical_loss": 3.5351719667177424, + "tokens_seen": 1546256384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026844533600802405, + "loss": 1.5099, + "theoretical_loss": 3.535158938933506, + "tokens_seen": 1546321920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002684353059177533, + "loss": 1.3811, + "theoretical_loss": 3.5351459118559916, + "tokens_seen": 1546387456 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002684252758274824, + "loss": 1.3108, + "theoretical_loss": 3.5351328854851305, + "tokens_seen": 1546452992 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026841524573721165, + "loss": 1.4021, + "theoretical_loss": 3.535119859820856, + "tokens_seen": 1546518528 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026840521564694083, + "loss": 1.2906, + "theoretical_loss": 3.5351068348630985, + "tokens_seen": 1546584064 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 564726, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.366960883140564, + "objective/train/theoretical_loss": 3.53509381061179, + "objective/train/tokens_used": 1567109600, + "theoretical_loss": 3.53509381061179, + "tokens_seen": 1546649600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026839518555667, + "loss": 1.3961, + "theoretical_loss": 3.53509381061179, + "tokens_seen": 1546649600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002683851554663992, + "loss": 1.4888, + "theoretical_loss": 3.535080787066863, + "tokens_seen": 1546715136 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026837512537612843, + "loss": 1.4598, + "theoretical_loss": 3.5350677642282484, + "tokens_seen": 1546780672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026836509528585756, + "loss": 1.3971, + "theoretical_loss": 3.5350547420958778, + "tokens_seen": 1546846208 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002683550651955868, + "loss": 1.4195, + "theoretical_loss": 3.5350417206696836, + "tokens_seen": 1546911744 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002683450351053159, + "loss": 1.4832, + "theoretical_loss": 3.535028699949598, + "tokens_seen": 1546977280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026833500501504515, + "loss": 1.4373, + "theoretical_loss": 3.535015679935552, + "tokens_seen": 1547042816 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026832497492477433, + "loss": 1.477, + "theoretical_loss": 3.535002660627477, + "tokens_seen": 1547108352 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002683149448345035, + "loss": 1.2996, + "theoretical_loss": 3.534989642025306, + "tokens_seen": 1547173888 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002683049147442327, + "loss": 1.45, + "theoretical_loss": 3.53497662412897, + "tokens_seen": 1547239424 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682948846539619, + "loss": 1.4143, + "theoretical_loss": 3.534963606938401, + "tokens_seen": 1547304960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026828485456369106, + "loss": 1.3572, + "theoretical_loss": 3.5349505904535317, + "tokens_seen": 1547370496 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682748244734203, + "loss": 1.4032, + "theoretical_loss": 3.5349375746742924, + "tokens_seen": 1547436032 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682647943831494, + "loss": 1.4994, + "theoretical_loss": 3.5349245596006162, + "tokens_seen": 1547501568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026825476429287866, + "loss": 1.321, + "theoretical_loss": 3.5349115452324344, + "tokens_seen": 1547567104 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682447342026078, + "loss": 1.382, + "theoretical_loss": 3.534898531569679, + "tokens_seen": 1547632640 + }, + { + "epoch": 0.47, + "learning_rate": 0.000268234704112337, + "loss": 1.4291, + "theoretical_loss": 3.534885518612282, + "tokens_seen": 1547698176 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682246740220662, + "loss": 1.5789, + "theoretical_loss": 3.5348725063601756, + "tokens_seen": 1547763712 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002682146439317954, + "loss": 1.4151, + "theoretical_loss": 3.5348594948132908, + "tokens_seen": 1547829248 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026820461384152456, + "loss": 1.5106, + "theoretical_loss": 3.5348464839715605, + "tokens_seen": 1547894784 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681945837512538, + "loss": 1.4904, + "theoretical_loss": 3.534833473834916, + "tokens_seen": 1547960320 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681845536609829, + "loss": 1.5013, + "theoretical_loss": 3.534820464403289, + "tokens_seen": 1548025856 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026817452357071216, + "loss": 1.4769, + "theoretical_loss": 3.5348074556766127, + "tokens_seen": 1548091392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681644934804413, + "loss": 1.3609, + "theoretical_loss": 3.534794447654818, + "tokens_seen": 1548156928 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681544633901705, + "loss": 1.3932, + "theoretical_loss": 3.534781440337836, + "tokens_seen": 1548222464 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681444332998997, + "loss": 1.5466, + "theoretical_loss": 3.5347684337256013, + "tokens_seen": 1548288000 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002681344032096289, + "loss": 1.43, + "theoretical_loss": 3.5347554278180437, + "tokens_seen": 1548353536 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026812437311935807, + "loss": 1.4779, + "theoretical_loss": 3.534742422615096, + "tokens_seen": 1548419072 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026811434302908725, + "loss": 1.4114, + "theoretical_loss": 3.53472941811669, + "tokens_seen": 1548484608 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026810431293881643, + "loss": 1.4048, + "theoretical_loss": 3.534716414322758, + "tokens_seen": 1548550144 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026809428284854566, + "loss": 1.4936, + "theoretical_loss": 3.534703411233232, + "tokens_seen": 1548615680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002680842527582748, + "loss": 1.3183, + "theoretical_loss": 3.5346904088480433, + "tokens_seen": 1548681216 + }, + { + "epoch": 0.47, + "learning_rate": 0.000268074222668004, + "loss": 1.3553, + "theoretical_loss": 3.534677407167125, + "tokens_seen": 1548746752 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026806419257773315, + "loss": 1.3788, + "theoretical_loss": 3.5346644061904082, + "tokens_seen": 1548812288 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002680541624874624, + "loss": 1.405, + "theoretical_loss": 3.534651405917826, + "tokens_seen": 1548877824 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026804413239719157, + "loss": 1.4782, + "theoretical_loss": 3.5346384063493095, + "tokens_seen": 1548943360 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026803410230692075, + "loss": 1.5046, + "theoretical_loss": 3.534625407484791, + "tokens_seen": 1549008896 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026802407221664993, + "loss": 1.2969, + "theoretical_loss": 3.534612409324203, + "tokens_seen": 1549074432 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026801404212637917, + "loss": 1.3232, + "theoretical_loss": 3.534599411867478, + "tokens_seen": 1549139968 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026800401203610835, + "loss": 1.4483, + "theoretical_loss": 3.5345864151145467, + "tokens_seen": 1549205504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026799398194583753, + "loss": 1.4741, + "theoretical_loss": 3.5345734190653424, + "tokens_seen": 1549271040 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679839518555667, + "loss": 1.2533, + "theoretical_loss": 3.534560423719797, + "tokens_seen": 1549336576 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679739217652959, + "loss": 1.478, + "theoretical_loss": 3.5345474290778416, + "tokens_seen": 1549402112 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679638916750251, + "loss": 1.5297, + "theoretical_loss": 3.5345344351394097, + "tokens_seen": 1549467648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026795386158475425, + "loss": 1.3841, + "theoretical_loss": 3.5345214419044333, + "tokens_seen": 1549533184 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679438314944835, + "loss": 1.5721, + "theoretical_loss": 3.534508449372844, + "tokens_seen": 1549598720 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679338014042126, + "loss": 1.4606, + "theoretical_loss": 3.5344954575445744, + "tokens_seen": 1549664256 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026792377131394185, + "loss": 1.3579, + "theoretical_loss": 3.5344824664195564, + "tokens_seen": 1549729792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026791374122367103, + "loss": 1.3901, + "theoretical_loss": 3.534469475997722, + "tokens_seen": 1549795328 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002679037111334002, + "loss": 1.4314, + "theoretical_loss": 3.534456486279004, + "tokens_seen": 1549860864 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 565981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.485386848449707, + "objective/train/theoretical_loss": 3.534443497263335, + "objective/train/tokens_used": 1570386400, + "theoretical_loss": 3.534443497263335, + "tokens_seen": 1549926400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002678936810431294, + "loss": 1.5036, + "theoretical_loss": 3.534443497263335, + "tokens_seen": 1549926400 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026788365095285863, + "loss": 1.3809, + "theoretical_loss": 3.5344305089506456, + "tokens_seen": 1549991936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026787362086258776, + "loss": 1.3003, + "theoretical_loss": 3.5344175213408695, + "tokens_seen": 1550057472 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267863590772317, + "loss": 1.3002, + "theoretical_loss": 3.534404534433938, + "tokens_seen": 1550123008 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002678535606820461, + "loss": 1.4776, + "theoretical_loss": 3.534391548229784, + "tokens_seen": 1550188544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026784353059177535, + "loss": 1.432, + "theoretical_loss": 3.5343785627283397, + "tokens_seen": 1550254080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026783350050150453, + "loss": 1.461, + "theoretical_loss": 3.5343655779295373, + "tokens_seen": 1550319616 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002678234704112337, + "loss": 1.3952, + "theoretical_loss": 3.5343525938333085, + "tokens_seen": 1550385152 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002678134403209629, + "loss": 1.523, + "theoretical_loss": 3.534339610439586, + "tokens_seen": 1550450688 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002678034102306921, + "loss": 1.456, + "theoretical_loss": 3.534326627748303, + "tokens_seen": 1550516224 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026779338014042126, + "loss": 1.3642, + "theoretical_loss": 3.5343136457593904, + "tokens_seen": 1550581760 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002677833500501505, + "loss": 1.4621, + "theoretical_loss": 3.5343006644727817, + "tokens_seen": 1550647296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002677733199598796, + "loss": 1.35, + "theoretical_loss": 3.534287683888408, + "tokens_seen": 1550712832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026776328986960886, + "loss": 1.421, + "theoretical_loss": 3.5342747040062026, + "tokens_seen": 1550778368 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267753259779338, + "loss": 1.3616, + "theoretical_loss": 3.534261724826097, + "tokens_seen": 1550843904 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002677432296890672, + "loss": 1.4018, + "theoretical_loss": 3.5342487463480246, + "tokens_seen": 1550909440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002677331995987964, + "loss": 1.4083, + "theoretical_loss": 3.534235768571917, + "tokens_seen": 1550974976 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002677231695085256, + "loss": 1.3087, + "theoretical_loss": 3.534222791497707, + "tokens_seen": 1551040512 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026771313941825476, + "loss": 1.3205, + "theoretical_loss": 3.534209815125327, + "tokens_seen": 1551106048 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267703109327984, + "loss": 1.3823, + "theoretical_loss": 3.5341968394547085, + "tokens_seen": 1551171584 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002676930792377131, + "loss": 1.3874, + "theoretical_loss": 3.534183864485785, + "tokens_seen": 1551237120 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026768304914744236, + "loss": 1.4193, + "theoretical_loss": 3.5341708902184887, + "tokens_seen": 1551302656 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002676730190571715, + "loss": 1.5105, + "theoretical_loss": 3.5341579166527515, + "tokens_seen": 1551368192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002676629889669007, + "loss": 1.3806, + "theoretical_loss": 3.5341449437885064, + "tokens_seen": 1551433728 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002676529588766299, + "loss": 1.3949, + "theoretical_loss": 3.534131971625685, + "tokens_seen": 1551499264 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002676429287863591, + "loss": 1.3289, + "theoretical_loss": 3.534119000164221, + "tokens_seen": 1551564800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026763289869608827, + "loss": 1.3242, + "theoretical_loss": 3.5341060294040463, + "tokens_seen": 1551630336 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026762286860581745, + "loss": 1.4347, + "theoretical_loss": 3.5340930593450928, + "tokens_seen": 1551695872 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026761283851554663, + "loss": 1.3409, + "theoretical_loss": 3.5340800899872935, + "tokens_seen": 1551761408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026760280842527586, + "loss": 1.3749, + "theoretical_loss": 3.5340671213305814, + "tokens_seen": 1551826944 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267592778335005, + "loss": 1.3581, + "theoretical_loss": 3.5340541533748877, + "tokens_seen": 1551892480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002675827482447342, + "loss": 1.3695, + "theoretical_loss": 3.534041186120146, + "tokens_seen": 1551958016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026757271815446335, + "loss": 1.4245, + "theoretical_loss": 3.534028219566289, + "tokens_seen": 1552023552 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002675626880641926, + "loss": 1.2849, + "theoretical_loss": 3.5340152537132474, + "tokens_seen": 1552089088 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026755265797392177, + "loss": 1.4542, + "theoretical_loss": 3.5340022885609557, + "tokens_seen": 1552154624 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026754262788365095, + "loss": 1.4173, + "theoretical_loss": 3.5339893241093456, + "tokens_seen": 1552220160 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026753259779338013, + "loss": 1.2847, + "theoretical_loss": 3.5339763603583503, + "tokens_seen": 1552285696 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026752256770310937, + "loss": 1.3351, + "theoretical_loss": 3.533963397307901, + "tokens_seen": 1552351232 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002675125376128385, + "loss": 1.33, + "theoretical_loss": 3.533950434957932, + "tokens_seen": 1552416768 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026750250752256773, + "loss": 1.3333, + "theoretical_loss": 3.5339374733083746, + "tokens_seen": 1552482304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026749247743229686, + "loss": 1.369, + "theoretical_loss": 3.5339245123591616, + "tokens_seen": 1552547840 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002674824473420261, + "loss": 1.4496, + "theoretical_loss": 3.533911552110226, + "tokens_seen": 1552613376 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026747241725175527, + "loss": 1.4903, + "theoretical_loss": 3.5338985925615, + "tokens_seen": 1552678912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026746238716148445, + "loss": 1.4615, + "theoretical_loss": 3.533885633712917, + "tokens_seen": 1552744448 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026745235707121363, + "loss": 1.4219, + "theoretical_loss": 3.5338726755644085, + "tokens_seen": 1552809984 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002674423269809428, + "loss": 1.3938, + "theoretical_loss": 3.5338597181159077, + "tokens_seen": 1552875520 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267432296890672, + "loss": 1.3854, + "theoretical_loss": 3.5338467613673474, + "tokens_seen": 1552941056 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026742226680040123, + "loss": 1.3656, + "theoretical_loss": 3.53383380531866, + "tokens_seen": 1553006592 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026741223671013036, + "loss": 1.3531, + "theoretical_loss": 3.5338208499697785, + "tokens_seen": 1553072128 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002674022066198596, + "loss": 1.3534, + "theoretical_loss": 3.5338078953206358, + "tokens_seen": 1553137664 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 567156, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3997645378112793, + "objective/train/theoretical_loss": 3.533794941371163, + "objective/train/tokens_used": 1573663200, + "theoretical_loss": 3.533794941371163, + "tokens_seen": 1553203200 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002673921765295888, + "loss": 1.365, + "theoretical_loss": 3.533794941371163, + "tokens_seen": 1553203200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026738214643931796, + "loss": 1.4034, + "theoretical_loss": 3.5337819881212944, + "tokens_seen": 1553268736 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026737211634904714, + "loss": 1.2768, + "theoretical_loss": 3.533769035570962, + "tokens_seen": 1553334272 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002673620862587763, + "loss": 1.3532, + "theoretical_loss": 3.5337560837200988, + "tokens_seen": 1553399808 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002673520561685055, + "loss": 1.319, + "theoretical_loss": 3.5337431325686377, + "tokens_seen": 1553465344 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026734202607823474, + "loss": 1.4537, + "theoretical_loss": 3.5337301821165106, + "tokens_seen": 1553530880 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026733199598796386, + "loss": 1.4542, + "theoretical_loss": 3.533717232363651, + "tokens_seen": 1553596416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002673219658976931, + "loss": 1.4525, + "theoretical_loss": 3.533704283309992, + "tokens_seen": 1553661952 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002673119358074222, + "loss": 1.2451, + "theoretical_loss": 3.5336913349554653, + "tokens_seen": 1553727488 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026730190571715146, + "loss": 1.3629, + "theoretical_loss": 3.533678387300004, + "tokens_seen": 1553793024 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026729187562688064, + "loss": 1.3775, + "theoretical_loss": 3.5336654403435417, + "tokens_seen": 1553858560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672818455366098, + "loss": 1.3503, + "theoretical_loss": 3.5336524940860103, + "tokens_seen": 1553924096 + }, + { + "epoch": 0.47, + "learning_rate": 0.000267271815446339, + "loss": 1.3802, + "theoretical_loss": 3.533639548527342, + "tokens_seen": 1553989632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672617853560682, + "loss": 1.4044, + "theoretical_loss": 3.5336266036674715, + "tokens_seen": 1554055168 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672517552657974, + "loss": 1.2867, + "theoretical_loss": 3.53361365950633, + "tokens_seen": 1554120704 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672417251755266, + "loss": 1.2449, + "theoretical_loss": 3.5336007160438507, + "tokens_seen": 1554186240 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672316950852558, + "loss": 1.3236, + "theoretical_loss": 3.533587773279967, + "tokens_seen": 1554251776 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026722166499498496, + "loss": 1.4008, + "theoretical_loss": 3.533574831214611, + "tokens_seen": 1554317312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672116349047142, + "loss": 1.3426, + "theoretical_loss": 3.533561889847716, + "tokens_seen": 1554382848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002672016048144433, + "loss": 1.3465, + "theoretical_loss": 3.533548949179215, + "tokens_seen": 1554448384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026719157472417256, + "loss": 1.4522, + "theoretical_loss": 3.5335360092090404, + "tokens_seen": 1554513920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002671815446339017, + "loss": 1.4672, + "theoretical_loss": 3.533523069937125, + "tokens_seen": 1554579456 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002671715145436309, + "loss": 1.483, + "theoretical_loss": 3.5335101313634025, + "tokens_seen": 1554644992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002671614844533601, + "loss": 1.3737, + "theoretical_loss": 3.533497193487805, + "tokens_seen": 1554710528 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002671514543630893, + "loss": 1.3811, + "theoretical_loss": 3.5334842563102655, + "tokens_seen": 1554776064 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026714142427281847, + "loss": 1.4858, + "theoretical_loss": 3.533471319830717, + "tokens_seen": 1554841600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026713139418254765, + "loss": 1.4241, + "theoretical_loss": 3.533458384049093, + "tokens_seen": 1554907136 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026712136409227683, + "loss": 1.273, + "theoretical_loss": 3.5334454489653258, + "tokens_seen": 1554972672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026711133400200606, + "loss": 1.4544, + "theoretical_loss": 3.5334325145793484, + "tokens_seen": 1555038208 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002671013039117352, + "loss": 1.4177, + "theoretical_loss": 3.5334195808910938, + "tokens_seen": 1555103744 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002670912738214644, + "loss": 1.2854, + "theoretical_loss": 3.533406647900495, + "tokens_seen": 1555169280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026708124373119355, + "loss": 1.2683, + "theoretical_loss": 3.5333937156074846, + "tokens_seen": 1555234816 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002670712136409228, + "loss": 1.4818, + "theoretical_loss": 3.533380784011997, + "tokens_seen": 1555300352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026706118355065197, + "loss": 1.4112, + "theoretical_loss": 3.5333678531139627, + "tokens_seen": 1555365888 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026705115346038115, + "loss": 1.3722, + "theoretical_loss": 3.5333549229133174, + "tokens_seen": 1555431424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026704112337011033, + "loss": 1.3716, + "theoretical_loss": 3.5333419934099926, + "tokens_seen": 1555496960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026703109327983957, + "loss": 1.4801, + "theoretical_loss": 3.533329064603921, + "tokens_seen": 1555562496 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002670210631895687, + "loss": 1.4786, + "theoretical_loss": 3.5333161364950367, + "tokens_seen": 1555628032 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026701103309929793, + "loss": 1.4261, + "theoretical_loss": 3.5333032090832717, + "tokens_seen": 1555693568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026700100300902706, + "loss": 1.3893, + "theoretical_loss": 3.53329028236856, + "tokens_seen": 1555759104 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002669909729187563, + "loss": 1.3219, + "theoretical_loss": 3.533277356350834, + "tokens_seen": 1555824640 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026698094282848547, + "loss": 1.1712, + "theoretical_loss": 3.5332644310300276, + "tokens_seen": 1555890176 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026697091273821465, + "loss": 1.3489, + "theoretical_loss": 3.5332515064060726, + "tokens_seen": 1555955712 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026696088264794383, + "loss": 1.3348, + "theoretical_loss": 3.5332385824789028, + "tokens_seen": 1556021248 + }, + { + "epoch": 0.47, + "learning_rate": 0.000266950852557673, + "loss": 1.4431, + "theoretical_loss": 3.5332256592484512, + "tokens_seen": 1556086784 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002669408224674022, + "loss": 1.6179, + "theoretical_loss": 3.5332127367146513, + "tokens_seen": 1556152320 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026693079237713143, + "loss": 1.4662, + "theoretical_loss": 3.5331998148774355, + "tokens_seen": 1556217856 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026692076228686056, + "loss": 1.2553, + "theoretical_loss": 3.533186893736737, + "tokens_seen": 1556283392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002669107321965898, + "loss": 1.3859, + "theoretical_loss": 3.53317397329249, + "tokens_seen": 1556348928 + }, + { + "epoch": 0.47, + "learning_rate": 0.000266900702106319, + "loss": 1.2938, + "theoretical_loss": 3.5331610535446263, + "tokens_seen": 1556414464 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 568457, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2830537557601929, + "objective/train/theoretical_loss": 3.5331481344930795, + "objective/train/tokens_used": 1576940000, + "theoretical_loss": 3.5331481344930795, + "tokens_seen": 1556480000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026689067201604816, + "loss": 1.3262, + "theoretical_loss": 3.5331481344930795, + "tokens_seen": 1556480000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026688064192577734, + "loss": 1.3536, + "theoretical_loss": 3.533135216137783, + "tokens_seen": 1556545536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002668706118355065, + "loss": 1.3574, + "theoretical_loss": 3.53312229847867, + "tokens_seen": 1556611072 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002668605817452357, + "loss": 1.3768, + "theoretical_loss": 3.533109381515673, + "tokens_seen": 1556676608 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026685055165496494, + "loss": 1.3398, + "theoretical_loss": 3.5330964652487253, + "tokens_seen": 1556742144 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026684052156469406, + "loss": 1.5583, + "theoretical_loss": 3.5330835496777615, + "tokens_seen": 1556807680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002668304914744233, + "loss": 1.4447, + "theoretical_loss": 3.533070634802713, + "tokens_seen": 1556873216 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002668204613841524, + "loss": 1.2932, + "theoretical_loss": 3.533057720623514, + "tokens_seen": 1556938752 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026681043129388166, + "loss": 1.4297, + "theoretical_loss": 3.5330448071400973, + "tokens_seen": 1557004288 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026680040120361084, + "loss": 1.3205, + "theoretical_loss": 3.5330318943523964, + "tokens_seen": 1557069824 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026679037111334, + "loss": 1.3383, + "theoretical_loss": 3.5330189822603444, + "tokens_seen": 1557135360 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002667803410230692, + "loss": 1.3952, + "theoretical_loss": 3.5330060708638746, + "tokens_seen": 1557200896 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002667703109327984, + "loss": 1.3795, + "theoretical_loss": 3.5329931601629205, + "tokens_seen": 1557266432 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026676028084252757, + "loss": 1.474, + "theoretical_loss": 3.5329802501574146, + "tokens_seen": 1557331968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002667502507522568, + "loss": 1.3958, + "theoretical_loss": 3.532967340847291, + "tokens_seen": 1557397504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026674022066198593, + "loss": 1.3089, + "theoretical_loss": 3.532954432232483, + "tokens_seen": 1557463040 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026673019057171516, + "loss": 1.3954, + "theoretical_loss": 3.532941524312923, + "tokens_seen": 1557528576 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026672016048144434, + "loss": 1.214, + "theoretical_loss": 3.5329286170885448, + "tokens_seen": 1557594112 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002667101303911735, + "loss": 1.3279, + "theoretical_loss": 3.532915710559282, + "tokens_seen": 1557659648 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002667001003009027, + "loss": 1.3439, + "theoretical_loss": 3.532902804725068, + "tokens_seen": 1557725184 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002666900702106319, + "loss": 1.3519, + "theoretical_loss": 3.532889899585835, + "tokens_seen": 1557790720 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026668004012036107, + "loss": 1.4443, + "theoretical_loss": 3.532876995141518, + "tokens_seen": 1557856256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002666700100300903, + "loss": 1.3509, + "theoretical_loss": 3.5328640913920495, + "tokens_seen": 1557921792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026665997993981943, + "loss": 1.3771, + "theoretical_loss": 3.5328511883373626, + "tokens_seen": 1557987328 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026664994984954867, + "loss": 1.4388, + "theoretical_loss": 3.532838285977391, + "tokens_seen": 1558052864 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002666399197592778, + "loss": 1.2687, + "theoretical_loss": 3.5328253843120674, + "tokens_seen": 1558118400 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026662988966900703, + "loss": 1.3768, + "theoretical_loss": 3.532812483341327, + "tokens_seen": 1558183936 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002666198595787362, + "loss": 1.2835, + "theoretical_loss": 3.532799583065101, + "tokens_seen": 1558249472 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002666098294884654, + "loss": 1.4876, + "theoretical_loss": 3.532786683483324, + "tokens_seen": 1558315008 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026659979939819457, + "loss": 1.4155, + "theoretical_loss": 3.5327737845959293, + "tokens_seen": 1558380544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026658976930792375, + "loss": 1.4596, + "theoretical_loss": 3.5327608864028504, + "tokens_seen": 1558446080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026657973921765293, + "loss": 1.3834, + "theoretical_loss": 3.5327479889040205, + "tokens_seen": 1558511616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026656970912738217, + "loss": 1.3016, + "theoretical_loss": 3.532735092099373, + "tokens_seen": 1558577152 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002665596790371113, + "loss": 1.3877, + "theoretical_loss": 3.5327221959888417, + "tokens_seen": 1558642688 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026654964894684053, + "loss": 1.3373, + "theoretical_loss": 3.5327093005723595, + "tokens_seen": 1558708224 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002665396188565697, + "loss": 1.2464, + "theoretical_loss": 3.53269640584986, + "tokens_seen": 1558773760 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002665295887662989, + "loss": 1.4931, + "theoretical_loss": 3.5326835118212774, + "tokens_seen": 1558839296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002665195586760281, + "loss": 1.469, + "theoretical_loss": 3.532670618486544, + "tokens_seen": 1558904832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026650952858575726, + "loss": 1.2982, + "theoretical_loss": 3.5326577258455947, + "tokens_seen": 1558970368 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002664994984954865, + "loss": 1.3839, + "theoretical_loss": 3.5326448338983623, + "tokens_seen": 1559035904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026648946840521567, + "loss": 1.3202, + "theoretical_loss": 3.5326319426447794, + "tokens_seen": 1559101440 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026647943831494485, + "loss": 1.4316, + "theoretical_loss": 3.5326190520847813, + "tokens_seen": 1559166976 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026646940822467403, + "loss": 1.3583, + "theoretical_loss": 3.5326061622183, + "tokens_seen": 1559232512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002664593781344032, + "loss": 1.3497, + "theoretical_loss": 3.53259327304527, + "tokens_seen": 1559298048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002664493480441324, + "loss": 1.3861, + "theoretical_loss": 3.532580384565624, + "tokens_seen": 1559363584 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026643931795386163, + "loss": 1.3394, + "theoretical_loss": 3.5325674967792966, + "tokens_seen": 1559429120 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026642928786359076, + "loss": 1.3145, + "theoretical_loss": 3.532554609686221, + "tokens_seen": 1559494656 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026641925777332, + "loss": 1.4657, + "theoretical_loss": 3.53254172328633, + "tokens_seen": 1559560192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002664092276830492, + "loss": 1.4953, + "theoretical_loss": 3.5325288375795583, + "tokens_seen": 1559625728 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026639919759277836, + "loss": 1.2722, + "theoretical_loss": 3.5325159525658387, + "tokens_seen": 1559691264 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 569616, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3973890542984009, + "objective/train/theoretical_loss": 3.5325030682451053, + "objective/train/tokens_used": 1580216800, + "theoretical_loss": 3.5325030682451053, + "tokens_seen": 1559756800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026638916750250754, + "loss": 1.4093, + "theoretical_loss": 3.5325030682451053, + "tokens_seen": 1559756800 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002663791374122367, + "loss": 1.5031, + "theoretical_loss": 3.5324901846172914, + "tokens_seen": 1559822336 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002663691073219659, + "loss": 1.4176, + "theoretical_loss": 3.5324773016823308, + "tokens_seen": 1559887872 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026635907723169514, + "loss": 1.408, + "theoretical_loss": 3.5324644194401573, + "tokens_seen": 1559953408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026634904714142426, + "loss": 1.3821, + "theoretical_loss": 3.532451537890704, + "tokens_seen": 1560018944 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002663390170511535, + "loss": 1.3378, + "theoretical_loss": 3.532438657033905, + "tokens_seen": 1560084480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002663289869608826, + "loss": 1.3243, + "theoretical_loss": 3.532425776869694, + "tokens_seen": 1560150016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026631895687061186, + "loss": 1.4491, + "theoretical_loss": 3.532412897398004, + "tokens_seen": 1560215552 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026630892678034104, + "loss": 1.326, + "theoretical_loss": 3.5324000186187696, + "tokens_seen": 1560281088 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002662988966900702, + "loss": 1.4585, + "theoretical_loss": 3.532387140531924, + "tokens_seen": 1560346624 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002662888665997994, + "loss": 1.337, + "theoretical_loss": 3.5323742631374007, + "tokens_seen": 1560412160 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002662788365095286, + "loss": 1.381, + "theoretical_loss": 3.5323613864351344, + "tokens_seen": 1560477696 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026626880641925777, + "loss": 1.4724, + "theoretical_loss": 3.5323485104250576, + "tokens_seen": 1560543232 + }, + { + "epoch": 0.47, + "learning_rate": 0.000266258776328987, + "loss": 1.4269, + "theoretical_loss": 3.5323356351071045, + "tokens_seen": 1560608768 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026624874623871613, + "loss": 1.4404, + "theoretical_loss": 3.532322760481209, + "tokens_seen": 1560674304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026623871614844536, + "loss": 1.3, + "theoretical_loss": 3.5323098865473046, + "tokens_seen": 1560739840 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026622868605817454, + "loss": 1.4474, + "theoretical_loss": 3.532297013305325, + "tokens_seen": 1560805376 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002662186559679037, + "loss": 1.332, + "theoretical_loss": 3.532284140755204, + "tokens_seen": 1560870912 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002662086258776329, + "loss": 1.3543, + "theoretical_loss": 3.532271268896876, + "tokens_seen": 1560936448 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002661985957873621, + "loss": 1.2733, + "theoretical_loss": 3.532258397730274, + "tokens_seen": 1561001984 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026618856569709127, + "loss": 1.319, + "theoretical_loss": 3.532245527255332, + "tokens_seen": 1561067520 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002661785356068205, + "loss": 1.406, + "theoretical_loss": 3.532232657471984, + "tokens_seen": 1561133056 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026616850551654963, + "loss": 1.3372, + "theoretical_loss": 3.5322197883801634, + "tokens_seen": 1561198592 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026615847542627887, + "loss": 1.4035, + "theoretical_loss": 3.5322069199798043, + "tokens_seen": 1561264128 + }, + { + "epoch": 0.47, + "learning_rate": 0.000266148445336008, + "loss": 1.3526, + "theoretical_loss": 3.5321940522708406, + "tokens_seen": 1561329664 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026613841524573723, + "loss": 1.4558, + "theoretical_loss": 3.532181185253206, + "tokens_seen": 1561395200 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002661283851554664, + "loss": 1.4739, + "theoretical_loss": 3.5321683189268342, + "tokens_seen": 1561460736 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002661183550651956, + "loss": 1.4365, + "theoretical_loss": 3.5321554532916597, + "tokens_seen": 1561526272 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026610832497492477, + "loss": 1.4325, + "theoretical_loss": 3.5321425883476154, + "tokens_seen": 1561591808 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026609829488465395, + "loss": 1.4579, + "theoretical_loss": 3.5321297240946357, + "tokens_seen": 1561657344 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026608826479438313, + "loss": 1.3378, + "theoretical_loss": 3.5321168605326543, + "tokens_seen": 1561722880 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026607823470411237, + "loss": 1.3059, + "theoretical_loss": 3.532103997661606, + "tokens_seen": 1561788416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002660682046138415, + "loss": 1.3851, + "theoretical_loss": 3.532091135481423, + "tokens_seen": 1561853952 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026605817452357073, + "loss": 1.4715, + "theoretical_loss": 3.5320782739920404, + "tokens_seen": 1561919488 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002660481444332999, + "loss": 1.4799, + "theoretical_loss": 3.5320654131933917, + "tokens_seen": 1561985024 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002660381143430291, + "loss": 1.4913, + "theoretical_loss": 3.5320525530854114, + "tokens_seen": 1562050560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002660280842527583, + "loss": 1.4518, + "theoretical_loss": 3.532039693668033, + "tokens_seen": 1562116096 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026601805416248746, + "loss": 1.3689, + "theoretical_loss": 3.53202683494119, + "tokens_seen": 1562181632 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026600802407221664, + "loss": 1.3597, + "theoretical_loss": 3.532013976904817, + "tokens_seen": 1562247168 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026599799398194587, + "loss": 1.4477, + "theoretical_loss": 3.5320011195588483, + "tokens_seen": 1562312704 + }, + { + "epoch": 0.47, + "learning_rate": 0.000265987963891675, + "loss": 1.4071, + "theoretical_loss": 3.5319882629032167, + "tokens_seen": 1562378240 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026597793380140423, + "loss": 1.3412, + "theoretical_loss": 3.531975406937857, + "tokens_seen": 1562443776 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026596790371113336, + "loss": 1.3467, + "theoretical_loss": 3.5319625516627027, + "tokens_seen": 1562509312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002659578736208626, + "loss": 1.3107, + "theoretical_loss": 3.531949697077688, + "tokens_seen": 1562574848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002659478435305918, + "loss": 1.3863, + "theoretical_loss": 3.5319368431827476, + "tokens_seen": 1562640384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026593781344032096, + "loss": 1.3671, + "theoretical_loss": 3.5319239899778143, + "tokens_seen": 1562705920 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026592778335005014, + "loss": 1.4052, + "theoretical_loss": 3.5319111374628234, + "tokens_seen": 1562771456 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002659177532597794, + "loss": 1.4306, + "theoretical_loss": 3.5318982856377077, + "tokens_seen": 1562836992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002659077231695085, + "loss": 1.3257, + "theoretical_loss": 3.531885434502402, + "tokens_seen": 1562902528 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026589769307923774, + "loss": 1.4103, + "theoretical_loss": 3.53187258405684, + "tokens_seen": 1562968064 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 570864, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1843256950378418, + "objective/train/theoretical_loss": 3.5318597343009563, + "objective/train/tokens_used": 1583493600, + "theoretical_loss": 3.5318597343009563, + "tokens_seen": 1563033600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026588766298896686, + "loss": 1.4061, + "theoretical_loss": 3.5318597343009563, + "tokens_seen": 1563033600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002658776328986961, + "loss": 1.3715, + "theoretical_loss": 3.5318468852346845, + "tokens_seen": 1563099136 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002658676028084253, + "loss": 1.5278, + "theoretical_loss": 3.5318340368579584, + "tokens_seen": 1563164672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026585757271815446, + "loss": 1.3589, + "theoretical_loss": 3.531821189170713, + "tokens_seen": 1563230208 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026584754262788364, + "loss": 1.2512, + "theoretical_loss": 3.531808342172882, + "tokens_seen": 1563295744 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002658375125376128, + "loss": 1.3384, + "theoretical_loss": 3.5317954958643987, + "tokens_seen": 1563361280 + }, + { + "epoch": 0.47, + "learning_rate": 0.000265827482447342, + "loss": 1.3749, + "theoretical_loss": 3.531782650245198, + "tokens_seen": 1563426816 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026581745235707124, + "loss": 1.2027, + "theoretical_loss": 3.5317698053152142, + "tokens_seen": 1563492352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026580742226680037, + "loss": 1.4491, + "theoretical_loss": 3.5317569610743815, + "tokens_seen": 1563557888 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002657973921765296, + "loss": 1.3635, + "theoretical_loss": 3.531744117522633, + "tokens_seen": 1563623424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026578736208625873, + "loss": 1.3912, + "theoretical_loss": 3.5317312746599034, + "tokens_seen": 1563688960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026577733199598797, + "loss": 1.3744, + "theoretical_loss": 3.5317184324861275, + "tokens_seen": 1563754496 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026576730190571715, + "loss": 1.3878, + "theoretical_loss": 3.531705591001239, + "tokens_seen": 1563820032 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026575727181544633, + "loss": 1.3491, + "theoretical_loss": 3.531692750205172, + "tokens_seen": 1563885568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026574724172517556, + "loss": 1.2909, + "theoretical_loss": 3.5316799100978606, + "tokens_seen": 1563951104 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026573721163490474, + "loss": 1.3343, + "theoretical_loss": 3.531667070679239, + "tokens_seen": 1564016640 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002657271815446339, + "loss": 1.3662, + "theoretical_loss": 3.5316542319492417, + "tokens_seen": 1564082176 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002657171514543631, + "loss": 1.3462, + "theoretical_loss": 3.531641393907803, + "tokens_seen": 1564147712 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002657071213640923, + "loss": 1.3431, + "theoretical_loss": 3.5316285565548564, + "tokens_seen": 1564213248 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026569709127382147, + "loss": 1.4184, + "theoretical_loss": 3.531615719890337, + "tokens_seen": 1564278784 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002656870611835507, + "loss": 1.3507, + "theoretical_loss": 3.5316028839141786, + "tokens_seen": 1564344320 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026567703109327983, + "loss": 1.3008, + "theoretical_loss": 3.5315900486263154, + "tokens_seen": 1564409856 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026566700100300907, + "loss": 1.3597, + "theoretical_loss": 3.5315772140266817, + "tokens_seen": 1564475392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002656569709127382, + "loss": 1.293, + "theoretical_loss": 3.5315643801152117, + "tokens_seen": 1564540928 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026564694082246743, + "loss": 1.3811, + "theoretical_loss": 3.5315515468918406, + "tokens_seen": 1564606464 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002656369107321966, + "loss": 1.4371, + "theoretical_loss": 3.5315387143565014, + "tokens_seen": 1564672000 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002656268806419258, + "loss": 1.2762, + "theoretical_loss": 3.5315258825091282, + "tokens_seen": 1564737536 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026561685055165497, + "loss": 1.3642, + "theoretical_loss": 3.531513051349657, + "tokens_seen": 1564803072 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026560682046138415, + "loss": 1.3193, + "theoretical_loss": 3.531500220878021, + "tokens_seen": 1564868608 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026559679037111333, + "loss": 1.3523, + "theoretical_loss": 3.531487391094154, + "tokens_seen": 1564934144 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026558676028084257, + "loss": 1.4328, + "theoretical_loss": 3.5314745619979915, + "tokens_seen": 1564999680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002655767301905717, + "loss": 1.4318, + "theoretical_loss": 3.5314617335894676, + "tokens_seen": 1565065216 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026556670010030093, + "loss": 1.4866, + "theoretical_loss": 3.5314489058685155, + "tokens_seen": 1565130752 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002655566700100301, + "loss": 1.3421, + "theoretical_loss": 3.531436078835071, + "tokens_seen": 1565196288 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002655466399197593, + "loss": 1.4663, + "theoretical_loss": 3.531423252489068, + "tokens_seen": 1565261824 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002655366098294885, + "loss": 1.4672, + "theoretical_loss": 3.5314104268304405, + "tokens_seen": 1565327360 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026552657973921766, + "loss": 1.3687, + "theoretical_loss": 3.531397601859123, + "tokens_seen": 1565392896 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026551654964894684, + "loss": 1.3263, + "theoretical_loss": 3.5313847775750506, + "tokens_seen": 1565458432 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026550651955867607, + "loss": 1.4254, + "theoretical_loss": 3.5313719539781565, + "tokens_seen": 1565523968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002654964894684052, + "loss": 1.5352, + "theoretical_loss": 3.5313591310683763, + "tokens_seen": 1565589504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026548645937813443, + "loss": 1.421, + "theoretical_loss": 3.5313463088456434, + "tokens_seen": 1565655040 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026547642928786356, + "loss": 1.3327, + "theoretical_loss": 3.5313334873098934, + "tokens_seen": 1565720576 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002654663991975928, + "loss": 1.3201, + "theoretical_loss": 3.5313206664610597, + "tokens_seen": 1565786112 + }, + { + "epoch": 0.47, + "learning_rate": 0.000265456369107322, + "loss": 1.3885, + "theoretical_loss": 3.531307846299077, + "tokens_seen": 1565851648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026544633901705116, + "loss": 1.4188, + "theoretical_loss": 3.53129502682388, + "tokens_seen": 1565917184 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026543630892678034, + "loss": 1.3277, + "theoretical_loss": 3.531282208035403, + "tokens_seen": 1565982720 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002654262788365096, + "loss": 1.4682, + "theoretical_loss": 3.5312693899335805, + "tokens_seen": 1566048256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002654162487462387, + "loss": 1.4151, + "theoretical_loss": 3.5312565725183473, + "tokens_seen": 1566113792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026540621865596794, + "loss": 1.3255, + "theoretical_loss": 3.531243755789637, + "tokens_seen": 1566179328 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026539618856569707, + "loss": 1.2897, + "theoretical_loss": 3.5312309397473856, + "tokens_seen": 1566244864 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 572070, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3877062797546387, + "objective/train/theoretical_loss": 3.531218124391526, + "objective/train/tokens_used": 1586770400, + "theoretical_loss": 3.531218124391526, + "tokens_seen": 1566310400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002653861584754263, + "loss": 1.3331, + "theoretical_loss": 3.531218124391526, + "tokens_seen": 1566310400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002653761283851555, + "loss": 1.3429, + "theoretical_loss": 3.5312053097219938, + "tokens_seen": 1566375936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026536609829488466, + "loss": 1.358, + "theoretical_loss": 3.531192495738723, + "tokens_seen": 1566441472 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026535606820461384, + "loss": 1.3762, + "theoretical_loss": 3.531179682441648, + "tokens_seen": 1566507008 + }, + { + "epoch": 0.47, + "learning_rate": 0.000265346038114343, + "loss": 1.2598, + "theoretical_loss": 3.531166869830704, + "tokens_seen": 1566572544 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002653360080240722, + "loss": 1.3411, + "theoretical_loss": 3.5311540579058254, + "tokens_seen": 1566638080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026532597793380144, + "loss": 1.3953, + "theoretical_loss": 3.5311412466669463, + "tokens_seen": 1566703616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026531594784353057, + "loss": 1.4009, + "theoretical_loss": 3.5311284361140016, + "tokens_seen": 1566769152 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002653059177532598, + "loss": 1.416, + "theoretical_loss": 3.531115626246926, + "tokens_seen": 1566834688 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026529588766298893, + "loss": 1.2751, + "theoretical_loss": 3.531102817065654, + "tokens_seen": 1566900224 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026528585757271817, + "loss": 1.3764, + "theoretical_loss": 3.5310900085701196, + "tokens_seen": 1566965760 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026527582748244735, + "loss": 1.3398, + "theoretical_loss": 3.5310772007602584, + "tokens_seen": 1567031296 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026526579739217653, + "loss": 1.3766, + "theoretical_loss": 3.5310643936360044, + "tokens_seen": 1567096832 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002652557673019057, + "loss": 1.3934, + "theoretical_loss": 3.5310515871972927, + "tokens_seen": 1567162368 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026524573721163494, + "loss": 1.3019, + "theoretical_loss": 3.5310387814440576, + "tokens_seen": 1567227904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026523570712136407, + "loss": 1.375, + "theoretical_loss": 3.5310259763762333, + "tokens_seen": 1567293440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0002652256770310933, + "loss": 1.3311, + "theoretical_loss": 3.5310131719937554, + "tokens_seen": 1567358976 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026521564694082243, + "loss": 1.3691, + "theoretical_loss": 3.531000368296558, + "tokens_seen": 1567424512 + }, + { + "epoch": 0.47, + "learning_rate": 0.00026520561685055167, + "loss": 1.3285, + "theoretical_loss": 3.5309875652845757, + "tokens_seen": 1567490048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026519558676028085, + "loss": 1.4801, + "theoretical_loss": 3.5309747629577437, + "tokens_seen": 1567555584 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026518555667001003, + "loss": 1.3237, + "theoretical_loss": 3.530961961315996, + "tokens_seen": 1567621120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002651755265797392, + "loss": 1.3927, + "theoretical_loss": 3.5309491603592678, + "tokens_seen": 1567686656 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002651654964894684, + "loss": 1.3115, + "theoretical_loss": 3.5309363600874937, + "tokens_seen": 1567752192 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002651554663991976, + "loss": 1.3872, + "theoretical_loss": 3.530923560500608, + "tokens_seen": 1567817728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002651454363089268, + "loss": 1.414, + "theoretical_loss": 3.5309107615985464, + "tokens_seen": 1567883264 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026513540621865594, + "loss": 1.3426, + "theoretical_loss": 3.5308979633812427, + "tokens_seen": 1567948800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026512537612838517, + "loss": 1.3491, + "theoretical_loss": 3.530885165848632, + "tokens_seen": 1568014336 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002651153460381143, + "loss": 1.3862, + "theoretical_loss": 3.530872369000649, + "tokens_seen": 1568079872 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026510531594784353, + "loss": 1.4342, + "theoretical_loss": 3.5308595728372296, + "tokens_seen": 1568145408 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650952858575727, + "loss": 1.4285, + "theoretical_loss": 3.530846777358306, + "tokens_seen": 1568210944 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650852557673019, + "loss": 1.4714, + "theoretical_loss": 3.5308339825638155, + "tokens_seen": 1568276480 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650752256770311, + "loss": 1.3251, + "theoretical_loss": 3.5308211884536913, + "tokens_seen": 1568342016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650651955867603, + "loss": 1.3071, + "theoretical_loss": 3.5308083950278686, + "tokens_seen": 1568407552 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026505516549648944, + "loss": 1.3394, + "theoretical_loss": 3.5307956022862825, + "tokens_seen": 1568473088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650451354062187, + "loss": 1.3818, + "theoretical_loss": 3.5307828102288683, + "tokens_seen": 1568538624 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650351053159478, + "loss": 1.3952, + "theoretical_loss": 3.5307700188555593, + "tokens_seen": 1568604160 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026502507522567704, + "loss": 1.3224, + "theoretical_loss": 3.530757228166292, + "tokens_seen": 1568669696 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650150451354062, + "loss": 1.3634, + "theoretical_loss": 3.5307444381610003, + "tokens_seen": 1568735232 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002650050150451354, + "loss": 1.3832, + "theoretical_loss": 3.5307316488396188, + "tokens_seen": 1568800768 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026499498495486464, + "loss": 1.3825, + "theoretical_loss": 3.530718860202083, + "tokens_seen": 1568866304 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026498495486459376, + "loss": 1.3283, + "theoretical_loss": 3.5307060722483277, + "tokens_seen": 1568931840 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264974924774323, + "loss": 1.3562, + "theoretical_loss": 3.530693284978288, + "tokens_seen": 1568997376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002649648946840522, + "loss": 1.2885, + "theoretical_loss": 3.530680498391898, + "tokens_seen": 1569062912 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026495486459378136, + "loss": 1.4557, + "theoretical_loss": 3.530667712489093, + "tokens_seen": 1569128448 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026494483450351054, + "loss": 1.3439, + "theoretical_loss": 3.5306549272698082, + "tokens_seen": 1569193984 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002649348044132398, + "loss": 1.2852, + "theoretical_loss": 3.5306421427339783, + "tokens_seen": 1569259520 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002649247743229689, + "loss": 1.4201, + "theoretical_loss": 3.530629358881538, + "tokens_seen": 1569325056 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026491474423269814, + "loss": 1.3092, + "theoretical_loss": 3.5306165757124224, + "tokens_seen": 1569390592 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026490471414242727, + "loss": 1.3953, + "theoretical_loss": 3.5306037932265664, + "tokens_seen": 1569456128 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002648946840521565, + "loss": 1.3338, + "theoretical_loss": 3.5305910114239056, + "tokens_seen": 1569521664 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 573246, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3076233863830566, + "objective/train/theoretical_loss": 3.5305782303043736, + "objective/train/tokens_used": 1590047200, + "theoretical_loss": 3.5305782303043736, + "tokens_seen": 1569587200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002648846539618857, + "loss": 1.3948, + "theoretical_loss": 3.5305782303043736, + "tokens_seen": 1569587200 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026487462387161486, + "loss": 1.3703, + "theoretical_loss": 3.5305654498679067, + "tokens_seen": 1569652736 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026486459378134404, + "loss": 1.3907, + "theoretical_loss": 3.5305526701144387, + "tokens_seen": 1569718272 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002648545636910732, + "loss": 1.2463, + "theoretical_loss": 3.5305398910439054, + "tokens_seen": 1569783808 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002648445336008024, + "loss": 1.3711, + "theoretical_loss": 3.5305271126562423, + "tokens_seen": 1569849344 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026483450351053164, + "loss": 1.319, + "theoretical_loss": 3.5305143349513832, + "tokens_seen": 1569914880 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026482447342026077, + "loss": 1.2595, + "theoretical_loss": 3.530501557929264, + "tokens_seen": 1569980416 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026481444332999, + "loss": 1.2901, + "theoretical_loss": 3.5304887815898187, + "tokens_seen": 1570045952 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026480441323971913, + "loss": 1.4589, + "theoretical_loss": 3.530476005932983, + "tokens_seen": 1570111488 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026479438314944837, + "loss": 1.3299, + "theoretical_loss": 3.5304632309586923, + "tokens_seen": 1570177024 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026478435305917755, + "loss": 1.347, + "theoretical_loss": 3.5304504566668813, + "tokens_seen": 1570242560 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026477432296890673, + "loss": 1.4629, + "theoretical_loss": 3.5304376830574844, + "tokens_seen": 1570308096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002647642928786359, + "loss": 1.3257, + "theoretical_loss": 3.530424910130438, + "tokens_seen": 1570373632 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026475426278836514, + "loss": 1.3037, + "theoretical_loss": 3.5304121378856763, + "tokens_seen": 1570439168 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026474423269809427, + "loss": 1.3938, + "theoretical_loss": 3.5303993663231346, + "tokens_seen": 1570504704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002647342026078235, + "loss": 1.2487, + "theoretical_loss": 3.5303865954427476, + "tokens_seen": 1570570240 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026472417251755263, + "loss": 1.27, + "theoretical_loss": 3.530373825244451, + "tokens_seen": 1570635776 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026471414242728187, + "loss": 1.3071, + "theoretical_loss": 3.5303610557281795, + "tokens_seen": 1570701312 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026470411233701105, + "loss": 1.2456, + "theoretical_loss": 3.5303482868938687, + "tokens_seen": 1570766848 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026469408224674023, + "loss": 1.3239, + "theoretical_loss": 3.5303355187414533, + "tokens_seen": 1570832384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002646840521564694, + "loss": 1.2923, + "theoretical_loss": 3.5303227512708686, + "tokens_seen": 1570897920 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002646740220661986, + "loss": 1.3244, + "theoretical_loss": 3.530309984482049, + "tokens_seen": 1570963456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002646639919759278, + "loss": 1.3402, + "theoretical_loss": 3.5302972183749306, + "tokens_seen": 1571028992 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264653961885657, + "loss": 1.3877, + "theoretical_loss": 3.530284452949449, + "tokens_seen": 1571094528 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026464393179538614, + "loss": 1.4284, + "theoretical_loss": 3.5302716882055383, + "tokens_seen": 1571160064 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026463390170511537, + "loss": 1.2785, + "theoretical_loss": 3.530258924143134, + "tokens_seen": 1571225600 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002646238716148445, + "loss": 1.3325, + "theoretical_loss": 3.5302461607621716, + "tokens_seen": 1571291136 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026461384152457373, + "loss": 1.3982, + "theoretical_loss": 3.5302333980625855, + "tokens_seen": 1571356672 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002646038114343029, + "loss": 1.3632, + "theoretical_loss": 3.5302206360443122, + "tokens_seen": 1571422208 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645937813440321, + "loss": 1.399, + "theoretical_loss": 3.5302078747072856, + "tokens_seen": 1571487744 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645837512537613, + "loss": 1.3402, + "theoretical_loss": 3.530195114051442, + "tokens_seen": 1571553280 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645737211634905, + "loss": 1.2686, + "theoretical_loss": 3.5301823540767154, + "tokens_seen": 1571618816 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026456369107321964, + "loss": 1.3095, + "theoretical_loss": 3.5301695947830423, + "tokens_seen": 1571684352 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645536609829489, + "loss": 1.3721, + "theoretical_loss": 3.5301568361703577, + "tokens_seen": 1571749888 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264543630892678, + "loss": 1.4552, + "theoretical_loss": 3.530144078238596, + "tokens_seen": 1571815424 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026453360080240724, + "loss": 1.4186, + "theoretical_loss": 3.5301313209876937, + "tokens_seen": 1571880960 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645235707121364, + "loss": 1.3649, + "theoretical_loss": 3.530118564417585, + "tokens_seen": 1571946496 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645135406218656, + "loss": 1.3251, + "theoretical_loss": 3.530105808528206, + "tokens_seen": 1572012032 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002645035105315948, + "loss": 1.2836, + "theoretical_loss": 3.5300930533194914, + "tokens_seen": 1572077568 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026449348044132396, + "loss": 1.3475, + "theoretical_loss": 3.5300802987913764, + "tokens_seen": 1572143104 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026448345035105314, + "loss": 1.2379, + "theoretical_loss": 3.5300675449437975, + "tokens_seen": 1572208640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002644734202607824, + "loss": 1.4344, + "theoretical_loss": 3.5300547917766885, + "tokens_seen": 1572274176 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002644633901705115, + "loss": 1.3528, + "theoretical_loss": 3.530042039289986, + "tokens_seen": 1572339712 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026445336008024074, + "loss": 1.5047, + "theoretical_loss": 3.5300292874836243, + "tokens_seen": 1572405248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026444332998996987, + "loss": 1.4002, + "theoretical_loss": 3.5300165363575395, + "tokens_seen": 1572470784 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002644332998996991, + "loss": 1.4883, + "theoretical_loss": 3.530003785911666, + "tokens_seen": 1572536320 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002644232698094283, + "loss": 1.3802, + "theoretical_loss": 3.5299910361459403, + "tokens_seen": 1572601856 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026441323971915747, + "loss": 1.4112, + "theoretical_loss": 3.529978287060298, + "tokens_seen": 1572667392 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026440320962888665, + "loss": 1.4108, + "theoretical_loss": 3.5299655386546727, + "tokens_seen": 1572732928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002643931795386159, + "loss": 1.3728, + "theoretical_loss": 3.5299527909290016, + "tokens_seen": 1572798464 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 574423, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4527121782302856, + "objective/train/theoretical_loss": 3.529940043883219, + "objective/train/tokens_used": 1593324000, + "theoretical_loss": 3.529940043883219, + "tokens_seen": 1572864000 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264383149448345, + "loss": 1.473, + "theoretical_loss": 3.529940043883219, + "tokens_seen": 1572864000 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026437311935807424, + "loss": 1.2887, + "theoretical_loss": 3.529927297517261, + "tokens_seen": 1572929536 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026436308926780337, + "loss": 1.2845, + "theoretical_loss": 3.5299145518310624, + "tokens_seen": 1572995072 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002643530591775326, + "loss": 1.3937, + "theoretical_loss": 3.5299018068245593, + "tokens_seen": 1573060608 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002643430290872618, + "loss": 1.3153, + "theoretical_loss": 3.529889062497687, + "tokens_seen": 1573126144 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026433299899699097, + "loss": 1.3326, + "theoretical_loss": 3.5298763188503806, + "tokens_seen": 1573191680 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026432296890672015, + "loss": 1.2281, + "theoretical_loss": 3.529863575882575, + "tokens_seen": 1573257216 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026431293881644933, + "loss": 1.2919, + "theoretical_loss": 3.529850833594207, + "tokens_seen": 1573322752 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002643029087261785, + "loss": 1.3713, + "theoretical_loss": 3.5298380919852113, + "tokens_seen": 1573388288 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026429287863590775, + "loss": 1.4068, + "theoretical_loss": 3.529825351055524, + "tokens_seen": 1573453824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002642828485456369, + "loss": 1.362, + "theoretical_loss": 3.529812610805079, + "tokens_seen": 1573519360 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002642728184553661, + "loss": 1.2135, + "theoretical_loss": 3.529799871233814, + "tokens_seen": 1573584896 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026426278836509534, + "loss": 1.2826, + "theoretical_loss": 3.529787132341663, + "tokens_seen": 1573650432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026425275827482447, + "loss": 1.3218, + "theoretical_loss": 3.529774394128562, + "tokens_seen": 1573715968 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002642427281845537, + "loss": 1.5533, + "theoretical_loss": 3.529761656594447, + "tokens_seen": 1573781504 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026423269809428283, + "loss": 1.2686, + "theoretical_loss": 3.5297489197392524, + "tokens_seen": 1573847040 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026422266800401207, + "loss": 1.2559, + "theoretical_loss": 3.529736183562914, + "tokens_seen": 1573912576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026421263791374125, + "loss": 1.3663, + "theoretical_loss": 3.5297234480653685, + "tokens_seen": 1573978112 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026420260782347043, + "loss": 1.4257, + "theoretical_loss": 3.5297107132465504, + "tokens_seen": 1574043648 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641925777331996, + "loss": 1.3418, + "theoretical_loss": 3.529697979106395, + "tokens_seen": 1574109184 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641825476429288, + "loss": 1.2543, + "theoretical_loss": 3.5296852456448393, + "tokens_seen": 1574174720 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264172517552658, + "loss": 1.3836, + "theoretical_loss": 3.5296725128618176, + "tokens_seen": 1574240256 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641624874623872, + "loss": 1.3718, + "theoretical_loss": 3.5296597807572656, + "tokens_seen": 1574305792 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026415245737211634, + "loss": 1.3353, + "theoretical_loss": 3.5296470493311194, + "tokens_seen": 1574371328 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026414242728184557, + "loss": 1.4064, + "theoretical_loss": 3.529634318583315, + "tokens_seen": 1574436864 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641323971915747, + "loss": 1.2735, + "theoretical_loss": 3.5296215885137867, + "tokens_seen": 1574502400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026412236710130393, + "loss": 1.3901, + "theoretical_loss": 3.5296088591224715, + "tokens_seen": 1574567936 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641123370110331, + "loss": 1.2251, + "theoretical_loss": 3.529596130409304, + "tokens_seen": 1574633472 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002641023069207623, + "loss": 1.4176, + "theoretical_loss": 3.5295834023742203, + "tokens_seen": 1574699008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640922768304915, + "loss": 1.1388, + "theoretical_loss": 3.529570675017156, + "tokens_seen": 1574764544 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640822467402207, + "loss": 1.1767, + "theoretical_loss": 3.529557948338047, + "tokens_seen": 1574830080 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026407221664994984, + "loss": 1.3485, + "theoretical_loss": 3.5295452223368287, + "tokens_seen": 1574895616 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640621865596791, + "loss": 1.3847, + "theoretical_loss": 3.529532497013437, + "tokens_seen": 1574961152 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640521564694082, + "loss": 1.398, + "theoretical_loss": 3.5295197723678076, + "tokens_seen": 1575026688 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026404212637913744, + "loss": 1.3106, + "theoretical_loss": 3.5295070483998754, + "tokens_seen": 1575092224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640320962888666, + "loss": 1.2511, + "theoretical_loss": 3.529494325109577, + "tokens_seen": 1575157760 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002640220661985958, + "loss": 1.3561, + "theoretical_loss": 3.5294816024968485, + "tokens_seen": 1575223296 + }, + { + "epoch": 0.48, + "learning_rate": 0.000264012036108325, + "loss": 1.4077, + "theoretical_loss": 3.529468880561624, + "tokens_seen": 1575288832 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026400200601805416, + "loss": 1.4588, + "theoretical_loss": 3.529456159303841, + "tokens_seen": 1575354368 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026399197592778334, + "loss": 1.2742, + "theoretical_loss": 3.5294434387234346, + "tokens_seen": 1575419904 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002639819458375126, + "loss": 1.366, + "theoretical_loss": 3.52943071882034, + "tokens_seen": 1575485440 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002639719157472417, + "loss": 1.2591, + "theoretical_loss": 3.529417999594494, + "tokens_seen": 1575550976 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026396188565697094, + "loss": 1.3482, + "theoretical_loss": 3.529405281045831, + "tokens_seen": 1575616512 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026395185556670007, + "loss": 1.3327, + "theoretical_loss": 3.529392563174288, + "tokens_seen": 1575682048 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002639418254764293, + "loss": 1.382, + "theoretical_loss": 3.5293798459798005, + "tokens_seen": 1575747584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002639317953861585, + "loss": 1.188, + "theoretical_loss": 3.529367129462304, + "tokens_seen": 1575813120 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026392176529588767, + "loss": 1.3286, + "theoretical_loss": 3.5293544136217347, + "tokens_seen": 1575878656 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026391173520561685, + "loss": 1.3631, + "theoretical_loss": 3.529341698458028, + "tokens_seen": 1575944192 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002639017051153461, + "loss": 1.3375, + "theoretical_loss": 3.52932898397112, + "tokens_seen": 1576009728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002638916750250752, + "loss": 1.3917, + "theoretical_loss": 3.529316270160946, + "tokens_seen": 1576075264 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 575737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2089872360229492, + "objective/train/theoretical_loss": 3.529303557027443, + "objective/train/tokens_used": 1596600800, + "theoretical_loss": 3.529303557027443, + "tokens_seen": 1576140800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026388164493480444, + "loss": 1.2692, + "theoretical_loss": 3.529303557027443, + "tokens_seen": 1576140800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026387161484453357, + "loss": 1.435, + "theoretical_loss": 3.5292908445705455, + "tokens_seen": 1576206336 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002638615847542628, + "loss": 1.2649, + "theoretical_loss": 3.5292781327901905, + "tokens_seen": 1576271872 + }, + { + "epoch": 0.48, + "learning_rate": 0.000263851554663992, + "loss": 1.292, + "theoretical_loss": 3.5292654216863135, + "tokens_seen": 1576337408 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026384152457372117, + "loss": 1.3817, + "theoretical_loss": 3.5292527112588497, + "tokens_seen": 1576402944 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026383149448345035, + "loss": 1.367, + "theoretical_loss": 3.529240001507736, + "tokens_seen": 1576468480 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026382146439317953, + "loss": 1.3263, + "theoretical_loss": 3.529227292432908, + "tokens_seen": 1576534016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002638114343029087, + "loss": 1.328, + "theoretical_loss": 3.5292145840343006, + "tokens_seen": 1576599552 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026380140421263795, + "loss": 1.3269, + "theoretical_loss": 3.529201876311851, + "tokens_seen": 1576665088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637913741223671, + "loss": 1.302, + "theoretical_loss": 3.5291891692654946, + "tokens_seen": 1576730624 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637813440320963, + "loss": 1.3443, + "theoretical_loss": 3.529176462895168, + "tokens_seen": 1576796160 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637713139418255, + "loss": 1.4058, + "theoretical_loss": 3.529163757200806, + "tokens_seen": 1576861696 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026376128385155467, + "loss": 1.2787, + "theoretical_loss": 3.5291510521823453, + "tokens_seen": 1576927232 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026375125376128385, + "loss": 1.3715, + "theoretical_loss": 3.529138347839721, + "tokens_seen": 1576992768 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026374122367101303, + "loss": 1.3509, + "theoretical_loss": 3.5291256441728702, + "tokens_seen": 1577058304 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637311935807422, + "loss": 1.3111, + "theoretical_loss": 3.5291129411817286, + "tokens_seen": 1577123840 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026372116349047145, + "loss": 1.3036, + "theoretical_loss": 3.5291002388662323, + "tokens_seen": 1577189376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637111334002006, + "loss": 1.3319, + "theoretical_loss": 3.5290875372263164, + "tokens_seen": 1577254912 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002637011033099298, + "loss": 1.2928, + "theoretical_loss": 3.529074836261917, + "tokens_seen": 1577320448 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026369107321965894, + "loss": 1.2972, + "theoretical_loss": 3.5290621359729712, + "tokens_seen": 1577385984 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636810431293882, + "loss": 1.306, + "theoretical_loss": 3.5290494363594145, + "tokens_seen": 1577451520 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026367101303911736, + "loss": 1.3186, + "theoretical_loss": 3.5290367374211824, + "tokens_seen": 1577517056 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026366098294884654, + "loss": 1.3997, + "theoretical_loss": 3.5290240391582115, + "tokens_seen": 1577582592 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636509528585757, + "loss": 1.3645, + "theoretical_loss": 3.5290113415704383, + "tokens_seen": 1577648128 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636409227683049, + "loss": 1.4268, + "theoretical_loss": 3.5289986446577974, + "tokens_seen": 1577713664 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636308926780341, + "loss": 1.3833, + "theoretical_loss": 3.528985948420226, + "tokens_seen": 1577779200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636208625877633, + "loss": 1.3298, + "theoretical_loss": 3.52897325285766, + "tokens_seen": 1577844736 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026361083249749244, + "loss": 1.3392, + "theoretical_loss": 3.528960557970035, + "tokens_seen": 1577910272 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002636008024072217, + "loss": 1.2218, + "theoretical_loss": 3.5289478637572875, + "tokens_seen": 1577975808 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026359077231695086, + "loss": 1.3593, + "theoretical_loss": 3.528935170219354, + "tokens_seen": 1578041344 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026358074222668004, + "loss": 1.2836, + "theoretical_loss": 3.5289224773561703, + "tokens_seen": 1578106880 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002635707121364092, + "loss": 1.328, + "theoretical_loss": 3.528909785167672, + "tokens_seen": 1578172416 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002635606820461384, + "loss": 1.3384, + "theoretical_loss": 3.528897093653795, + "tokens_seen": 1578237952 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002635506519558676, + "loss": 1.2651, + "theoretical_loss": 3.5288844028144766, + "tokens_seen": 1578303488 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002635406218655968, + "loss": 1.3269, + "theoretical_loss": 3.5288717126496523, + "tokens_seen": 1578369024 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026353059177532595, + "loss": 1.3865, + "theoretical_loss": 3.5288590231592583, + "tokens_seen": 1578434560 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002635205616850552, + "loss": 1.383, + "theoretical_loss": 3.528846334343231, + "tokens_seen": 1578500096 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026351053159478436, + "loss": 1.3105, + "theoretical_loss": 3.528833646201506, + "tokens_seen": 1578565632 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026350050150451354, + "loss": 1.4098, + "theoretical_loss": 3.52882095873402, + "tokens_seen": 1578631168 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634904714142428, + "loss": 1.3678, + "theoretical_loss": 3.528808271940709, + "tokens_seen": 1578696704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634804413239719, + "loss": 1.4037, + "theoretical_loss": 3.5287955858215083, + "tokens_seen": 1578762240 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026347041123370114, + "loss": 1.3633, + "theoretical_loss": 3.5287829003763562, + "tokens_seen": 1578827776 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026346038114343027, + "loss": 1.4277, + "theoretical_loss": 3.5287702156051868, + "tokens_seen": 1578893312 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634503510531595, + "loss": 1.3571, + "theoretical_loss": 3.528757531507938, + "tokens_seen": 1578958848 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634403209628887, + "loss": 1.3979, + "theoretical_loss": 3.5287448480845445, + "tokens_seen": 1579024384 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026343029087261787, + "loss": 1.4466, + "theoretical_loss": 3.5287321653349437, + "tokens_seen": 1579089920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026342026078234705, + "loss": 1.3753, + "theoretical_loss": 3.5287194832590707, + "tokens_seen": 1579155456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634102306920763, + "loss": 1.3454, + "theoretical_loss": 3.5287068018568633, + "tokens_seen": 1579220992 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002634002006018054, + "loss": 1.3817, + "theoretical_loss": 3.5286941211282565, + "tokens_seen": 1579286528 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026339017051153464, + "loss": 1.3704, + "theoretical_loss": 3.5286814410731875, + "tokens_seen": 1579352064 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 576827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3818645477294922, + "objective/train/theoretical_loss": 3.5286687616915913, + "objective/train/tokens_used": 1599877600, + "theoretical_loss": 3.5286687616915913, + "tokens_seen": 1579417600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026338014042126377, + "loss": 1.3969, + "theoretical_loss": 3.5286687616915913, + "tokens_seen": 1579417600 + }, + { + "epoch": 0.48, + "learning_rate": 0.000263370110330993, + "loss": 1.4312, + "theoretical_loss": 3.5286560829834057, + "tokens_seen": 1579483136 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002633600802407222, + "loss": 1.28, + "theoretical_loss": 3.5286434049485655, + "tokens_seen": 1579548672 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026335005015045137, + "loss": 1.2845, + "theoretical_loss": 3.528630727587008, + "tokens_seen": 1579614208 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026334002006018055, + "loss": 1.3533, + "theoretical_loss": 3.52861805089867, + "tokens_seen": 1579679744 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026332998996990973, + "loss": 1.3533, + "theoretical_loss": 3.5286053748834862, + "tokens_seen": 1579745280 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002633199598796389, + "loss": 1.2242, + "theoretical_loss": 3.5285926995413943, + "tokens_seen": 1579810816 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026330992978936815, + "loss": 1.4922, + "theoretical_loss": 3.52858002487233, + "tokens_seen": 1579876352 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002632998996990973, + "loss": 1.4124, + "theoretical_loss": 3.5285673508762296, + "tokens_seen": 1579941888 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002632898696088265, + "loss": 1.3346, + "theoretical_loss": 3.52855467755303, + "tokens_seen": 1580007424 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002632798395185557, + "loss": 1.3556, + "theoretical_loss": 3.528542004902667, + "tokens_seen": 1580072960 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026326980942828487, + "loss": 1.3856, + "theoretical_loss": 3.528529332925077, + "tokens_seen": 1580138496 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026325977933801405, + "loss": 1.3138, + "theoretical_loss": 3.5285166616201966, + "tokens_seen": 1580204032 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026324974924774323, + "loss": 1.349, + "theoretical_loss": 3.5285039909879625, + "tokens_seen": 1580269568 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002632397191574724, + "loss": 1.4015, + "theoretical_loss": 3.5284913210283113, + "tokens_seen": 1580335104 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026322968906720165, + "loss": 1.2915, + "theoretical_loss": 3.528478651741178, + "tokens_seen": 1580400640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002632196589769308, + "loss": 1.4171, + "theoretical_loss": 3.5284659831265, + "tokens_seen": 1580466176 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026320962888666, + "loss": 1.3249, + "theoretical_loss": 3.528453315184214, + "tokens_seen": 1580531712 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026319959879638914, + "loss": 1.5432, + "theoretical_loss": 3.528440647914256, + "tokens_seen": 1580597248 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631895687061184, + "loss": 1.4914, + "theoretical_loss": 3.528427981316562, + "tokens_seen": 1580662784 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026317953861584756, + "loss": 1.3707, + "theoretical_loss": 3.5284153153910696, + "tokens_seen": 1580728320 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026316950852557674, + "loss": 1.3469, + "theoretical_loss": 3.528402650137714, + "tokens_seen": 1580793856 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631594784353059, + "loss": 1.3897, + "theoretical_loss": 3.528389985556433, + "tokens_seen": 1580859392 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631494483450351, + "loss": 1.4149, + "theoretical_loss": 3.528377321647162, + "tokens_seen": 1580924928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631394182547643, + "loss": 1.3211, + "theoretical_loss": 3.5283646584098376, + "tokens_seen": 1580990464 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631293881644935, + "loss": 1.3924, + "theoretical_loss": 3.528351995844397, + "tokens_seen": 1581056000 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026311935807422264, + "loss": 1.3273, + "theoretical_loss": 3.5283393339507763, + "tokens_seen": 1581121536 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002631093279839519, + "loss": 1.4877, + "theoretical_loss": 3.5283266727289115, + "tokens_seen": 1581187072 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026309929789368106, + "loss": 1.3525, + "theoretical_loss": 3.5283140121787397, + "tokens_seen": 1581252608 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026308926780341024, + "loss": 1.3157, + "theoretical_loss": 3.5283013523001974, + "tokens_seen": 1581318144 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002630792377131394, + "loss": 1.3402, + "theoretical_loss": 3.528288693093221, + "tokens_seen": 1581383680 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002630692076228686, + "loss": 1.3191, + "theoretical_loss": 3.528276034557747, + "tokens_seen": 1581449216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002630591775325978, + "loss": 1.4011, + "theoretical_loss": 3.5282633766937126, + "tokens_seen": 1581514752 + }, + { + "epoch": 0.48, + "learning_rate": 0.000263049147442327, + "loss": 1.2876, + "theoretical_loss": 3.5282507195010533, + "tokens_seen": 1581580288 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026303911735205615, + "loss": 1.4098, + "theoretical_loss": 3.528238062979706, + "tokens_seen": 1581645824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002630290872617854, + "loss": 1.375, + "theoretical_loss": 3.5282254071296077, + "tokens_seen": 1581711360 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002630190571715145, + "loss": 1.3149, + "theoretical_loss": 3.528212751950695, + "tokens_seen": 1581776896 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026300902708124374, + "loss": 1.3871, + "theoretical_loss": 3.528200097442904, + "tokens_seen": 1581842432 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002629989969909729, + "loss": 1.3666, + "theoretical_loss": 3.5281874436061713, + "tokens_seen": 1581907968 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002629889669007021, + "loss": 1.4822, + "theoretical_loss": 3.5281747904404344, + "tokens_seen": 1581973504 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002629789368104313, + "loss": 1.3518, + "theoretical_loss": 3.528162137945629, + "tokens_seen": 1582039040 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026296890672016047, + "loss": 1.4129, + "theoretical_loss": 3.528149486121692, + "tokens_seen": 1582104576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026295887662988965, + "loss": 1.3269, + "theoretical_loss": 3.52813683496856, + "tokens_seen": 1582170112 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002629488465396189, + "loss": 1.3349, + "theoretical_loss": 3.52812418448617, + "tokens_seen": 1582235648 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262938816449348, + "loss": 1.4647, + "theoretical_loss": 3.528111534674458, + "tokens_seen": 1582301184 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026292878635907725, + "loss": 1.4466, + "theoretical_loss": 3.5280988855333613, + "tokens_seen": 1582366720 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026291875626880643, + "loss": 1.3435, + "theoretical_loss": 3.5280862370628165, + "tokens_seen": 1582432256 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002629087261785356, + "loss": 1.3791, + "theoretical_loss": 3.5280735892627595, + "tokens_seen": 1582497792 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628986960882648, + "loss": 1.4839, + "theoretical_loss": 3.528060942133128, + "tokens_seen": 1582563328 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026288866599799397, + "loss": 1.4254, + "theoretical_loss": 3.528048295673859, + "tokens_seen": 1582628864 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 578056, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1964266300201416, + "objective/train/theoretical_loss": 3.5280356498848873, + "objective/train/tokens_used": 1603154400, + "theoretical_loss": 3.5280356498848873, + "tokens_seen": 1582694400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026287863590772315, + "loss": 1.2554, + "theoretical_loss": 3.5280356498848873, + "tokens_seen": 1582694400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628686058174524, + "loss": 1.4376, + "theoretical_loss": 3.528023004766151, + "tokens_seen": 1582759936 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628585757271815, + "loss": 1.3926, + "theoretical_loss": 3.5280103603175874, + "tokens_seen": 1582825472 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026284854563691075, + "loss": 1.3372, + "theoretical_loss": 3.527997716539132, + "tokens_seen": 1582891008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628385155466399, + "loss": 1.3962, + "theoretical_loss": 3.527985073430722, + "tokens_seen": 1582956544 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628284854563691, + "loss": 1.3317, + "theoretical_loss": 3.5279724309922944, + "tokens_seen": 1583022080 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628184553660983, + "loss": 1.3215, + "theoretical_loss": 3.527959789223786, + "tokens_seen": 1583087616 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002628084252758275, + "loss": 1.3212, + "theoretical_loss": 3.5279471481251328, + "tokens_seen": 1583153152 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026279839518555666, + "loss": 1.2993, + "theoretical_loss": 3.5279345076962727, + "tokens_seen": 1583218688 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002627883650952859, + "loss": 1.4009, + "theoretical_loss": 3.527921867937142, + "tokens_seen": 1583284224 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262778335005015, + "loss": 1.4219, + "theoretical_loss": 3.527909228847676, + "tokens_seen": 1583349760 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026276830491474425, + "loss": 1.3817, + "theoretical_loss": 3.5278965904278143, + "tokens_seen": 1583415296 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026275827482447343, + "loss": 1.4208, + "theoretical_loss": 3.5278839526774917, + "tokens_seen": 1583480832 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002627482447342026, + "loss": 1.4082, + "theoretical_loss": 3.527871315596646, + "tokens_seen": 1583546368 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026273821464393185, + "loss": 1.3662, + "theoretical_loss": 3.527858679185213, + "tokens_seen": 1583611904 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262728184553661, + "loss": 1.378, + "theoretical_loss": 3.527846043443131, + "tokens_seen": 1583677440 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002627181544633902, + "loss": 1.3665, + "theoretical_loss": 3.527833408370336, + "tokens_seen": 1583742976 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026270812437311934, + "loss": 1.4231, + "theoretical_loss": 3.527820773966764, + "tokens_seen": 1583808512 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626980942828486, + "loss": 1.394, + "theoretical_loss": 3.5278081402323536, + "tokens_seen": 1583874048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026268806419257776, + "loss": 1.3826, + "theoretical_loss": 3.5277955071670406, + "tokens_seen": 1583939584 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026267803410230694, + "loss": 1.397, + "theoretical_loss": 3.5277828747707622, + "tokens_seen": 1584005120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626680040120361, + "loss": 1.3276, + "theoretical_loss": 3.527770243043455, + "tokens_seen": 1584070656 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626579739217653, + "loss": 1.3745, + "theoretical_loss": 3.5277576119850567, + "tokens_seen": 1584136192 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626479438314945, + "loss": 1.4555, + "theoretical_loss": 3.527744981595503, + "tokens_seen": 1584201728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626379137412237, + "loss": 1.2486, + "theoretical_loss": 3.5277323518747314, + "tokens_seen": 1584267264 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026262788365095284, + "loss": 1.3469, + "theoretical_loss": 3.5277197228226793, + "tokens_seen": 1584332800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002626178535606821, + "loss": 1.4621, + "theoretical_loss": 3.5277070944392825, + "tokens_seen": 1584398336 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026260782347041126, + "loss": 1.3537, + "theoretical_loss": 3.5276944667244794, + "tokens_seen": 1584463872 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026259779338014044, + "loss": 1.3528, + "theoretical_loss": 3.527681839678206, + "tokens_seen": 1584529408 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625877632898696, + "loss": 1.3955, + "theoretical_loss": 3.527669213300399, + "tokens_seen": 1584594944 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625777331995988, + "loss": 1.4485, + "theoretical_loss": 3.5276565875909967, + "tokens_seen": 1584660480 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262567703109328, + "loss": 1.3387, + "theoretical_loss": 3.5276439625499343, + "tokens_seen": 1584726016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625576730190572, + "loss": 1.3358, + "theoretical_loss": 3.5276313381771502, + "tokens_seen": 1584791552 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026254764292878635, + "loss": 1.398, + "theoretical_loss": 3.5276187144725806, + "tokens_seen": 1584857088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625376128385156, + "loss": 1.2999, + "theoretical_loss": 3.527606091436163, + "tokens_seen": 1584922624 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625275827482447, + "loss": 1.3713, + "theoretical_loss": 3.527593469067834, + "tokens_seen": 1584988160 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026251755265797394, + "loss": 1.3599, + "theoretical_loss": 3.5275808473675307, + "tokens_seen": 1585053696 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002625075225677031, + "loss": 1.3262, + "theoretical_loss": 3.52756822633519, + "tokens_seen": 1585119232 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002624974924774323, + "loss": 1.3165, + "theoretical_loss": 3.52755560597075, + "tokens_seen": 1585184768 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002624874623871615, + "loss": 1.2851, + "theoretical_loss": 3.5275429862741463, + "tokens_seen": 1585250304 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026247743229689067, + "loss": 1.4802, + "theoretical_loss": 3.5275303672453164, + "tokens_seen": 1585315840 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026246740220661985, + "loss": 1.4965, + "theoretical_loss": 3.5275177488841982, + "tokens_seen": 1585381376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002624573721163491, + "loss": 1.2387, + "theoretical_loss": 3.5275051311907277, + "tokens_seen": 1585446912 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002624473420260782, + "loss": 1.2706, + "theoretical_loss": 3.5274925141648423, + "tokens_seen": 1585512448 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026243731193580745, + "loss": 1.2873, + "theoretical_loss": 3.5274798978064794, + "tokens_seen": 1585577984 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026242728184553663, + "loss": 1.3735, + "theoretical_loss": 3.527467282115576, + "tokens_seen": 1585643520 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002624172517552658, + "loss": 1.3426, + "theoretical_loss": 3.5274546670920683, + "tokens_seen": 1585709056 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262407221664995, + "loss": 1.3688, + "theoretical_loss": 3.5274420527358945, + "tokens_seen": 1585774592 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026239719157472417, + "loss": 1.405, + "theoretical_loss": 3.5274294390469914, + "tokens_seen": 1585840128 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026238716148445335, + "loss": 1.2697, + "theoretical_loss": 3.5274168260252967, + "tokens_seen": 1585905664 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 579278, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2067229747772217, + "objective/train/theoretical_loss": 3.527404213670746, + "objective/train/tokens_used": 1606431200, + "theoretical_loss": 3.527404213670746, + "tokens_seen": 1585971200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623771313941826, + "loss": 1.2758, + "theoretical_loss": 3.527404213670746, + "tokens_seen": 1585971200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623671013039117, + "loss": 1.3727, + "theoretical_loss": 3.527391601983278, + "tokens_seen": 1586036736 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026235707121364095, + "loss": 1.4052, + "theoretical_loss": 3.527378990962829, + "tokens_seen": 1586102272 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623470411233701, + "loss": 1.4291, + "theoretical_loss": 3.5273663806093367, + "tokens_seen": 1586167808 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623370110330993, + "loss": 1.2544, + "theoretical_loss": 3.527353770922738, + "tokens_seen": 1586233344 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623269809428285, + "loss": 1.4784, + "theoretical_loss": 3.52734116190297, + "tokens_seen": 1586298880 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002623169508525577, + "loss": 1.3113, + "theoretical_loss": 3.5273285535499697, + "tokens_seen": 1586364416 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026230692076228686, + "loss": 1.3427, + "theoretical_loss": 3.5273159458636747, + "tokens_seen": 1586429952 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622968906720161, + "loss": 1.3219, + "theoretical_loss": 3.527303338844022, + "tokens_seen": 1586495488 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622868605817452, + "loss": 1.4708, + "theoretical_loss": 3.527290732490949, + "tokens_seen": 1586561024 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026227683049147445, + "loss": 1.2819, + "theoretical_loss": 3.5272781268043927, + "tokens_seen": 1586626560 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622668004012036, + "loss": 1.3182, + "theoretical_loss": 3.5272655217842908, + "tokens_seen": 1586692096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622567703109328, + "loss": 1.4337, + "theoretical_loss": 3.5272529174305802, + "tokens_seen": 1586757632 + }, + { + "epoch": 0.48, + "learning_rate": 0.000262246740220662, + "loss": 1.328, + "theoretical_loss": 3.5272403137431976, + "tokens_seen": 1586823168 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622367101303912, + "loss": 1.4677, + "theoretical_loss": 3.527227710722081, + "tokens_seen": 1586888704 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026222668004012036, + "loss": 1.4192, + "theoretical_loss": 3.5272151083671677, + "tokens_seen": 1586954240 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026221664994984954, + "loss": 1.4524, + "theoretical_loss": 3.5272025066783943, + "tokens_seen": 1587019776 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002622066198595787, + "loss": 1.4664, + "theoretical_loss": 3.527189905655699, + "tokens_seen": 1587085312 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026219658976930796, + "loss": 1.3256, + "theoretical_loss": 3.527177305299018, + "tokens_seen": 1587150848 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002621865596790371, + "loss": 1.3615, + "theoretical_loss": 3.5271647056082895, + "tokens_seen": 1587216384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002621765295887663, + "loss": 1.347, + "theoretical_loss": 3.5271521065834506, + "tokens_seen": 1587281920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026216649949849545, + "loss": 1.3826, + "theoretical_loss": 3.5271395082244386, + "tokens_seen": 1587347456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002621564694082247, + "loss": 1.3458, + "theoretical_loss": 3.52712691053119, + "tokens_seen": 1587412992 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026214643931795386, + "loss": 1.3556, + "theoretical_loss": 3.5271143135036436, + "tokens_seen": 1587478528 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026213640922768304, + "loss": 1.3606, + "theoretical_loss": 3.5271017171417363, + "tokens_seen": 1587544064 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002621263791374122, + "loss": 1.3401, + "theoretical_loss": 3.5270891214454045, + "tokens_seen": 1587609600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026211634904714146, + "loss": 1.402, + "theoretical_loss": 3.5270765264145867, + "tokens_seen": 1587675136 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002621063189568706, + "loss": 1.2858, + "theoretical_loss": 3.527063932049219, + "tokens_seen": 1587740672 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620962888665998, + "loss": 1.3529, + "theoretical_loss": 3.5270513383492403, + "tokens_seen": 1587806208 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026208625877632895, + "loss": 1.2888, + "theoretical_loss": 3.5270387453145875, + "tokens_seen": 1587871744 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620762286860582, + "loss": 1.2665, + "theoretical_loss": 3.527026152945197, + "tokens_seen": 1587937280 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026206619859578737, + "loss": 1.2801, + "theoretical_loss": 3.5270135612410076, + "tokens_seen": 1588002816 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026205616850551655, + "loss": 1.3109, + "theoretical_loss": 3.5270009702019562, + "tokens_seen": 1588068352 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026204613841524573, + "loss": 1.4232, + "theoretical_loss": 3.5269883798279795, + "tokens_seen": 1588133888 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620361083249749, + "loss": 1.368, + "theoretical_loss": 3.5269757901190157, + "tokens_seen": 1588199424 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620260782347041, + "loss": 1.2452, + "theoretical_loss": 3.5269632010750023, + "tokens_seen": 1588264960 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620160481444333, + "loss": 1.2894, + "theoretical_loss": 3.526950612695876, + "tokens_seen": 1588330496 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002620060180541625, + "loss": 1.3511, + "theoretical_loss": 3.5269380249815754, + "tokens_seen": 1588396032 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002619959879638917, + "loss": 1.3361, + "theoretical_loss": 3.5269254379320367, + "tokens_seen": 1588461568 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026198595787362087, + "loss": 1.3231, + "theoretical_loss": 3.5269128515471984, + "tokens_seen": 1588527104 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026197592778335005, + "loss": 1.3828, + "theoretical_loss": 3.5269002658269972, + "tokens_seen": 1588592640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002619658976930793, + "loss": 1.2782, + "theoretical_loss": 3.526887680771371, + "tokens_seen": 1588658176 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002619558676028084, + "loss": 1.3325, + "theoretical_loss": 3.526875096380257, + "tokens_seen": 1588723712 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026194583751253765, + "loss": 1.3303, + "theoretical_loss": 3.5268625126535937, + "tokens_seen": 1588789248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026193580742226683, + "loss": 1.345, + "theoretical_loss": 3.526849929591317, + "tokens_seen": 1588854784 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261925777331996, + "loss": 1.3769, + "theoretical_loss": 3.526837347193366, + "tokens_seen": 1588920320 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002619157472417252, + "loss": 1.3474, + "theoretical_loss": 3.5268247654596774, + "tokens_seen": 1588985856 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026190571715145437, + "loss": 1.3708, + "theoretical_loss": 3.5268121843901885, + "tokens_seen": 1589051392 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026189568706118355, + "loss": 1.4345, + "theoretical_loss": 3.526799603984837, + "tokens_seen": 1589116928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618856569709128, + "loss": 1.3808, + "theoretical_loss": 3.526787024243561, + "tokens_seen": 1589182464 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 580493, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3095253705978394, + "objective/train/theoretical_loss": 3.5267744451662977, + "objective/train/tokens_used": 1609708000, + "theoretical_loss": 3.5267744451662977, + "tokens_seen": 1589248000 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618756268806419, + "loss": 1.2842, + "theoretical_loss": 3.5267744451662977, + "tokens_seen": 1589248000 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026186559679037115, + "loss": 1.2781, + "theoretical_loss": 3.5267618667529845, + "tokens_seen": 1589313536 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618555667001003, + "loss": 1.2921, + "theoretical_loss": 3.5267492890035594, + "tokens_seen": 1589379072 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618455366098295, + "loss": 1.3883, + "theoretical_loss": 3.52673671191796, + "tokens_seen": 1589444608 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618355065195587, + "loss": 1.308, + "theoretical_loss": 3.5267241354961225, + "tokens_seen": 1589510144 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618254764292879, + "loss": 1.5092, + "theoretical_loss": 3.5267115597379863, + "tokens_seen": 1589575680 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026181544633901706, + "loss": 1.3864, + "theoretical_loss": 3.5266989846434886, + "tokens_seen": 1589641216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002618054162487463, + "loss": 1.3666, + "theoretical_loss": 3.526686410212567, + "tokens_seen": 1589706752 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002617953861584754, + "loss": 1.2791, + "theoretical_loss": 3.526673836445158, + "tokens_seen": 1589772288 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026178535606820465, + "loss": 1.3424, + "theoretical_loss": 3.5266612633412007, + "tokens_seen": 1589837824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002617753259779338, + "loss": 1.2869, + "theoretical_loss": 3.526648690900632, + "tokens_seen": 1589903360 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261765295887663, + "loss": 1.2645, + "theoretical_loss": 3.52663611912339, + "tokens_seen": 1589968896 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002617552657973922, + "loss": 1.3391, + "theoretical_loss": 3.526623548009412, + "tokens_seen": 1590034432 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002617452357071214, + "loss": 1.2932, + "theoretical_loss": 3.5266109775586356, + "tokens_seen": 1590099968 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026173520561685056, + "loss": 1.3505, + "theoretical_loss": 3.5265984077709986, + "tokens_seen": 1590165504 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026172517552657974, + "loss": 1.4207, + "theoretical_loss": 3.526585838646439, + "tokens_seen": 1590231040 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002617151454363089, + "loss": 1.406, + "theoretical_loss": 3.526573270184894, + "tokens_seen": 1590296576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026170511534603816, + "loss": 1.3802, + "theoretical_loss": 3.526560702386302, + "tokens_seen": 1590362112 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002616950852557673, + "loss": 1.4915, + "theoretical_loss": 3.5265481352505996, + "tokens_seen": 1590427648 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002616850551654965, + "loss": 1.3153, + "theoretical_loss": 3.5265355687777253, + "tokens_seen": 1590493184 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026167502507522565, + "loss": 1.3421, + "theoretical_loss": 3.5265230029676173, + "tokens_seen": 1590558720 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002616649949849549, + "loss": 1.3945, + "theoretical_loss": 3.526510437820212, + "tokens_seen": 1590624256 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026165496489468406, + "loss": 1.2933, + "theoretical_loss": 3.5264978733354484, + "tokens_seen": 1590689792 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026164493480441324, + "loss": 1.3348, + "theoretical_loss": 3.526485309513264, + "tokens_seen": 1590755328 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002616349047141424, + "loss": 1.3815, + "theoretical_loss": 3.526472746353596, + "tokens_seen": 1590820864 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026162487462387166, + "loss": 1.4265, + "theoretical_loss": 3.5264601838563823, + "tokens_seen": 1590886400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002616148445336008, + "loss": 1.3928, + "theoretical_loss": 3.526447622021561, + "tokens_seen": 1590951936 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026160481444333, + "loss": 1.3263, + "theoretical_loss": 3.52643506084907, + "tokens_seen": 1591017472 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026159478435305915, + "loss": 1.4562, + "theoretical_loss": 3.526422500338846, + "tokens_seen": 1591083008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002615847542627884, + "loss": 1.2983, + "theoretical_loss": 3.5264099404908285, + "tokens_seen": 1591148544 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026157472417251757, + "loss": 1.3929, + "theoretical_loss": 3.5263973813049545, + "tokens_seen": 1591214080 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026156469408224675, + "loss": 1.3328, + "theoretical_loss": 3.5263848227811616, + "tokens_seen": 1591279616 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026155466399197593, + "loss": 1.3054, + "theoretical_loss": 3.526372264919388, + "tokens_seen": 1591345152 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002615446339017051, + "loss": 1.2394, + "theoretical_loss": 3.526359707719571, + "tokens_seen": 1591410688 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002615346038114343, + "loss": 1.3623, + "theoretical_loss": 3.5263471511816493, + "tokens_seen": 1591476224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002615245737211635, + "loss": 1.2182, + "theoretical_loss": 3.52633459530556, + "tokens_seen": 1591541760 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026151454363089265, + "loss": 1.4388, + "theoretical_loss": 3.5263220400912414, + "tokens_seen": 1591607296 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002615045135406219, + "loss": 1.2428, + "theoretical_loss": 3.526309485538631, + "tokens_seen": 1591672832 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261494483450351, + "loss": 1.3775, + "theoretical_loss": 3.5262969316476673, + "tokens_seen": 1591738368 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026148445336008025, + "loss": 1.3443, + "theoretical_loss": 3.5262843784182873, + "tokens_seen": 1591803904 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026147442326980943, + "loss": 1.2577, + "theoretical_loss": 3.5262718258504298, + "tokens_seen": 1591869440 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002614643931795386, + "loss": 1.4536, + "theoretical_loss": 3.5262592739440324, + "tokens_seen": 1591934976 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002614543630892678, + "loss": 1.3581, + "theoretical_loss": 3.5262467226990326, + "tokens_seen": 1592000512 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026144433299899703, + "loss": 1.2473, + "theoretical_loss": 3.5262341721153687, + "tokens_seen": 1592066048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026143430290872616, + "loss": 1.3333, + "theoretical_loss": 3.5262216221929785, + "tokens_seen": 1592131584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002614242728184554, + "loss": 1.3403, + "theoretical_loss": 3.5262090729318, + "tokens_seen": 1592197120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002614142427281845, + "loss": 1.2874, + "theoretical_loss": 3.5261965243317714, + "tokens_seen": 1592262656 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026140421263791375, + "loss": 1.3063, + "theoretical_loss": 3.52618397639283, + "tokens_seen": 1592328192 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026139418254764293, + "loss": 1.3089, + "theoretical_loss": 3.526171429114915, + "tokens_seen": 1592393728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613841524573721, + "loss": 1.2878, + "theoretical_loss": 3.526158882497963, + "tokens_seen": 1592459264 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 581545, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.483906626701355, + "objective/train/theoretical_loss": 3.5261463365419123, + "objective/train/tokens_used": 1612984800, + "theoretical_loss": 3.5261463365419123, + "tokens_seen": 1592524800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613741223671013, + "loss": 1.4246, + "theoretical_loss": 3.5261463365419123, + "tokens_seen": 1592524800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613640922768305, + "loss": 1.3284, + "theoretical_loss": 3.5261337912467017, + "tokens_seen": 1592590336 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026135406218655966, + "loss": 1.2487, + "theoretical_loss": 3.5261212466122682, + "tokens_seen": 1592655872 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613440320962889, + "loss": 1.3719, + "theoretical_loss": 3.5261087026385507, + "tokens_seen": 1592721408 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261334002006018, + "loss": 1.4546, + "theoretical_loss": 3.526096159325486, + "tokens_seen": 1592786944 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026132397191574726, + "loss": 1.3715, + "theoretical_loss": 3.526083616673014, + "tokens_seen": 1592852480 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613139418254764, + "loss": 1.383, + "theoretical_loss": 3.526071074681071, + "tokens_seen": 1592918016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002613039117352056, + "loss": 1.4145, + "theoretical_loss": 3.5260585333495955, + "tokens_seen": 1592983552 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002612938816449348, + "loss": 1.2886, + "theoretical_loss": 3.526045992678526, + "tokens_seen": 1593049088 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261283851554664, + "loss": 1.4088, + "theoretical_loss": 3.5260334526678, + "tokens_seen": 1593114624 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026127382146439316, + "loss": 1.3041, + "theoretical_loss": 3.526020913317356, + "tokens_seen": 1593180160 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002612637913741224, + "loss": 1.2787, + "theoretical_loss": 3.526008374627132, + "tokens_seen": 1593245696 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002612537612838516, + "loss": 1.3429, + "theoretical_loss": 3.525995836597066, + "tokens_seen": 1593311232 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026124373119358076, + "loss": 1.3454, + "theoretical_loss": 3.5259832992270956, + "tokens_seen": 1593376768 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026123370110330994, + "loss": 1.3683, + "theoretical_loss": 3.52597076251716, + "tokens_seen": 1593442304 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002612236710130391, + "loss": 1.3177, + "theoretical_loss": 3.5259582264671963, + "tokens_seen": 1593507840 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026121364092276836, + "loss": 1.4215, + "theoretical_loss": 3.525945691077143, + "tokens_seen": 1593573376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002612036108324975, + "loss": 1.4266, + "theoretical_loss": 3.5259331563469387, + "tokens_seen": 1593638912 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002611935807422267, + "loss": 1.4018, + "theoretical_loss": 3.525920622276521, + "tokens_seen": 1593704448 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026118355065195585, + "loss": 1.3997, + "theoretical_loss": 3.525908088865828, + "tokens_seen": 1593769984 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002611735205616851, + "loss": 1.3979, + "theoretical_loss": 3.5258955561147975, + "tokens_seen": 1593835520 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026116349047141426, + "loss": 1.3742, + "theoretical_loss": 3.5258830240233685, + "tokens_seen": 1593901056 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026115346038114344, + "loss": 1.264, + "theoretical_loss": 3.525870492591479, + "tokens_seen": 1593966592 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002611434302908726, + "loss": 1.3629, + "theoretical_loss": 3.5258579618190664, + "tokens_seen": 1594032128 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026113340020060186, + "loss": 1.4013, + "theoretical_loss": 3.5258454317060695, + "tokens_seen": 1594097664 + }, + { + "epoch": 0.48, + "learning_rate": 0.000261123370110331, + "loss": 1.3876, + "theoretical_loss": 3.525832902252427, + "tokens_seen": 1594163200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002611133400200602, + "loss": 1.2791, + "theoretical_loss": 3.5258203734580764, + "tokens_seen": 1594228736 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026110330992978935, + "loss": 1.3543, + "theoretical_loss": 3.5258078453229555, + "tokens_seen": 1594294272 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610932798395186, + "loss": 1.2337, + "theoretical_loss": 3.5257953178470034, + "tokens_seen": 1594359808 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026108324974924777, + "loss": 1.3209, + "theoretical_loss": 3.5257827910301582, + "tokens_seen": 1594425344 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026107321965897695, + "loss": 1.3179, + "theoretical_loss": 3.5257702648723575, + "tokens_seen": 1594490880 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026106318956870613, + "loss": 1.3419, + "theoretical_loss": 3.5257577393735406, + "tokens_seen": 1594556416 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610531594784353, + "loss": 1.3115, + "theoretical_loss": 3.5257452145336443, + "tokens_seen": 1594621952 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610431293881645, + "loss": 1.3042, + "theoretical_loss": 3.525732690352608, + "tokens_seen": 1594687488 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610330992978937, + "loss": 1.4026, + "theoretical_loss": 3.5257201668303697, + "tokens_seen": 1594753024 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026102306920762285, + "loss": 1.3815, + "theoretical_loss": 3.525707643966868, + "tokens_seen": 1594818560 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610130391173521, + "loss": 1.3659, + "theoretical_loss": 3.5256951217620403, + "tokens_seen": 1594884096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002610030090270812, + "loss": 1.3302, + "theoretical_loss": 3.5256826002158252, + "tokens_seen": 1594949632 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026099297893681045, + "loss": 1.2759, + "theoretical_loss": 3.5256700793281617, + "tokens_seen": 1595015168 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026098294884653963, + "loss": 1.1478, + "theoretical_loss": 3.5256575590989874, + "tokens_seen": 1595080704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002609729187562688, + "loss": 1.3427, + "theoretical_loss": 3.5256450395282406, + "tokens_seen": 1595146240 + }, + { + "epoch": 0.48, + "learning_rate": 0.000260962888665998, + "loss": 1.3185, + "theoretical_loss": 3.52563252061586, + "tokens_seen": 1595211776 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026095285857572723, + "loss": 1.235, + "theoretical_loss": 3.5256200023617836, + "tokens_seen": 1595277312 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026094282848545636, + "loss": 1.2614, + "theoretical_loss": 3.52560748476595, + "tokens_seen": 1595342848 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002609327983951856, + "loss": 1.3622, + "theoretical_loss": 3.5255949678282974, + "tokens_seen": 1595408384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002609227683049147, + "loss": 1.3952, + "theoretical_loss": 3.5255824515487646, + "tokens_seen": 1595473920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026091273821464395, + "loss": 1.387, + "theoretical_loss": 3.5255699359272894, + "tokens_seen": 1595539456 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026090270812437313, + "loss": 1.3131, + "theoretical_loss": 3.52555742096381, + "tokens_seen": 1595604992 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608926780341023, + "loss": 1.234, + "theoretical_loss": 3.5255449066582654, + "tokens_seen": 1595670528 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608826479438315, + "loss": 1.2514, + "theoretical_loss": 3.525532393010594, + "tokens_seen": 1595736064 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 582715, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3253862857818604, + "objective/train/theoretical_loss": 3.5255198800207332, + "objective/train/tokens_used": 1616261600, + "theoretical_loss": 3.5255198800207332, + "tokens_seen": 1595801600 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608726178535607, + "loss": 1.2615, + "theoretical_loss": 3.5255198800207332, + "tokens_seen": 1595801600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026086258776328986, + "loss": 1.3132, + "theoretical_loss": 3.525507367688623, + "tokens_seen": 1595867136 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608525576730191, + "loss": 1.2592, + "theoretical_loss": 3.5254948560142, + "tokens_seen": 1595932672 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608425275827482, + "loss": 1.2995, + "theoretical_loss": 3.525482344997404, + "tokens_seen": 1595998208 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026083249749247746, + "loss": 1.3221, + "theoretical_loss": 3.5254698346381734, + "tokens_seen": 1596063744 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608224674022066, + "loss": 1.4414, + "theoretical_loss": 3.525457324936446, + "tokens_seen": 1596129280 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002608124373119358, + "loss": 1.3686, + "theoretical_loss": 3.5254448158921603, + "tokens_seen": 1596194816 + }, + { + "epoch": 0.48, + "learning_rate": 0.000260802407221665, + "loss": 1.3151, + "theoretical_loss": 3.5254323075052554, + "tokens_seen": 1596260352 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607923771313942, + "loss": 1.345, + "theoretical_loss": 3.525419799775669, + "tokens_seen": 1596325888 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026078234704112336, + "loss": 1.2506, + "theoretical_loss": 3.52540729270334, + "tokens_seen": 1596391424 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607723169508526, + "loss": 1.2471, + "theoretical_loss": 3.5253947862882065, + "tokens_seen": 1596456960 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607622868605817, + "loss": 1.3352, + "theoretical_loss": 3.5253822805302075, + "tokens_seen": 1596522496 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026075225677031096, + "loss": 1.4148, + "theoretical_loss": 3.525369775429281, + "tokens_seen": 1596588032 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607422266800401, + "loss": 1.3013, + "theoretical_loss": 3.525357270985366, + "tokens_seen": 1596653568 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607321965897693, + "loss": 1.3642, + "theoretical_loss": 3.5253447671984013, + "tokens_seen": 1596719104 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607221664994985, + "loss": 1.2951, + "theoretical_loss": 3.5253322640683242, + "tokens_seen": 1596784640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002607121364092277, + "loss": 1.3795, + "theoretical_loss": 3.5253197615950747, + "tokens_seen": 1596850176 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026070210631895687, + "loss": 1.2874, + "theoretical_loss": 3.5253072597785895, + "tokens_seen": 1596915712 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026069207622868605, + "loss": 1.2922, + "theoretical_loss": 3.5252947586188093, + "tokens_seen": 1596981248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026068204613841523, + "loss": 1.3821, + "theoretical_loss": 3.525282258115671, + "tokens_seen": 1597046784 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026067201604814446, + "loss": 1.3356, + "theoretical_loss": 3.525269758269114, + "tokens_seen": 1597112320 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002606619859578736, + "loss": 1.3304, + "theoretical_loss": 3.5252572590790763, + "tokens_seen": 1597177856 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002606519558676028, + "loss": 1.2868, + "theoretical_loss": 3.5252447605454966, + "tokens_seen": 1597243392 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026064192577733195, + "loss": 1.2624, + "theoretical_loss": 3.5252322626683146, + "tokens_seen": 1597308928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002606318956870612, + "loss": 1.2614, + "theoretical_loss": 3.525219765447467, + "tokens_seen": 1597374464 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026062186559679037, + "loss": 1.3895, + "theoretical_loss": 3.525207268882894, + "tokens_seen": 1597440000 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026061183550651955, + "loss": 1.4257, + "theoretical_loss": 3.5251947729745337, + "tokens_seen": 1597505536 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026060180541624873, + "loss": 1.3033, + "theoretical_loss": 3.5251822777223243, + "tokens_seen": 1597571072 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026059177532597797, + "loss": 1.2958, + "theoretical_loss": 3.525169783126205, + "tokens_seen": 1597636608 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002605817452357071, + "loss": 1.264, + "theoretical_loss": 3.5251572891861143, + "tokens_seen": 1597702144 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026057171514543633, + "loss": 1.4133, + "theoretical_loss": 3.5251447959019906, + "tokens_seen": 1597767680 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026056168505516546, + "loss": 1.3117, + "theoretical_loss": 3.525132303273773, + "tokens_seen": 1597833216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002605516549648947, + "loss": 1.3212, + "theoretical_loss": 3.5251198113013995, + "tokens_seen": 1597898752 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026054162487462387, + "loss": 1.3411, + "theoretical_loss": 3.5251073199848095, + "tokens_seen": 1597964288 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026053159478435305, + "loss": 1.3104, + "theoretical_loss": 3.5250948293239412, + "tokens_seen": 1598029824 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026052156469408223, + "loss": 1.3074, + "theoretical_loss": 3.5250823393187334, + "tokens_seen": 1598095360 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002605115346038114, + "loss": 1.2655, + "theoretical_loss": 3.5250698499691246, + "tokens_seen": 1598160896 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026050150451354065, + "loss": 1.2907, + "theoretical_loss": 3.5250573612750538, + "tokens_seen": 1598226432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026049147442326983, + "loss": 1.269, + "theoretical_loss": 3.52504487323646, + "tokens_seen": 1598291968 + }, + { + "epoch": 0.48, + "learning_rate": 0.000260481444332999, + "loss": 1.3896, + "theoretical_loss": 3.5250323858532813, + "tokens_seen": 1598357504 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002604714142427282, + "loss": 1.4237, + "theoretical_loss": 3.5250198991254567, + "tokens_seen": 1598423040 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026046138415245743, + "loss": 1.3232, + "theoretical_loss": 3.5250074130529248, + "tokens_seen": 1598488576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026045135406218656, + "loss": 1.36, + "theoretical_loss": 3.5249949276356247, + "tokens_seen": 1598554112 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002604413239719158, + "loss": 1.3814, + "theoretical_loss": 3.524982442873495, + "tokens_seen": 1598619648 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002604312938816449, + "loss": 1.3518, + "theoretical_loss": 3.5249699587664742, + "tokens_seen": 1598685184 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026042126379137415, + "loss": 1.4027, + "theoretical_loss": 3.524957475314501, + "tokens_seen": 1598750720 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026041123370110333, + "loss": 1.3413, + "theoretical_loss": 3.524944992517515, + "tokens_seen": 1598816256 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002604012036108325, + "loss": 1.4485, + "theoretical_loss": 3.5249325103754545, + "tokens_seen": 1598881792 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603911735205617, + "loss": 1.2431, + "theoretical_loss": 3.524920028888258, + "tokens_seen": 1598947328 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603811434302909, + "loss": 1.3427, + "theoretical_loss": 3.5249075480558645, + "tokens_seen": 1599012864 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 583998, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4529955387115479, + "objective/train/theoretical_loss": 3.5248950678782127, + "objective/train/tokens_used": 1619538400, + "theoretical_loss": 3.5248950678782127, + "tokens_seen": 1599078400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026037111334002006, + "loss": 1.3584, + "theoretical_loss": 3.5248950678782127, + "tokens_seen": 1599078400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603610832497493, + "loss": 1.2791, + "theoretical_loss": 3.5248825883552417, + "tokens_seen": 1599143936 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603510531594784, + "loss": 1.2932, + "theoretical_loss": 3.5248701094868906, + "tokens_seen": 1599209472 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026034102306920766, + "loss": 1.4572, + "theoretical_loss": 3.5248576312730973, + "tokens_seen": 1599275008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603309929789368, + "loss": 1.3106, + "theoretical_loss": 3.5248451537138012, + "tokens_seen": 1599340544 + }, + { + "epoch": 0.48, + "learning_rate": 0.000260320962888666, + "loss": 1.3904, + "theoretical_loss": 3.5248326768089413, + "tokens_seen": 1599406080 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603109327983952, + "loss": 1.3978, + "theoretical_loss": 3.524820200558456, + "tokens_seen": 1599471616 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002603009027081244, + "loss": 1.3392, + "theoretical_loss": 3.524807724962285, + "tokens_seen": 1599537152 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026029087261785356, + "loss": 1.3984, + "theoretical_loss": 3.5247952500203663, + "tokens_seen": 1599602688 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602808425275828, + "loss": 1.3524, + "theoretical_loss": 3.5247827757326395, + "tokens_seen": 1599668224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602708124373119, + "loss": 1.3725, + "theoretical_loss": 3.524770302099043, + "tokens_seen": 1599733760 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026026078234704116, + "loss": 1.3547, + "theoretical_loss": 3.5247578291195154, + "tokens_seen": 1599799296 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602507522567703, + "loss": 1.439, + "theoretical_loss": 3.5247453567939964, + "tokens_seen": 1599864832 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602407221664995, + "loss": 1.4486, + "theoretical_loss": 3.5247328851224244, + "tokens_seen": 1599930368 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602306920762287, + "loss": 1.3388, + "theoretical_loss": 3.5247204141047392, + "tokens_seen": 1599995904 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002602206619859579, + "loss": 1.3234, + "theoretical_loss": 3.524707943740878, + "tokens_seen": 1600061440 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026021063189568707, + "loss": 1.3363, + "theoretical_loss": 3.524695474030781, + "tokens_seen": 1600126976 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026020060180541625, + "loss": 1.4298, + "theoretical_loss": 3.5246830049743876, + "tokens_seen": 1600192512 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026019057171514543, + "loss": 1.3955, + "theoretical_loss": 3.5246705365716355, + "tokens_seen": 1600258048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00026018054162487466, + "loss": 1.3287, + "theoretical_loss": 3.5246580688224647, + "tokens_seen": 1600323584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0002601705115346038, + "loss": 1.2785, + "theoretical_loss": 3.524645601726813, + "tokens_seen": 1600389120 + }, + { + "epoch": 0.48, + "learning_rate": 0.000260160481444333, + "loss": 1.4766, + "theoretical_loss": 3.5246331352846205, + "tokens_seen": 1600454656 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026015045135406215, + "loss": 1.2976, + "theoretical_loss": 3.524620669495826, + "tokens_seen": 1600520192 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002601404212637914, + "loss": 1.468, + "theoretical_loss": 3.524608204360368, + "tokens_seen": 1600585728 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026013039117352057, + "loss": 1.4624, + "theoretical_loss": 3.524595739878186, + "tokens_seen": 1600651264 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026012036108324975, + "loss": 1.2864, + "theoretical_loss": 3.524583276049219, + "tokens_seen": 1600716800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026011033099297893, + "loss": 1.3075, + "theoretical_loss": 3.5245708128734057, + "tokens_seen": 1600782336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026010030090270817, + "loss": 1.3886, + "theoretical_loss": 3.524558350350685, + "tokens_seen": 1600847872 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002600902708124373, + "loss": 1.4295, + "theoretical_loss": 3.524545888480996, + "tokens_seen": 1600913408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026008024072216653, + "loss": 1.3441, + "theoretical_loss": 3.524533427264279, + "tokens_seen": 1600978944 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026007021063189566, + "loss": 1.3635, + "theoretical_loss": 3.5245209667004715, + "tokens_seen": 1601044480 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002600601805416249, + "loss": 1.3804, + "theoretical_loss": 3.524508506789513, + "tokens_seen": 1601110016 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026005015045135407, + "loss": 1.3045, + "theoretical_loss": 3.5244960475313425, + "tokens_seen": 1601175552 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026004012036108325, + "loss": 1.3246, + "theoretical_loss": 3.5244835889259, + "tokens_seen": 1601241088 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026003009027081243, + "loss": 1.4521, + "theoretical_loss": 3.524471130973123, + "tokens_seen": 1601306624 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002600200601805416, + "loss": 1.228, + "theoretical_loss": 3.524458673672952, + "tokens_seen": 1601372160 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002600100300902708, + "loss": 1.4079, + "theoretical_loss": 3.524446217025326, + "tokens_seen": 1601437696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00026000000000000003, + "loss": 1.2287, + "theoretical_loss": 3.524433761030183, + "tokens_seen": 1601503232 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025998996990972916, + "loss": 1.2874, + "theoretical_loss": 3.524421305687463, + "tokens_seen": 1601568768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002599799398194584, + "loss": 1.4317, + "theoretical_loss": 3.524408850997105, + "tokens_seen": 1601634304 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002599699097291876, + "loss": 1.4434, + "theoretical_loss": 3.5243963969590477, + "tokens_seen": 1601699840 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025995987963891676, + "loss": 1.3616, + "theoretical_loss": 3.524383943573231, + "tokens_seen": 1601765376 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025994984954864594, + "loss": 1.3383, + "theoretical_loss": 3.5243714908395933, + "tokens_seen": 1601830912 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002599398194583751, + "loss": 1.3894, + "theoretical_loss": 3.5243590387580745, + "tokens_seen": 1601896448 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002599297893681043, + "loss": 1.4262, + "theoretical_loss": 3.5243465873286137, + "tokens_seen": 1601961984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025991975927783353, + "loss": 1.3924, + "theoretical_loss": 3.5243341365511496, + "tokens_seen": 1602027520 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025990972918756266, + "loss": 1.3056, + "theoretical_loss": 3.5243216864256213, + "tokens_seen": 1602093056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002598996990972919, + "loss": 1.5095, + "theoretical_loss": 3.524309236951969, + "tokens_seen": 1602158592 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259889669007021, + "loss": 1.328, + "theoretical_loss": 3.5242967881301306, + "tokens_seen": 1602224128 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025987963891675026, + "loss": 1.2974, + "theoretical_loss": 3.5242843399600456, + "tokens_seen": 1602289664 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 585126, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3694276809692383, + "objective/train/theoretical_loss": 3.524271892441654, + "objective/train/tokens_used": 1622815200, + "theoretical_loss": 3.524271892441654, + "tokens_seen": 1602355200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025986960882647944, + "loss": 1.3617, + "theoretical_loss": 3.524271892441654, + "tokens_seen": 1602355200 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002598595787362086, + "loss": 1.3653, + "theoretical_loss": 3.5242594455748946, + "tokens_seen": 1602420736 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002598495486459378, + "loss": 1.2698, + "theoretical_loss": 3.5242469993597068, + "tokens_seen": 1602486272 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259839518555667, + "loss": 1.3561, + "theoretical_loss": 3.5242345537960293, + "tokens_seen": 1602551808 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025982948846539616, + "loss": 1.3559, + "theoretical_loss": 3.5242221088838015, + "tokens_seen": 1602617344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002598194583751254, + "loss": 1.3225, + "theoretical_loss": 3.5242096646229633, + "tokens_seen": 1602682880 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002598094282848545, + "loss": 1.3765, + "theoretical_loss": 3.5241972210134533, + "tokens_seen": 1602748416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025979939819458376, + "loss": 1.3346, + "theoretical_loss": 3.5241847780552114, + "tokens_seen": 1602813952 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025978936810431294, + "loss": 1.286, + "theoretical_loss": 3.524172335748176, + "tokens_seen": 1602879488 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597793380140421, + "loss": 1.1265, + "theoretical_loss": 3.5241598940922874, + "tokens_seen": 1602945024 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597693079237713, + "loss": 1.3871, + "theoretical_loss": 3.524147453087484, + "tokens_seen": 1603010560 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597592778335005, + "loss": 1.4586, + "theoretical_loss": 3.524135012733706, + "tokens_seen": 1603076096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597492477432297, + "loss": 1.4058, + "theoretical_loss": 3.5241225730308914, + "tokens_seen": 1603141632 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597392176529589, + "loss": 1.2779, + "theoretical_loss": 3.524110133978981, + "tokens_seen": 1603207168 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002597291875626881, + "loss": 1.2498, + "theoretical_loss": 3.5240976955779137, + "tokens_seen": 1603272704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025971915747241727, + "loss": 1.3048, + "theoretical_loss": 3.524085257827628, + "tokens_seen": 1603338240 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025970912738214645, + "loss": 1.3975, + "theoretical_loss": 3.5240728207280645, + "tokens_seen": 1603403776 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025969909729187563, + "loss": 1.336, + "theoretical_loss": 3.5240603842791614, + "tokens_seen": 1603469312 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025968906720160486, + "loss": 1.3767, + "theoretical_loss": 3.5240479484808587, + "tokens_seen": 1603534848 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259679037111334, + "loss": 1.3461, + "theoretical_loss": 3.524035513333096, + "tokens_seen": 1603600384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002596690070210632, + "loss": 1.3336, + "theoretical_loss": 3.5240230788358122, + "tokens_seen": 1603665920 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025965897693079235, + "loss": 1.4271, + "theoretical_loss": 3.5240106449889472, + "tokens_seen": 1603731456 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002596489468405216, + "loss": 1.3648, + "theoretical_loss": 3.5239982117924398, + "tokens_seen": 1603796992 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025963891675025077, + "loss": 1.2529, + "theoretical_loss": 3.52398577924623, + "tokens_seen": 1603862528 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025962888665997995, + "loss": 1.4421, + "theoretical_loss": 3.5239733473502564, + "tokens_seen": 1603928064 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025961885656970913, + "loss": 1.358, + "theoretical_loss": 3.5239609161044596, + "tokens_seen": 1603993600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025960882647943837, + "loss": 1.3881, + "theoretical_loss": 3.5239484855087775, + "tokens_seen": 1604059136 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002595987963891675, + "loss": 1.3404, + "theoretical_loss": 3.523936055563151, + "tokens_seen": 1604124672 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025958876629889673, + "loss": 1.4839, + "theoretical_loss": 3.5239236262675186, + "tokens_seen": 1604190208 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025957873620862586, + "loss": 1.3831, + "theoretical_loss": 3.5239111976218203, + "tokens_seen": 1604255744 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002595687061183551, + "loss": 1.335, + "theoretical_loss": 3.523898769625996, + "tokens_seen": 1604321280 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025955867602808427, + "loss": 1.2795, + "theoretical_loss": 3.523886342279984, + "tokens_seen": 1604386816 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025954864593781345, + "loss": 1.4194, + "theoretical_loss": 3.523873915583724, + "tokens_seen": 1604452352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025953861584754263, + "loss": 1.2743, + "theoretical_loss": 3.5238614895371563, + "tokens_seen": 1604517888 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002595285857572718, + "loss": 1.4372, + "theoretical_loss": 3.5238490641402196, + "tokens_seen": 1604583424 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259518555667001, + "loss": 1.2663, + "theoretical_loss": 3.523836639392854, + "tokens_seen": 1604648960 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025950852557673023, + "loss": 1.2995, + "theoretical_loss": 3.5238242152949986, + "tokens_seen": 1604714496 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025949849548645936, + "loss": 1.3508, + "theoretical_loss": 3.5238117918465934, + "tokens_seen": 1604780032 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002594884653961886, + "loss": 1.2521, + "theoretical_loss": 3.5237993690475773, + "tokens_seen": 1604845568 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002594784353059178, + "loss": 1.3876, + "theoretical_loss": 3.52378694689789, + "tokens_seen": 1604911104 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025946840521564696, + "loss": 1.4745, + "theoretical_loss": 3.5237745253974717, + "tokens_seen": 1604976640 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025945837512537614, + "loss": 1.3261, + "theoretical_loss": 3.523762104546261, + "tokens_seen": 1605042176 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002594483450351053, + "loss": 1.4268, + "theoretical_loss": 3.523749684344198, + "tokens_seen": 1605107712 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002594383149448345, + "loss": 1.4307, + "theoretical_loss": 3.5237372647912224, + "tokens_seen": 1605173248 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025942828485456373, + "loss": 1.34, + "theoretical_loss": 3.5237248458872736, + "tokens_seen": 1605238784 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025941825476429286, + "loss": 1.3304, + "theoretical_loss": 3.5237124276322906, + "tokens_seen": 1605304320 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002594082246740221, + "loss": 1.2568, + "theoretical_loss": 3.523700010026214, + "tokens_seen": 1605369856 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002593981945837512, + "loss": 1.3318, + "theoretical_loss": 3.5236875930689826, + "tokens_seen": 1605435392 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025938816449348046, + "loss": 1.378, + "theoretical_loss": 3.5236751767605368, + "tokens_seen": 1605500928 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025937813440320964, + "loss": 1.2472, + "theoretical_loss": 3.5236627611008156, + "tokens_seen": 1605566464 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 586391, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3439435958862305, + "objective/train/theoretical_loss": 3.5236503460897586, + "objective/train/tokens_used": 1626092000, + "theoretical_loss": 3.5236503460897586, + "tokens_seen": 1605632000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002593681043129388, + "loss": 1.3455, + "theoretical_loss": 3.5236503460897586, + "tokens_seen": 1605632000 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259358074222668, + "loss": 1.3409, + "theoretical_loss": 3.523637931727306, + "tokens_seen": 1605697536 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002593480441323972, + "loss": 1.3277, + "theoretical_loss": 3.5236255180133966, + "tokens_seen": 1605763072 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025933801404212636, + "loss": 1.2472, + "theoretical_loss": 3.5236131049479713, + "tokens_seen": 1605828608 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002593279839518556, + "loss": 1.3626, + "theoretical_loss": 3.5236006925309686, + "tokens_seen": 1605894144 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002593179538615847, + "loss": 1.3737, + "theoretical_loss": 3.523588280762328, + "tokens_seen": 1605959680 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025930792377131396, + "loss": 1.393, + "theoretical_loss": 3.5235758696419905, + "tokens_seen": 1606025216 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025929789368104314, + "loss": 1.3958, + "theoretical_loss": 3.5235634591698948, + "tokens_seen": 1606090752 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002592878635907723, + "loss": 1.2642, + "theoretical_loss": 3.523551049345981, + "tokens_seen": 1606156288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002592778335005015, + "loss": 1.3016, + "theoretical_loss": 3.5235386401701883, + "tokens_seen": 1606221824 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002592678034102307, + "loss": 1.3707, + "theoretical_loss": 3.5235262316424567, + "tokens_seen": 1606287360 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025925777331995987, + "loss": 1.3318, + "theoretical_loss": 3.523513823762726, + "tokens_seen": 1606352896 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002592477432296891, + "loss": 1.2275, + "theoretical_loss": 3.5235014165309364, + "tokens_seen": 1606418432 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025923771313941823, + "loss": 1.3342, + "theoretical_loss": 3.5234890099470264, + "tokens_seen": 1606483968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025922768304914747, + "loss": 1.3497, + "theoretical_loss": 3.523476604010937, + "tokens_seen": 1606549504 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002592176529588766, + "loss": 1.3614, + "theoretical_loss": 3.523464198722607, + "tokens_seen": 1606615040 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025920762286860583, + "loss": 1.3419, + "theoretical_loss": 3.5234517940819767, + "tokens_seen": 1606680576 + }, + { + "epoch": 0.49, + "learning_rate": 0.000259197592778335, + "loss": 1.3707, + "theoretical_loss": 3.5234393900889858, + "tokens_seen": 1606746112 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002591875626880642, + "loss": 1.2948, + "theoretical_loss": 3.5234269867435737, + "tokens_seen": 1606811648 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025917753259779337, + "loss": 1.1428, + "theoretical_loss": 3.523414584045681, + "tokens_seen": 1606877184 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025916750250752255, + "loss": 1.392, + "theoretical_loss": 3.5234021819952464, + "tokens_seen": 1606942720 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025915747241725173, + "loss": 1.2899, + "theoretical_loss": 3.5233897805922103, + "tokens_seen": 1607008256 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025914744232698097, + "loss": 1.4526, + "theoretical_loss": 3.523377379836513, + "tokens_seen": 1607073792 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002591374122367101, + "loss": 1.3793, + "theoretical_loss": 3.5233649797280933, + "tokens_seen": 1607139328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025912738214643933, + "loss": 1.2974, + "theoretical_loss": 3.5233525802668915, + "tokens_seen": 1607204864 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002591173520561685, + "loss": 1.3334, + "theoretical_loss": 3.5233401814528476, + "tokens_seen": 1607270400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002591073219658977, + "loss": 1.3141, + "theoretical_loss": 3.523327783285901, + "tokens_seen": 1607335936 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002590972918756269, + "loss": 1.3539, + "theoretical_loss": 3.5233153857659922, + "tokens_seen": 1607401472 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025908726178535606, + "loss": 1.2768, + "theoretical_loss": 3.523302988893061, + "tokens_seen": 1607467008 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025907723169508524, + "loss": 1.5205, + "theoretical_loss": 3.5232905926670464, + "tokens_seen": 1607532544 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025906720160481447, + "loss": 1.2529, + "theoretical_loss": 3.5232781970878886, + "tokens_seen": 1607598080 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002590571715145436, + "loss": 1.2998, + "theoretical_loss": 3.523265802155528, + "tokens_seen": 1607663616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025904714142427283, + "loss": 1.3049, + "theoretical_loss": 3.5232534078699036, + "tokens_seen": 1607729152 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025903711133400196, + "loss": 1.4274, + "theoretical_loss": 3.5232410142309565, + "tokens_seen": 1607794688 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002590270812437312, + "loss": 1.4096, + "theoretical_loss": 3.523228621238626, + "tokens_seen": 1607860224 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002590170511534604, + "loss": 1.3933, + "theoretical_loss": 3.523216228892852, + "tokens_seen": 1607925760 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025900702106318956, + "loss": 1.2972, + "theoretical_loss": 3.5232038371935737, + "tokens_seen": 1607991296 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002589969909729188, + "loss": 1.2828, + "theoretical_loss": 3.523191446140732, + "tokens_seen": 1608056832 + }, + { + "epoch": 0.49, + "learning_rate": 0.000258986960882648, + "loss": 1.2789, + "theoretical_loss": 3.523179055734267, + "tokens_seen": 1608122368 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025897693079237716, + "loss": 1.4154, + "theoretical_loss": 3.523166665974118, + "tokens_seen": 1608187904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025896690070210634, + "loss": 1.2713, + "theoretical_loss": 3.523154276860225, + "tokens_seen": 1608253440 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002589568706118355, + "loss": 1.3713, + "theoretical_loss": 3.5231418883925283, + "tokens_seen": 1608318976 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002589468405215647, + "loss": 1.367, + "theoretical_loss": 3.523129500570967, + "tokens_seen": 1608384512 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025893681043129393, + "loss": 1.2799, + "theoretical_loss": 3.5231171133954824, + "tokens_seen": 1608450048 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025892678034102306, + "loss": 1.4277, + "theoretical_loss": 3.5231047268660136, + "tokens_seen": 1608515584 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002589167502507523, + "loss": 1.4027, + "theoretical_loss": 3.523092340982501, + "tokens_seen": 1608581120 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002589067201604814, + "loss": 1.4294, + "theoretical_loss": 3.523079955744884, + "tokens_seen": 1608646656 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025889669007021066, + "loss": 1.3527, + "theoretical_loss": 3.5230675711531036, + "tokens_seen": 1608712192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025888665997993984, + "loss": 1.2519, + "theoretical_loss": 3.523055187207099, + "tokens_seen": 1608777728 + }, + { + "epoch": 0.49, + "learning_rate": 0.000258876629889669, + "loss": 1.4411, + "theoretical_loss": 3.5230428039068102, + "tokens_seen": 1608843264 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 587617, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5783393383026123, + "objective/train/theoretical_loss": 3.5230304212521775, + "objective/train/tokens_used": 1629368800, + "theoretical_loss": 3.5230304212521775, + "tokens_seen": 1608908800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002588665997993982, + "loss": 1.4913, + "theoretical_loss": 3.5230304212521775, + "tokens_seen": 1608908800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002588565697091274, + "loss": 1.3967, + "theoretical_loss": 3.523018039243141, + "tokens_seen": 1608974336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025884653961885657, + "loss": 1.4864, + "theoretical_loss": 3.523005657879641, + "tokens_seen": 1609039872 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002588365095285858, + "loss": 1.4032, + "theoretical_loss": 3.522993277161617, + "tokens_seen": 1609105408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025882647943831493, + "loss": 1.3334, + "theoretical_loss": 3.522980897089009, + "tokens_seen": 1609170944 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025881644934804416, + "loss": 1.4386, + "theoretical_loss": 3.5229685176617576, + "tokens_seen": 1609236480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025880641925777334, + "loss": 1.459, + "theoretical_loss": 3.5229561388798025, + "tokens_seen": 1609302016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587963891675025, + "loss": 1.3872, + "theoretical_loss": 3.5229437607430842, + "tokens_seen": 1609367552 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587863590772317, + "loss": 1.4117, + "theoretical_loss": 3.5229313832515423, + "tokens_seen": 1609433088 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587763289869609, + "loss": 1.2866, + "theoretical_loss": 3.5229190064051172, + "tokens_seen": 1609498624 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025876629889669007, + "loss": 1.2838, + "theoretical_loss": 3.522906630203749, + "tokens_seen": 1609564160 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587562688064193, + "loss": 1.3363, + "theoretical_loss": 3.522894254647378, + "tokens_seen": 1609629696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025874623871614843, + "loss": 1.3548, + "theoretical_loss": 3.522881879735943, + "tokens_seen": 1609695232 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025873620862587767, + "loss": 1.4118, + "theoretical_loss": 3.5228695054693864, + "tokens_seen": 1609760768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587261785356068, + "loss": 1.3121, + "theoretical_loss": 3.5228571318476467, + "tokens_seen": 1609826304 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025871614844533603, + "loss": 1.3728, + "theoretical_loss": 3.522844758870664, + "tokens_seen": 1609891840 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002587061183550652, + "loss": 1.4689, + "theoretical_loss": 3.5228323865383793, + "tokens_seen": 1609957376 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002586960882647944, + "loss": 1.401, + "theoretical_loss": 3.5228200148507325, + "tokens_seen": 1610022912 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025868605817452357, + "loss": 1.2942, + "theoretical_loss": 3.5228076438076643, + "tokens_seen": 1610088448 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025867602808425275, + "loss": 1.409, + "theoretical_loss": 3.5227952734091135, + "tokens_seen": 1610153984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025866599799398193, + "loss": 1.3389, + "theoretical_loss": 3.5227829036550213, + "tokens_seen": 1610219520 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025865596790371117, + "loss": 1.3951, + "theoretical_loss": 3.5227705345453275, + "tokens_seen": 1610285056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002586459378134403, + "loss": 1.3846, + "theoretical_loss": 3.5227581660799725, + "tokens_seen": 1610350592 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025863590772316953, + "loss": 1.3751, + "theoretical_loss": 3.5227457982588968, + "tokens_seen": 1610416128 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002586258776328987, + "loss": 1.3574, + "theoretical_loss": 3.52273343108204, + "tokens_seen": 1610481664 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002586158475426279, + "loss": 1.379, + "theoretical_loss": 3.522721064549343, + "tokens_seen": 1610547200 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002586058174523571, + "loss": 1.3376, + "theoretical_loss": 3.5227086986607454, + "tokens_seen": 1610612736 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025859578736208626, + "loss": 1.3333, + "theoretical_loss": 3.522696333416188, + "tokens_seen": 1610678272 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025858575727181544, + "loss": 1.3827, + "theoretical_loss": 3.5226839688156106, + "tokens_seen": 1610743808 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025857572718154467, + "loss": 1.4643, + "theoretical_loss": 3.522671604858953, + "tokens_seen": 1610809344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002585656970912738, + "loss": 1.3853, + "theoretical_loss": 3.5226592415461573, + "tokens_seen": 1610874880 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025855566700100303, + "loss": 1.4131, + "theoretical_loss": 3.522646878877162, + "tokens_seen": 1610940416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025854563691073216, + "loss": 1.2847, + "theoretical_loss": 3.522634516851908, + "tokens_seen": 1611005952 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002585356068204614, + "loss": 1.2463, + "theoretical_loss": 3.5226221554703354, + "tokens_seen": 1611071488 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002585255767301906, + "loss": 1.3535, + "theoretical_loss": 3.522609794732385, + "tokens_seen": 1611137024 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025851554663991976, + "loss": 1.2698, + "theoretical_loss": 3.5225974346379965, + "tokens_seen": 1611202560 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025850551654964894, + "loss": 1.4673, + "theoretical_loss": 3.52258507518711, + "tokens_seen": 1611268096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584954864593782, + "loss": 1.4335, + "theoretical_loss": 3.5225727163796674, + "tokens_seen": 1611333632 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584854563691073, + "loss": 1.3012, + "theoretical_loss": 3.5225603582156078, + "tokens_seen": 1611399168 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025847542627883654, + "loss": 1.3258, + "theoretical_loss": 3.5225480006948713, + "tokens_seen": 1611464704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025846539618856566, + "loss": 1.2906, + "theoretical_loss": 3.5225356438173985, + "tokens_seen": 1611530240 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584553660982949, + "loss": 1.4901, + "theoretical_loss": 3.5225232875831303, + "tokens_seen": 1611595776 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584453360080241, + "loss": 1.4018, + "theoretical_loss": 3.5225109319920067, + "tokens_seen": 1611661312 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025843530591775326, + "loss": 1.3637, + "theoretical_loss": 3.522498577043968, + "tokens_seen": 1611726848 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025842527582748244, + "loss": 1.4485, + "theoretical_loss": 3.522486222738954, + "tokens_seen": 1611792384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584152457372116, + "loss": 1.5143, + "theoretical_loss": 3.522473869076906, + "tokens_seen": 1611857920 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002584052156469408, + "loss": 1.3151, + "theoretical_loss": 3.5224615160577644, + "tokens_seen": 1611923456 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025839518555667004, + "loss": 1.4198, + "theoretical_loss": 3.5224491636814697, + "tokens_seen": 1611988992 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025838515546639917, + "loss": 1.3006, + "theoretical_loss": 3.522436811947961, + "tokens_seen": 1612054528 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002583751253761284, + "loss": 1.3795, + "theoretical_loss": 3.522424460857181, + "tokens_seen": 1612120064 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 588885, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.410120964050293, + "objective/train/theoretical_loss": 3.5224121104090678, + "objective/train/tokens_used": 1632645600, + "theoretical_loss": 3.5224121104090678, + "tokens_seen": 1612185600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025836509528585753, + "loss": 1.3372, + "theoretical_loss": 3.5224121104090678, + "tokens_seen": 1612185600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025835506519558677, + "loss": 1.3622, + "theoretical_loss": 3.5223997606035624, + "tokens_seen": 1612251136 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025834503510531595, + "loss": 1.3202, + "theoretical_loss": 3.5223874114406066, + "tokens_seen": 1612316672 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025833500501504513, + "loss": 1.4726, + "theoretical_loss": 3.5223750629201396, + "tokens_seen": 1612382208 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002583249749247743, + "loss": 1.3936, + "theoretical_loss": 3.522362715042102, + "tokens_seen": 1612447744 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025831494483450354, + "loss": 1.2038, + "theoretical_loss": 3.5223503678064345, + "tokens_seen": 1612513280 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025830491474423267, + "loss": 1.3591, + "theoretical_loss": 3.522338021213078, + "tokens_seen": 1612578816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002582948846539619, + "loss": 1.2431, + "theoretical_loss": 3.522325675261972, + "tokens_seen": 1612644352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025828485456369103, + "loss": 1.4423, + "theoretical_loss": 3.5223133299530573, + "tokens_seen": 1612709888 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025827482447342027, + "loss": 1.4061, + "theoretical_loss": 3.522300985286275, + "tokens_seen": 1612775424 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025826479438314945, + "loss": 1.3653, + "theoretical_loss": 3.5222886412615653, + "tokens_seen": 1612840960 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025825476429287863, + "loss": 1.3583, + "theoretical_loss": 3.522276297878869, + "tokens_seen": 1612906496 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025824473420260787, + "loss": 1.3978, + "theoretical_loss": 3.5222639551381256, + "tokens_seen": 1612972032 + }, + { + "epoch": 0.49, + "learning_rate": 0.000258234704112337, + "loss": 1.3729, + "theoretical_loss": 3.5222516130392765, + "tokens_seen": 1613037568 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025822467402206623, + "loss": 1.3642, + "theoretical_loss": 3.522239271582262, + "tokens_seen": 1613103104 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002582146439317954, + "loss": 1.2928, + "theoretical_loss": 3.5222269307670224, + "tokens_seen": 1613168640 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002582046138415246, + "loss": 1.4253, + "theoretical_loss": 3.522214590593499, + "tokens_seen": 1613234176 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025819458375125377, + "loss": 1.3665, + "theoretical_loss": 3.5222022510616315, + "tokens_seen": 1613299712 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025818455366098295, + "loss": 1.3794, + "theoretical_loss": 3.5221899121713616, + "tokens_seen": 1613365248 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025817452357071213, + "loss": 1.4272, + "theoretical_loss": 3.522177573922628, + "tokens_seen": 1613430784 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025816449348044137, + "loss": 1.4351, + "theoretical_loss": 3.5221652363153733, + "tokens_seen": 1613496320 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002581544633901705, + "loss": 1.3633, + "theoretical_loss": 3.5221528993495372, + "tokens_seen": 1613561856 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025814443329989973, + "loss": 1.3191, + "theoretical_loss": 3.5221405630250597, + "tokens_seen": 1613627392 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002581344032096289, + "loss": 1.368, + "theoretical_loss": 3.5221282273418826, + "tokens_seen": 1613692928 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002581243731193581, + "loss": 1.3646, + "theoretical_loss": 3.5221158922999463, + "tokens_seen": 1613758464 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002581143430290873, + "loss": 1.4064, + "theoretical_loss": 3.5221035578991904, + "tokens_seen": 1613824000 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025810431293881646, + "loss": 1.3892, + "theoretical_loss": 3.5220912241395563, + "tokens_seen": 1613889536 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025809428284854564, + "loss": 1.4481, + "theoretical_loss": 3.522078891020985, + "tokens_seen": 1613955072 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025808425275827487, + "loss": 1.3573, + "theoretical_loss": 3.5220665585434165, + "tokens_seen": 1614020608 + }, + { + "epoch": 0.49, + "learning_rate": 0.000258074222668004, + "loss": 1.4313, + "theoretical_loss": 3.522054226706792, + "tokens_seen": 1614086144 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025806419257773323, + "loss": 1.4801, + "theoretical_loss": 3.522041895511051, + "tokens_seen": 1614151680 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025805416248746236, + "loss": 1.4446, + "theoretical_loss": 3.522029564956136, + "tokens_seen": 1614217216 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002580441323971916, + "loss": 1.5045, + "theoretical_loss": 3.522017235041986, + "tokens_seen": 1614282752 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002580341023069208, + "loss": 1.4171, + "theoretical_loss": 3.5220049057685427, + "tokens_seen": 1614348288 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025802407221664996, + "loss": 1.4181, + "theoretical_loss": 3.5219925771357463, + "tokens_seen": 1614413824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025801404212637914, + "loss": 1.4151, + "theoretical_loss": 3.521980249143538, + "tokens_seen": 1614479360 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002580040120361084, + "loss": 1.4636, + "theoretical_loss": 3.521967921791858, + "tokens_seen": 1614544896 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002579939819458375, + "loss": 1.2892, + "theoretical_loss": 3.521955595080647, + "tokens_seen": 1614610432 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025798395185556674, + "loss": 1.4478, + "theoretical_loss": 3.5219432690098467, + "tokens_seen": 1614675968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025797392176529586, + "loss": 1.4057, + "theoretical_loss": 3.521930943579397, + "tokens_seen": 1614741504 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002579638916750251, + "loss": 1.4097, + "theoretical_loss": 3.521918618789238, + "tokens_seen": 1614807040 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002579538615847543, + "loss": 1.335, + "theoretical_loss": 3.521906294639312, + "tokens_seen": 1614872576 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025794383149448346, + "loss": 1.3611, + "theoretical_loss": 3.5218939711295585, + "tokens_seen": 1614938112 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025793380140421264, + "loss": 1.4947, + "theoretical_loss": 3.521881648259919, + "tokens_seen": 1615003648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002579237713139418, + "loss": 1.3327, + "theoretical_loss": 3.521869326030334, + "tokens_seen": 1615069184 + }, + { + "epoch": 0.49, + "learning_rate": 0.000257913741223671, + "loss": 1.3841, + "theoretical_loss": 3.521857004440744, + "tokens_seen": 1615134720 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025790371113340024, + "loss": 1.2964, + "theoretical_loss": 3.52184468349109, + "tokens_seen": 1615200256 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025789368104312937, + "loss": 1.3391, + "theoretical_loss": 3.5218323631813133, + "tokens_seen": 1615265792 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002578836509528586, + "loss": 1.4843, + "theoretical_loss": 3.5218200435113545, + "tokens_seen": 1615331328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025787362086258773, + "loss": 1.3349, + "theoretical_loss": 3.521807724481154, + "tokens_seen": 1615396864 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 590099, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3131601810455322, + "objective/train/theoretical_loss": 3.5217954060906527, + "objective/train/tokens_used": 1635922400, + "theoretical_loss": 3.5217954060906527, + "tokens_seen": 1615462400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025786359077231697, + "loss": 1.3475, + "theoretical_loss": 3.5217954060906527, + "tokens_seen": 1615462400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025785356068204615, + "loss": 1.285, + "theoretical_loss": 3.5217830883397916, + "tokens_seen": 1615527936 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025784353059177533, + "loss": 1.382, + "theoretical_loss": 3.521770771228512, + "tokens_seen": 1615593472 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002578335005015045, + "loss": 1.4362, + "theoretical_loss": 3.5217584547567538, + "tokens_seen": 1615659008 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025782347041123374, + "loss": 1.4913, + "theoretical_loss": 3.521746138924458, + "tokens_seen": 1615724544 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025781344032096287, + "loss": 1.3686, + "theoretical_loss": 3.5217338237315667, + "tokens_seen": 1615790080 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002578034102306921, + "loss": 1.2711, + "theoretical_loss": 3.521721509178019, + "tokens_seen": 1615855616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025779338014042123, + "loss": 1.4065, + "theoretical_loss": 3.5217091952637567, + "tokens_seen": 1615921152 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025778335005015047, + "loss": 1.3308, + "theoretical_loss": 3.5216968819887216, + "tokens_seen": 1615986688 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025777331995987965, + "loss": 1.2742, + "theoretical_loss": 3.5216845693528525, + "tokens_seen": 1616052224 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025776328986960883, + "loss": 1.4478, + "theoretical_loss": 3.5216722573560917, + "tokens_seen": 1616117760 + }, + { + "epoch": 0.49, + "learning_rate": 0.000257753259779338, + "loss": 1.3063, + "theoretical_loss": 3.5216599459983806, + "tokens_seen": 1616183296 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002577432296890672, + "loss": 1.3399, + "theoretical_loss": 3.5216476352796584, + "tokens_seen": 1616248832 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002577331995987964, + "loss": 1.3648, + "theoretical_loss": 3.5216353251998673, + "tokens_seen": 1616314368 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002577231695085256, + "loss": 1.4306, + "theoretical_loss": 3.5216230157589483, + "tokens_seen": 1616379904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025771313941825474, + "loss": 1.4004, + "theoretical_loss": 3.5216107069568414, + "tokens_seen": 1616445440 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025770310932798397, + "loss": 1.3667, + "theoretical_loss": 3.5215983987934885, + "tokens_seen": 1616510976 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576930792377131, + "loss": 1.4839, + "theoretical_loss": 3.52158609126883, + "tokens_seen": 1616576512 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025768304914744233, + "loss": 1.3329, + "theoretical_loss": 3.5215737843828068, + "tokens_seen": 1616642048 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576730190571715, + "loss": 1.4196, + "theoretical_loss": 3.521561478135361, + "tokens_seen": 1616707584 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576629889669007, + "loss": 1.3509, + "theoretical_loss": 3.5215491725264316, + "tokens_seen": 1616773120 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576529588766299, + "loss": 1.3294, + "theoretical_loss": 3.5215368675559615, + "tokens_seen": 1616838656 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576429287863591, + "loss": 1.3322, + "theoretical_loss": 3.52152456322389, + "tokens_seen": 1616904192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025763289869608824, + "loss": 1.3768, + "theoretical_loss": 3.52151225953016, + "tokens_seen": 1616969728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576228686058175, + "loss": 1.4384, + "theoretical_loss": 3.521499956474711, + "tokens_seen": 1617035264 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002576128385155466, + "loss": 1.475, + "theoretical_loss": 3.521487654057484, + "tokens_seen": 1617100800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025760280842527584, + "loss": 1.4009, + "theoretical_loss": 3.5214753522784212, + "tokens_seen": 1617166336 + }, + { + "epoch": 0.49, + "learning_rate": 0.000257592778335005, + "loss": 1.4434, + "theoretical_loss": 3.521463051137463, + "tokens_seen": 1617231872 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002575827482447342, + "loss": 1.3443, + "theoretical_loss": 3.5214507506345507, + "tokens_seen": 1617297408 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002575727181544634, + "loss": 1.3569, + "theoretical_loss": 3.5214384507696246, + "tokens_seen": 1617362944 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025756268806419256, + "loss": 1.4309, + "theoretical_loss": 3.5214261515426264, + "tokens_seen": 1617428480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025755265797392174, + "loss": 1.3583, + "theoretical_loss": 3.521413852953497, + "tokens_seen": 1617494016 + }, + { + "epoch": 0.49, + "learning_rate": 0.000257542627883651, + "loss": 1.2919, + "theoretical_loss": 3.5214015550021776, + "tokens_seen": 1617559552 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002575325977933801, + "loss": 1.2863, + "theoretical_loss": 3.5213892576886088, + "tokens_seen": 1617625088 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025752256770310934, + "loss": 1.2626, + "theoretical_loss": 3.5213769610127326, + "tokens_seen": 1617690624 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025751253761283847, + "loss": 1.3735, + "theoretical_loss": 3.521364664974489, + "tokens_seen": 1617756160 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002575025075225677, + "loss": 1.3387, + "theoretical_loss": 3.52135236957382, + "tokens_seen": 1617821696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025749247743229694, + "loss": 1.3643, + "theoretical_loss": 3.5213400748106665, + "tokens_seen": 1617887232 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025748244734202606, + "loss": 1.4104, + "theoretical_loss": 3.5213277806849694, + "tokens_seen": 1617952768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002574724172517553, + "loss": 1.3655, + "theoretical_loss": 3.5213154871966696, + "tokens_seen": 1618018304 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002574623871614845, + "loss": 1.4266, + "theoretical_loss": 3.521303194345709, + "tokens_seen": 1618083840 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025745235707121366, + "loss": 1.3806, + "theoretical_loss": 3.5212909021320282, + "tokens_seen": 1618149376 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025744232698094284, + "loss": 1.3366, + "theoretical_loss": 3.5212786105555685, + "tokens_seen": 1618214912 + }, + { + "epoch": 0.49, + "learning_rate": 0.000257432296890672, + "loss": 1.4691, + "theoretical_loss": 3.521266319616271, + "tokens_seen": 1618280448 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002574222668004012, + "loss": 1.3742, + "theoretical_loss": 3.521254029314077, + "tokens_seen": 1618345984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025741223671013044, + "loss": 1.3682, + "theoretical_loss": 3.5212417396489273, + "tokens_seen": 1618411520 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025740220661985957, + "loss": 1.3919, + "theoretical_loss": 3.5212294506207638, + "tokens_seen": 1618477056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002573921765295888, + "loss": 1.3051, + "theoretical_loss": 3.5212171622295267, + "tokens_seen": 1618542592 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025738214643931793, + "loss": 1.3583, + "theoretical_loss": 3.521204874475158, + "tokens_seen": 1618608128 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025737211634904717, + "loss": 1.4331, + "theoretical_loss": 3.521192587357599, + "tokens_seen": 1618673664 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 591377, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2145819664001465, + "objective/train/theoretical_loss": 3.5211803008767903, + "objective/train/tokens_used": 1639199200, + "theoretical_loss": 3.5211803008767903, + "tokens_seen": 1618739200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025736208625877635, + "loss": 1.1805, + "theoretical_loss": 3.5211803008767903, + "tokens_seen": 1618739200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025735205616850553, + "loss": 1.3935, + "theoretical_loss": 3.521168015032674, + "tokens_seen": 1618804736 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002573420260782347, + "loss": 1.4002, + "theoretical_loss": 3.52115572982519, + "tokens_seen": 1618870272 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025733199598796394, + "loss": 1.433, + "theoretical_loss": 3.521143445254281, + "tokens_seen": 1618935808 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025732196589769307, + "loss": 1.4255, + "theoretical_loss": 3.521131161319887, + "tokens_seen": 1619001344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002573119358074223, + "loss": 1.367, + "theoretical_loss": 3.52111887802195, + "tokens_seen": 1619066880 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025730190571715143, + "loss": 1.4224, + "theoretical_loss": 3.5211065953604104, + "tokens_seen": 1619132416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025729187562688067, + "loss": 1.2446, + "theoretical_loss": 3.521094313335211, + "tokens_seen": 1619197952 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025728184553660985, + "loss": 1.3534, + "theoretical_loss": 3.5210820319462917, + "tokens_seen": 1619263488 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025727181544633903, + "loss": 1.3942, + "theoretical_loss": 3.521069751193594, + "tokens_seen": 1619329024 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002572617853560682, + "loss": 1.2724, + "theoretical_loss": 3.521057471077061, + "tokens_seen": 1619394560 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002572517552657974, + "loss": 1.3168, + "theoretical_loss": 3.521045191596631, + "tokens_seen": 1619460096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002572417251755266, + "loss": 1.4822, + "theoretical_loss": 3.521032912752247, + "tokens_seen": 1619525632 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002572316950852558, + "loss": 1.3868, + "theoretical_loss": 3.521020634543851, + "tokens_seen": 1619591168 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025722166499498494, + "loss": 1.326, + "theoretical_loss": 3.5210083569713824, + "tokens_seen": 1619656704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025721163490471417, + "loss": 1.4537, + "theoretical_loss": 3.5209960800347844, + "tokens_seen": 1619722240 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002572016048144433, + "loss": 1.3957, + "theoretical_loss": 3.520983803733997, + "tokens_seen": 1619787776 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025719157472417253, + "loss": 1.3538, + "theoretical_loss": 3.520971528068962, + "tokens_seen": 1619853312 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571815446339017, + "loss": 1.3936, + "theoretical_loss": 3.5209592530396217, + "tokens_seen": 1619918848 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571715145436309, + "loss": 1.3928, + "theoretical_loss": 3.5209469786459158, + "tokens_seen": 1619984384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571614844533601, + "loss": 1.3414, + "theoretical_loss": 3.5209347048877864, + "tokens_seen": 1620049920 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571514543630893, + "loss": 1.3713, + "theoretical_loss": 3.5209224317651753, + "tokens_seen": 1620115456 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025714142427281844, + "loss": 1.4057, + "theoretical_loss": 3.5209101592780234, + "tokens_seen": 1620180992 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571313941825477, + "loss": 1.5087, + "theoretical_loss": 3.520897887426272, + "tokens_seen": 1620246528 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571213640922768, + "loss": 1.3207, + "theoretical_loss": 3.5208856162098634, + "tokens_seen": 1620312064 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025711133400200604, + "loss": 1.3851, + "theoretical_loss": 3.520873345628738, + "tokens_seen": 1620377600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002571013039117352, + "loss": 1.2649, + "theoretical_loss": 3.5208610756828373, + "tokens_seen": 1620443136 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570912738214644, + "loss": 1.3668, + "theoretical_loss": 3.5208488063721033, + "tokens_seen": 1620508672 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570812437311936, + "loss": 1.3496, + "theoretical_loss": 3.5208365376964768, + "tokens_seen": 1620574208 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025707121364092276, + "loss": 1.4476, + "theoretical_loss": 3.5208242696558996, + "tokens_seen": 1620639744 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025706118355065194, + "loss": 1.2646, + "theoretical_loss": 3.520812002250314, + "tokens_seen": 1620705280 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570511534603812, + "loss": 1.3099, + "theoretical_loss": 3.5207997354796596, + "tokens_seen": 1620770816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570411233701103, + "loss": 1.3241, + "theoretical_loss": 3.5207874693438788, + "tokens_seen": 1620836352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025703109327983954, + "loss": 1.3156, + "theoretical_loss": 3.5207752038429136, + "tokens_seen": 1620901888 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025702106318956867, + "loss": 1.3874, + "theoretical_loss": 3.520762938976705, + "tokens_seen": 1620967424 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570110330992979, + "loss": 1.3464, + "theoretical_loss": 3.5207506747451944, + "tokens_seen": 1621032960 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002570010030090271, + "loss": 1.3858, + "theoretical_loss": 3.520738411148323, + "tokens_seen": 1621098496 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025699097291875626, + "loss": 1.3598, + "theoretical_loss": 3.520726148186033, + "tokens_seen": 1621164032 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025698094282848545, + "loss": 1.3668, + "theoretical_loss": 3.520713885858266, + "tokens_seen": 1621229568 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002569709127382147, + "loss": 1.3142, + "theoretical_loss": 3.5207016241649627, + "tokens_seen": 1621295104 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002569608826479438, + "loss": 1.3955, + "theoretical_loss": 3.520689363106065, + "tokens_seen": 1621360640 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025695085255767304, + "loss": 1.4042, + "theoretical_loss": 3.520677102681515, + "tokens_seen": 1621426176 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025694082246740217, + "loss": 1.469, + "theoretical_loss": 3.520664842891253, + "tokens_seen": 1621491712 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002569307923771314, + "loss": 1.3459, + "theoretical_loss": 3.5206525837352216, + "tokens_seen": 1621557248 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002569207622868606, + "loss": 1.4494, + "theoretical_loss": 3.5206403252133622, + "tokens_seen": 1621622784 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025691073219658977, + "loss": 1.436, + "theoretical_loss": 3.5206280673256156, + "tokens_seen": 1621688320 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025690070210631895, + "loss": 1.3109, + "theoretical_loss": 3.5206158100719245, + "tokens_seen": 1621753856 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025689067201604813, + "loss": 1.4148, + "theoretical_loss": 3.52060355345223, + "tokens_seen": 1621819392 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002568806419257773, + "loss": 1.3076, + "theoretical_loss": 3.5205912974664737, + "tokens_seen": 1621884928 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025687061183550655, + "loss": 1.3804, + "theoretical_loss": 3.520579042114597, + "tokens_seen": 1621950464 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 592532, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.285015344619751, + "objective/train/theoretical_loss": 3.5205667873965414, + "objective/train/tokens_used": 1642476000, + "theoretical_loss": 3.5205667873965414, + "tokens_seen": 1622016000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002568605817452357, + "loss": 1.3199, + "theoretical_loss": 3.5205667873965414, + "tokens_seen": 1622016000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002568505516549649, + "loss": 1.3436, + "theoretical_loss": 3.520554533312249, + "tokens_seen": 1622081536 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025684052156469404, + "loss": 1.3328, + "theoretical_loss": 3.5205422798616617, + "tokens_seen": 1622147072 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025683049147442327, + "loss": 1.3607, + "theoretical_loss": 3.52053002704472, + "tokens_seen": 1622212608 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025682046138415245, + "loss": 1.319, + "theoretical_loss": 3.520517774861366, + "tokens_seen": 1622278144 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025681043129388163, + "loss": 1.3983, + "theoretical_loss": 3.520505523311542, + "tokens_seen": 1622343680 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002568004012036108, + "loss": 1.4257, + "theoretical_loss": 3.520493272395189, + "tokens_seen": 1622409216 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025679037111334005, + "loss": 1.3131, + "theoretical_loss": 3.5204810221122487, + "tokens_seen": 1622474752 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002567803410230692, + "loss": 1.3303, + "theoretical_loss": 3.520468772462663, + "tokens_seen": 1622540288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002567703109327984, + "loss": 1.4328, + "theoretical_loss": 3.5204565234463736, + "tokens_seen": 1622605824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025676028084252754, + "loss": 1.4183, + "theoretical_loss": 3.520444275063322, + "tokens_seen": 1622671360 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002567502507522568, + "loss": 1.412, + "theoretical_loss": 3.5204320273134493, + "tokens_seen": 1622736896 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256740220661986, + "loss": 1.3631, + "theoretical_loss": 3.5204197801966983, + "tokens_seen": 1622802432 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025673019057171514, + "loss": 1.3428, + "theoretical_loss": 3.5204075337130103, + "tokens_seen": 1622867968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025672016048144437, + "loss": 1.2186, + "theoretical_loss": 3.5203952878623266, + "tokens_seen": 1622933504 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002567101303911735, + "loss": 1.3923, + "theoretical_loss": 3.52038304264459, + "tokens_seen": 1622999040 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025670010030090273, + "loss": 1.2982, + "theoretical_loss": 3.5203707980597407, + "tokens_seen": 1623064576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566900702106319, + "loss": 1.3664, + "theoretical_loss": 3.5203585541077214, + "tokens_seen": 1623130112 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566800401203611, + "loss": 1.3418, + "theoretical_loss": 3.520346310788474, + "tokens_seen": 1623195648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566700100300903, + "loss": 1.3712, + "theoretical_loss": 3.5203340681019397, + "tokens_seen": 1623261184 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566599799398195, + "loss": 1.3594, + "theoretical_loss": 3.5203218260480607, + "tokens_seen": 1623326720 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025664994984954864, + "loss": 1.365, + "theoretical_loss": 3.5203095846267782, + "tokens_seen": 1623392256 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566399197592779, + "loss": 1.4399, + "theoretical_loss": 3.5202973438380347, + "tokens_seen": 1623457792 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256629889669007, + "loss": 1.4189, + "theoretical_loss": 3.5202851036817715, + "tokens_seen": 1623523328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025661985957873624, + "loss": 1.4193, + "theoretical_loss": 3.5202728641579304, + "tokens_seen": 1623588864 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002566098294884654, + "loss": 1.4277, + "theoretical_loss": 3.5202606252664532, + "tokens_seen": 1623654400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565997993981946, + "loss": 1.4379, + "theoretical_loss": 3.520248387007282, + "tokens_seen": 1623719936 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565897693079238, + "loss": 1.3965, + "theoretical_loss": 3.5202361493803584, + "tokens_seen": 1623785472 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025657973921765296, + "loss": 1.3043, + "theoretical_loss": 3.520223912385624, + "tokens_seen": 1623851008 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025656970912738214, + "loss": 1.4286, + "theoretical_loss": 3.5202116760230213, + "tokens_seen": 1623916544 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565596790371114, + "loss": 1.3197, + "theoretical_loss": 3.5201994402924917, + "tokens_seen": 1623982080 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565496489468405, + "loss": 1.326, + "theoretical_loss": 3.5201872051939764, + "tokens_seen": 1624047616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025653961885656974, + "loss": 1.441, + "theoretical_loss": 3.5201749707274184, + "tokens_seen": 1624113152 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025652958876629887, + "loss": 1.4097, + "theoretical_loss": 3.5201627368927593, + "tokens_seen": 1624178688 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565195586760281, + "loss": 1.3984, + "theoretical_loss": 3.5201505036899405, + "tokens_seen": 1624244224 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002565095285857573, + "loss": 1.3388, + "theoretical_loss": 3.520138271118904, + "tokens_seen": 1624309760 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025649949849548647, + "loss": 1.4298, + "theoretical_loss": 3.520126039179592, + "tokens_seen": 1624375296 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025648946840521565, + "loss": 1.4332, + "theoretical_loss": 3.520113807871946, + "tokens_seen": 1624440832 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002564794383149449, + "loss": 1.2544, + "theoretical_loss": 3.5201015771959083, + "tokens_seen": 1624506368 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256469408224674, + "loss": 1.3915, + "theoretical_loss": 3.52008934715142, + "tokens_seen": 1624571904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025645937813440324, + "loss": 1.3191, + "theoretical_loss": 3.520077117738424, + "tokens_seen": 1624637440 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025644934804413237, + "loss": 1.3539, + "theoretical_loss": 3.5200648889568624, + "tokens_seen": 1624702976 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002564393179538616, + "loss": 1.3406, + "theoretical_loss": 3.520052660806676, + "tokens_seen": 1624768512 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002564292878635908, + "loss": 1.3874, + "theoretical_loss": 3.5200404332878072, + "tokens_seen": 1624834048 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025641925777331997, + "loss": 1.4952, + "theoretical_loss": 3.520028206400198, + "tokens_seen": 1624899584 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025640922768304915, + "loss": 1.3427, + "theoretical_loss": 3.5200159801437905, + "tokens_seen": 1624965120 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025639919759277833, + "loss": 1.4397, + "theoretical_loss": 3.5200037545185268, + "tokens_seen": 1625030656 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002563891675025075, + "loss": 1.3525, + "theoretical_loss": 3.519991529524348, + "tokens_seen": 1625096192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025637913741223675, + "loss": 1.4355, + "theoretical_loss": 3.5199793051611974, + "tokens_seen": 1625161728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002563691073219659, + "loss": 1.3382, + "theoretical_loss": 3.519967081429016, + "tokens_seen": 1625227264 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 593743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5765501260757446, + "objective/train/theoretical_loss": 3.519954858327746, + "objective/train/tokens_used": 1645752800, + "theoretical_loss": 3.519954858327746, + "tokens_seen": 1625292800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002563590772316951, + "loss": 1.4546, + "theoretical_loss": 3.519954858327746, + "tokens_seen": 1625292800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025634904714142424, + "loss": 1.3798, + "theoretical_loss": 3.5199426358573294, + "tokens_seen": 1625358336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025633901705115347, + "loss": 1.4292, + "theoretical_loss": 3.5199304140177086, + "tokens_seen": 1625423872 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025632898696088265, + "loss": 1.5631, + "theoretical_loss": 3.5199181928088246, + "tokens_seen": 1625489408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025631895687061183, + "loss": 1.3408, + "theoretical_loss": 3.519905972230621, + "tokens_seen": 1625554944 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256308926780341, + "loss": 1.4016, + "theoretical_loss": 3.5198937522830382, + "tokens_seen": 1625620480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025629889669007025, + "loss": 1.3805, + "theoretical_loss": 3.5198815329660196, + "tokens_seen": 1625686016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002562888665997994, + "loss": 1.3724, + "theoretical_loss": 3.519869314279506, + "tokens_seen": 1625751552 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002562788365095286, + "loss": 1.3597, + "theoretical_loss": 3.51985709622344, + "tokens_seen": 1625817088 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025626880641925774, + "loss": 1.2495, + "theoretical_loss": 3.5198448787977643, + "tokens_seen": 1625882624 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256258776328987, + "loss": 1.3948, + "theoretical_loss": 3.51983266200242, + "tokens_seen": 1625948160 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025624874623871616, + "loss": 1.3375, + "theoretical_loss": 3.51982044583735, + "tokens_seen": 1626013696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025623871614844534, + "loss": 1.3913, + "theoretical_loss": 3.5198082303024956, + "tokens_seen": 1626079232 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002562286860581745, + "loss": 1.3131, + "theoretical_loss": 3.519796015397799, + "tokens_seen": 1626144768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002562186559679037, + "loss": 1.443, + "theoretical_loss": 3.5197838011232028, + "tokens_seen": 1626210304 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002562086258776329, + "loss": 1.4626, + "theoretical_loss": 3.5197715874786493, + "tokens_seen": 1626275840 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002561985957873621, + "loss": 1.3368, + "theoretical_loss": 3.51975937446408, + "tokens_seen": 1626341376 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025618856569709124, + "loss": 1.3695, + "theoretical_loss": 3.5197471620794367, + "tokens_seen": 1626406912 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002561785356068205, + "loss": 1.3282, + "theoretical_loss": 3.519734950324662, + "tokens_seen": 1626472448 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025616850551654966, + "loss": 1.3179, + "theoretical_loss": 3.519722739199698, + "tokens_seen": 1626537984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025615847542627884, + "loss": 1.417, + "theoretical_loss": 3.519710528704487, + "tokens_seen": 1626603520 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256148445336008, + "loss": 1.5159, + "theoretical_loss": 3.5196983188389717, + "tokens_seen": 1626669056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002561384152457372, + "loss": 1.3163, + "theoretical_loss": 3.519686109603093, + "tokens_seen": 1626734592 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002561283851554664, + "loss": 1.3741, + "theoretical_loss": 3.5196739009967937, + "tokens_seen": 1626800128 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002561183550651956, + "loss": 1.423, + "theoretical_loss": 3.519661693020016, + "tokens_seen": 1626865664 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025610832497492475, + "loss": 1.2838, + "theoretical_loss": 3.5196494856727023, + "tokens_seen": 1626931200 + }, + { + "epoch": 0.49, + "learning_rate": 0.000256098294884654, + "loss": 1.464, + "theoretical_loss": 3.519637278954794, + "tokens_seen": 1626996736 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560882647943831, + "loss": 1.4209, + "theoretical_loss": 3.5196250728662344, + "tokens_seen": 1627062272 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025607823470411234, + "loss": 1.3868, + "theoretical_loss": 3.5196128674069644, + "tokens_seen": 1627127808 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560682046138415, + "loss": 1.3476, + "theoretical_loss": 3.519600662576927, + "tokens_seen": 1627193344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560581745235707, + "loss": 1.2959, + "theoretical_loss": 3.519588458376065, + "tokens_seen": 1627258880 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560481444332999, + "loss": 1.3215, + "theoretical_loss": 3.5195762548043197, + "tokens_seen": 1627324416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025603811434302907, + "loss": 1.399, + "theoretical_loss": 3.5195640518616336, + "tokens_seen": 1627389952 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025602808425275825, + "loss": 1.3504, + "theoretical_loss": 3.5195518495479488, + "tokens_seen": 1627455488 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560180541624875, + "loss": 1.3498, + "theoretical_loss": 3.5195396478632075, + "tokens_seen": 1627521024 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002560080240722166, + "loss": 1.4693, + "theoretical_loss": 3.519527446807353, + "tokens_seen": 1627586560 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025599799398194585, + "loss": 1.4195, + "theoretical_loss": 3.519515246380326, + "tokens_seen": 1627652096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002559879638916751, + "loss": 1.4734, + "theoretical_loss": 3.519503046582069, + "tokens_seen": 1627717632 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002559779338014042, + "loss": 1.3033, + "theoretical_loss": 3.519490847412526, + "tokens_seen": 1627783168 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025596790371113344, + "loss": 1.4244, + "theoretical_loss": 3.5194786488716376, + "tokens_seen": 1627848704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025595787362086257, + "loss": 1.3355, + "theoretical_loss": 3.5194664509593463, + "tokens_seen": 1627914240 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002559478435305918, + "loss": 1.378, + "theoretical_loss": 3.519454253675595, + "tokens_seen": 1627979776 + }, + { + "epoch": 0.49, + "learning_rate": 0.000255937813440321, + "loss": 1.3895, + "theoretical_loss": 3.519442057020325, + "tokens_seen": 1628045312 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025592778335005017, + "loss": 1.4563, + "theoretical_loss": 3.51942986099348, + "tokens_seen": 1628110848 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025591775325977935, + "loss": 1.3248, + "theoretical_loss": 3.519417665595002, + "tokens_seen": 1628176384 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025590772316950853, + "loss": 1.2847, + "theoretical_loss": 3.5194054708248323, + "tokens_seen": 1628241920 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002558976930792377, + "loss": 1.3831, + "theoretical_loss": 3.519393276682914, + "tokens_seen": 1628307456 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025588766298896695, + "loss": 1.298, + "theoretical_loss": 3.5193810831691894, + "tokens_seen": 1628372992 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002558776328986961, + "loss": 1.403, + "theoretical_loss": 3.5193688902836007, + "tokens_seen": 1628438528 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002558676028084253, + "loss": 1.387, + "theoretical_loss": 3.5193566980260904, + "tokens_seen": 1628504064 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 595003, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3390653133392334, + "objective/train/theoretical_loss": 3.5193445063966013, + "objective/train/tokens_used": 1649029600, + "theoretical_loss": 3.5193445063966013, + "tokens_seen": 1628569600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025585757271815444, + "loss": 1.2927, + "theoretical_loss": 3.5193445063966013, + "tokens_seen": 1628569600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025584754262788367, + "loss": 1.3803, + "theoretical_loss": 3.519332315395075, + "tokens_seen": 1628635136 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025583751253761285, + "loss": 1.4585, + "theoretical_loss": 3.519320125021454, + "tokens_seen": 1628700672 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025582748244734203, + "loss": 1.3419, + "theoretical_loss": 3.5193079352756813, + "tokens_seen": 1628766208 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002558174523570712, + "loss": 1.3245, + "theoretical_loss": 3.519295746157699, + "tokens_seen": 1628831744 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025580742226680045, + "loss": 1.4285, + "theoretical_loss": 3.519283557667449, + "tokens_seen": 1628897280 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557973921765296, + "loss": 1.3408, + "theoretical_loss": 3.519271369804874, + "tokens_seen": 1628962816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557873620862588, + "loss": 1.4021, + "theoretical_loss": 3.5192591825699173, + "tokens_seen": 1629028352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025577733199598794, + "loss": 1.3783, + "theoretical_loss": 3.5192469959625203, + "tokens_seen": 1629093888 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557673019057172, + "loss": 1.3575, + "theoretical_loss": 3.5192348099826254, + "tokens_seen": 1629159424 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025575727181544636, + "loss": 1.4586, + "theoretical_loss": 3.5192226246301757, + "tokens_seen": 1629224960 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025574724172517554, + "loss": 1.4572, + "theoretical_loss": 3.5192104399051134, + "tokens_seen": 1629290496 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557372116349047, + "loss": 1.3094, + "theoretical_loss": 3.519198255807381, + "tokens_seen": 1629356032 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557271815446339, + "loss": 1.3723, + "theoretical_loss": 3.5191860723369204, + "tokens_seen": 1629421568 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557171514543631, + "loss": 1.3627, + "theoretical_loss": 3.5191738894936746, + "tokens_seen": 1629487104 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002557071213640923, + "loss": 1.3682, + "theoretical_loss": 3.5191617072775863, + "tokens_seen": 1629552640 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025569709127382144, + "loss": 1.3081, + "theoretical_loss": 3.519149525688598, + "tokens_seen": 1629618176 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556870611835507, + "loss": 1.2869, + "theoretical_loss": 3.519137344726652, + "tokens_seen": 1629683712 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025567703109327986, + "loss": 1.387, + "theoretical_loss": 3.5191251643916903, + "tokens_seen": 1629749248 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025566700100300904, + "loss": 1.3293, + "theoretical_loss": 3.519112984683656, + "tokens_seen": 1629814784 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556569709127382, + "loss": 1.3141, + "theoretical_loss": 3.5191008056024917, + "tokens_seen": 1629880320 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556469408224674, + "loss": 1.3635, + "theoretical_loss": 3.5190886271481396, + "tokens_seen": 1629945856 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556369107321966, + "loss": 1.3118, + "theoretical_loss": 3.5190764493205426, + "tokens_seen": 1630011392 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556268806419258, + "loss": 1.3982, + "theoretical_loss": 3.5190642721196426, + "tokens_seen": 1630076928 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025561685055165495, + "loss": 1.4239, + "theoretical_loss": 3.5190520955453826, + "tokens_seen": 1630142464 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002556068204613842, + "loss": 1.3156, + "theoretical_loss": 3.519039919597706, + "tokens_seen": 1630208000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555967903711133, + "loss": 1.4023, + "theoretical_loss": 3.5190277442765536, + "tokens_seen": 1630273536 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025558676028084254, + "loss": 1.3661, + "theoretical_loss": 3.519015569581869, + "tokens_seen": 1630339072 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555767301905717, + "loss": 1.2544, + "theoretical_loss": 3.5190033955135944, + "tokens_seen": 1630404608 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555667001003009, + "loss": 1.4105, + "theoretical_loss": 3.5189912220716733, + "tokens_seen": 1630470144 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555566700100301, + "loss": 1.3299, + "theoretical_loss": 3.518979049256047, + "tokens_seen": 1630535680 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025554663991975927, + "loss": 1.3665, + "theoretical_loss": 3.5189668770666596, + "tokens_seen": 1630601216 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025553660982948845, + "loss": 1.3576, + "theoretical_loss": 3.518954705503452, + "tokens_seen": 1630666752 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555265797392177, + "loss": 1.4286, + "theoretical_loss": 3.5189425345663685, + "tokens_seen": 1630732288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002555165496489468, + "loss": 1.4141, + "theoretical_loss": 3.5189303642553504, + "tokens_seen": 1630797824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025550651955867605, + "loss": 1.4, + "theoretical_loss": 3.5189181945703414, + "tokens_seen": 1630863360 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025549648946840523, + "loss": 1.3524, + "theoretical_loss": 3.518906025511283, + "tokens_seen": 1630928896 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554864593781344, + "loss": 1.3125, + "theoretical_loss": 3.518893857078119, + "tokens_seen": 1630994432 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554764292878636, + "loss": 1.409, + "theoretical_loss": 3.518881689270791, + "tokens_seen": 1631059968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025546639919759277, + "loss": 1.3105, + "theoretical_loss": 3.518869522089243, + "tokens_seen": 1631125504 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025545636910732195, + "loss": 1.3549, + "theoretical_loss": 3.5188573555334166, + "tokens_seen": 1631191040 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554463390170512, + "loss": 1.2251, + "theoretical_loss": 3.518845189603254, + "tokens_seen": 1631256576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554363089267803, + "loss": 1.3018, + "theoretical_loss": 3.5188330242987, + "tokens_seen": 1631322112 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025542627883650955, + "loss": 1.4317, + "theoretical_loss": 3.5188208596196944, + "tokens_seen": 1631387648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554162487462387, + "loss": 1.2753, + "theoretical_loss": 3.5188086955661824, + "tokens_seen": 1631453184 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002554062186559679, + "loss": 1.4373, + "theoretical_loss": 3.518796532138106, + "tokens_seen": 1631518720 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553961885656971, + "loss": 1.3938, + "theoretical_loss": 3.518784369335407, + "tokens_seen": 1631584256 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553861584754263, + "loss": 1.4464, + "theoretical_loss": 3.5187722071580296, + "tokens_seen": 1631649792 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025537612838515546, + "loss": 1.3395, + "theoretical_loss": 3.518760045605915, + "tokens_seen": 1631715328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025536609829488464, + "loss": 1.3422, + "theoretical_loss": 3.518747884679007, + "tokens_seen": 1631780864 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 596193, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4781241416931152, + "objective/train/theoretical_loss": 3.5187357243772484, + "objective/train/tokens_used": 1652306400, + "theoretical_loss": 3.5187357243772484, + "tokens_seen": 1631846400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553560682046138, + "loss": 1.4168, + "theoretical_loss": 3.5187357243772484, + "tokens_seen": 1631846400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025534603811434305, + "loss": 1.4088, + "theoretical_loss": 3.518723564700581, + "tokens_seen": 1631911936 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553360080240722, + "loss": 1.4071, + "theoretical_loss": 3.518711405648949, + "tokens_seen": 1631977472 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553259779338014, + "loss": 1.3477, + "theoretical_loss": 3.518699247222293, + "tokens_seen": 1632043008 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553159478435306, + "loss": 1.3996, + "theoretical_loss": 3.5186870894205584, + "tokens_seen": 1632108544 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002553059177532598, + "loss": 1.409, + "theoretical_loss": 3.518674932243686, + "tokens_seen": 1632174080 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025529588766298896, + "loss": 1.3304, + "theoretical_loss": 3.51866277569162, + "tokens_seen": 1632239616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025528585757271814, + "loss": 1.3869, + "theoretical_loss": 3.518650619764302, + "tokens_seen": 1632305152 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552758274824473, + "loss": 1.3825, + "theoretical_loss": 3.5186384644616755, + "tokens_seen": 1632370688 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025526579739217656, + "loss": 1.4484, + "theoretical_loss": 3.518626309783683, + "tokens_seen": 1632436224 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552557673019057, + "loss": 1.4339, + "theoretical_loss": 3.5186141557302677, + "tokens_seen": 1632501760 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552457372116349, + "loss": 1.4422, + "theoretical_loss": 3.518602002301372, + "tokens_seen": 1632567296 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552357071213641, + "loss": 1.413, + "theoretical_loss": 3.518589849496939, + "tokens_seen": 1632632832 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552256770310933, + "loss": 1.3878, + "theoretical_loss": 3.518577697316912, + "tokens_seen": 1632698368 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002552156469408225, + "loss": 1.4068, + "theoretical_loss": 3.5185655457612324, + "tokens_seen": 1632763904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025520561685055164, + "loss": 1.3814, + "theoretical_loss": 3.518553394829845, + "tokens_seen": 1632829440 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551955867602809, + "loss": 1.2711, + "theoretical_loss": 3.518541244522691, + "tokens_seen": 1632894976 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025518555667001006, + "loss": 1.2872, + "theoretical_loss": 3.5185290948397148, + "tokens_seen": 1632960512 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025517552657973924, + "loss": 1.4074, + "theoretical_loss": 3.518516945780858, + "tokens_seen": 1633026048 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551654964894684, + "loss": 1.3709, + "theoretical_loss": 3.518504797346064, + "tokens_seen": 1633091584 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551554663991976, + "loss": 1.3317, + "theoretical_loss": 3.5184926495352755, + "tokens_seen": 1633157120 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551454363089268, + "loss": 1.2979, + "theoretical_loss": 3.518480502348435, + "tokens_seen": 1633222656 + }, + { + "epoch": 0.49, + "learning_rate": 0.000255135406218656, + "loss": 1.3436, + "theoretical_loss": 3.518468355785487, + "tokens_seen": 1633288192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00025512537612838515, + "loss": 1.3619, + "theoretical_loss": 3.518456209846373, + "tokens_seen": 1633353728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551153460381144, + "loss": 1.462, + "theoretical_loss": 3.5184440645310366, + "tokens_seen": 1633419264 + }, + { + "epoch": 0.49, + "learning_rate": 0.0002551053159478435, + "loss": 1.341, + "theoretical_loss": 3.5184319198394203, + "tokens_seen": 1633484800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025509528585757274, + "loss": 1.3792, + "theoretical_loss": 3.5184197757714673, + "tokens_seen": 1633550336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002550852557673019, + "loss": 1.4269, + "theoretical_loss": 3.5184076323271203, + "tokens_seen": 1633615872 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002550752256770311, + "loss": 1.2929, + "theoretical_loss": 3.5183954895063225, + "tokens_seen": 1633681408 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002550651955867603, + "loss": 1.3509, + "theoretical_loss": 3.518383347309017, + "tokens_seen": 1633746944 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025505516549648947, + "loss": 1.3644, + "theoretical_loss": 3.5183712057351464, + "tokens_seen": 1633812480 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025504513540621865, + "loss": 1.2805, + "theoretical_loss": 3.5183590647846543, + "tokens_seen": 1633878016 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002550351053159479, + "loss": 1.3297, + "theoretical_loss": 3.5183469244574828, + "tokens_seen": 1633943552 + }, + { + "epoch": 0.5, + "learning_rate": 0.000255025075225677, + "loss": 1.3451, + "theoretical_loss": 3.5183347847535753, + "tokens_seen": 1634009088 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025501504513540625, + "loss": 1.4042, + "theoretical_loss": 3.5183226456728756, + "tokens_seen": 1634074624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025500501504513543, + "loss": 1.3459, + "theoretical_loss": 3.5183105072153253, + "tokens_seen": 1634140160 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549949849548646, + "loss": 1.3496, + "theoretical_loss": 3.5182983693808683, + "tokens_seen": 1634205696 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549849548645938, + "loss": 1.383, + "theoretical_loss": 3.518286232169447, + "tokens_seen": 1634271232 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025497492477432297, + "loss": 1.3105, + "theoretical_loss": 3.5182740955810057, + "tokens_seen": 1634336768 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025496489468405215, + "loss": 1.3612, + "theoretical_loss": 3.5182619596154865, + "tokens_seen": 1634402304 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549548645937814, + "loss": 1.4112, + "theoretical_loss": 3.518249824272832, + "tokens_seen": 1634467840 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549448345035105, + "loss": 1.3407, + "theoretical_loss": 3.5182376895529863, + "tokens_seen": 1634533376 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025493480441323975, + "loss": 1.3397, + "theoretical_loss": 3.518225555455892, + "tokens_seen": 1634598912 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549247743229689, + "loss": 1.292, + "theoretical_loss": 3.5182134219814922, + "tokens_seen": 1634664448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549147442326981, + "loss": 1.3343, + "theoretical_loss": 3.51820128912973, + "tokens_seen": 1634729984 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002549047141424273, + "loss": 1.4132, + "theoretical_loss": 3.518189156900548, + "tokens_seen": 1634795520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002548946840521565, + "loss": 1.351, + "theoretical_loss": 3.51817702529389, + "tokens_seen": 1634861056 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025488465396188566, + "loss": 1.3767, + "theoretical_loss": 3.5181648943096993, + "tokens_seen": 1634926592 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025487462387161484, + "loss": 1.3941, + "theoretical_loss": 3.518152763947918, + "tokens_seen": 1634992128 + }, + { + "epoch": 0.5, + "learning_rate": 0.000254864593781344, + "loss": 1.3808, + "theoretical_loss": 3.51814063420849, + "tokens_seen": 1635057664 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 597373, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3879289627075195, + "objective/train/theoretical_loss": 3.518128505091358, + "objective/train/tokens_used": 1655583200, + "theoretical_loss": 3.518128505091358, + "tokens_seen": 1635123200 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025485456369107325, + "loss": 1.3448, + "theoretical_loss": 3.518128505091358, + "tokens_seen": 1635123200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002548445336008024, + "loss": 1.3657, + "theoretical_loss": 3.5181163765964656, + "tokens_seen": 1635188736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002548345035105316, + "loss": 1.3904, + "theoretical_loss": 3.5181042487237555, + "tokens_seen": 1635254272 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002548244734202608, + "loss": 1.4246, + "theoretical_loss": 3.5180921214731713, + "tokens_seen": 1635319808 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025481444332999, + "loss": 1.2342, + "theoretical_loss": 3.5180799948446557, + "tokens_seen": 1635385344 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025480441323971916, + "loss": 1.4722, + "theoretical_loss": 3.5180678688381524, + "tokens_seen": 1635450880 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025479438314944834, + "loss": 1.2997, + "theoretical_loss": 3.518055743453604, + "tokens_seen": 1635516416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002547843530591775, + "loss": 1.3019, + "theoretical_loss": 3.5180436186909536, + "tokens_seen": 1635581952 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025477432296890676, + "loss": 1.2796, + "theoretical_loss": 3.518031494550145, + "tokens_seen": 1635647488 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002547642928786359, + "loss": 1.3053, + "theoretical_loss": 3.518019371031121, + "tokens_seen": 1635713024 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002547542627883651, + "loss": 1.2997, + "theoretical_loss": 3.518007248133825, + "tokens_seen": 1635778560 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025474423269809425, + "loss": 1.3413, + "theoretical_loss": 3.5179951258582, + "tokens_seen": 1635844096 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002547342026078235, + "loss": 1.3362, + "theoretical_loss": 3.517983004204189, + "tokens_seen": 1635909632 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025472417251755266, + "loss": 1.2031, + "theoretical_loss": 3.5179708831717362, + "tokens_seen": 1635975168 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025471414242728184, + "loss": 1.2863, + "theoretical_loss": 3.5179587627607836, + "tokens_seen": 1636040704 + }, + { + "epoch": 0.5, + "learning_rate": 0.000254704112337011, + "loss": 1.3973, + "theoretical_loss": 3.5179466429712756, + "tokens_seen": 1636106240 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025469408224674026, + "loss": 1.4185, + "theoretical_loss": 3.5179345238031545, + "tokens_seen": 1636171776 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002546840521564694, + "loss": 1.4748, + "theoretical_loss": 3.517922405256364, + "tokens_seen": 1636237312 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002546740220661986, + "loss": 1.4643, + "theoretical_loss": 3.5179102873308468, + "tokens_seen": 1636302848 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025466399197592775, + "loss": 1.1649, + "theoretical_loss": 3.5178981700265473, + "tokens_seen": 1636368384 + }, + { + "epoch": 0.5, + "learning_rate": 0.000254653961885657, + "loss": 1.2907, + "theoretical_loss": 3.5178860533434078, + "tokens_seen": 1636433920 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025464393179538616, + "loss": 1.3556, + "theoretical_loss": 3.5178739372813723, + "tokens_seen": 1636499456 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025463390170511535, + "loss": 1.3048, + "theoretical_loss": 3.517861821840383, + "tokens_seen": 1636564992 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025462387161484453, + "loss": 1.3267, + "theoretical_loss": 3.5178497070203845, + "tokens_seen": 1636630528 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002546138415245737, + "loss": 1.2922, + "theoretical_loss": 3.5178375928213192, + "tokens_seen": 1636696064 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002546038114343029, + "loss": 1.3733, + "theoretical_loss": 3.5178254792431307, + "tokens_seen": 1636761600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545937813440321, + "loss": 1.4751, + "theoretical_loss": 3.517813366285762, + "tokens_seen": 1636827136 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025458375125376125, + "loss": 1.257, + "theoretical_loss": 3.517801253949157, + "tokens_seen": 1636892672 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545737211634905, + "loss": 1.4478, + "theoretical_loss": 3.5177891422332594, + "tokens_seen": 1636958208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545636910732196, + "loss": 1.4338, + "theoretical_loss": 3.5177770311380114, + "tokens_seen": 1637023744 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025455366098294885, + "loss": 1.3955, + "theoretical_loss": 3.5177649206633568, + "tokens_seen": 1637089280 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025454363089267803, + "loss": 1.3131, + "theoretical_loss": 3.5177528108092395, + "tokens_seen": 1637154816 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545336008024072, + "loss": 1.3031, + "theoretical_loss": 3.517740701575602, + "tokens_seen": 1637220352 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545235707121364, + "loss": 1.3262, + "theoretical_loss": 3.5177285929623885, + "tokens_seen": 1637285888 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025451354062186563, + "loss": 1.2694, + "theoretical_loss": 3.5177164849695415, + "tokens_seen": 1637351424 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002545035105315948, + "loss": 1.3387, + "theoretical_loss": 3.5177043775970054, + "tokens_seen": 1637416960 + }, + { + "epoch": 0.5, + "learning_rate": 0.000254493480441324, + "loss": 1.3455, + "theoretical_loss": 3.517692270844722, + "tokens_seen": 1637482496 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025448345035105317, + "loss": 1.2939, + "theoretical_loss": 3.517680164712637, + "tokens_seen": 1637548032 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025447342026078235, + "loss": 1.3286, + "theoretical_loss": 3.517668059200692, + "tokens_seen": 1637613568 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544633901705116, + "loss": 1.3476, + "theoretical_loss": 3.5176559543088306, + "tokens_seen": 1637679104 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544533600802407, + "loss": 1.2079, + "theoretical_loss": 3.517643850036997, + "tokens_seen": 1637744640 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025444332998996995, + "loss": 1.3564, + "theoretical_loss": 3.5176317463851343, + "tokens_seen": 1637810176 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544332998996991, + "loss": 1.4992, + "theoretical_loss": 3.5176196433531857, + "tokens_seen": 1637875712 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544232698094283, + "loss": 1.4364, + "theoretical_loss": 3.517607540941095, + "tokens_seen": 1637941248 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544132397191575, + "loss": 1.3299, + "theoretical_loss": 3.5175954391488053, + "tokens_seen": 1638006784 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002544032096288867, + "loss": 1.332, + "theoretical_loss": 3.5175833379762604, + "tokens_seen": 1638072320 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025439317953861586, + "loss": 1.4762, + "theoretical_loss": 3.5175712374234034, + "tokens_seen": 1638137856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025438314944834504, + "loss": 1.4519, + "theoretical_loss": 3.5175591374901782, + "tokens_seen": 1638203392 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002543731193580742, + "loss": 1.3655, + "theoretical_loss": 3.517547038176528, + "tokens_seen": 1638268928 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025436308926780345, + "loss": 1.4029, + "theoretical_loss": 3.5175349394823963, + "tokens_seen": 1638334464 + }, + { + "debugging/distinct-1-grams": 0.7474626180525646, + "debugging/entropy-1-grams": 5.599013278347353, + "debugging/length": 499.42857142857144, + "debugging/num_segments": 14, + "debugging/score": 0.0110880360129877, + "debugging/score_std": 0.01573414919444883, + "epoch": 0.5, + "objective/train/docs_used": 598618, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3606061935424805, + "objective/train/theoretical_loss": 3.5175228414077266, + "objective/train/tokens_used": 1658860000, + "theoretical_loss": 3.5175228414077266, + "tokens_seen": 1638400000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002543530591775326, + "loss": 1.296, + "theoretical_loss": 3.5175228414077266, + "tokens_seen": 1638400000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002543430290872618, + "loss": 1.4068, + "theoretical_loss": 3.517510743952463, + "tokens_seen": 1638465536 + }, + { + "epoch": 0.5, + "learning_rate": 0.000254332998996991, + "loss": 1.4276, + "theoretical_loss": 3.5174986471165477, + "tokens_seen": 1638531072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002543229689067202, + "loss": 1.4629, + "theoretical_loss": 3.5174865508999256, + "tokens_seen": 1638596608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025431293881644936, + "loss": 1.4731, + "theoretical_loss": 3.5174744553025388, + "tokens_seen": 1638662144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025430290872617854, + "loss": 1.4282, + "theoretical_loss": 3.5174623603243322, + "tokens_seen": 1638727680 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002542928786359077, + "loss": 1.3843, + "theoretical_loss": 3.517450265965249, + "tokens_seen": 1638793216 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025428284854563696, + "loss": 1.1585, + "theoretical_loss": 3.5174381722252326, + "tokens_seen": 1638858752 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002542728184553661, + "loss": 1.3726, + "theoretical_loss": 3.517426079104226, + "tokens_seen": 1638924288 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002542627883650953, + "loss": 1.3709, + "theoretical_loss": 3.5174139866021736, + "tokens_seen": 1638989824 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025425275827482445, + "loss": 1.3868, + "theoretical_loss": 3.5174018947190193, + "tokens_seen": 1639055360 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002542427281845537, + "loss": 1.3915, + "theoretical_loss": 3.517389803454705, + "tokens_seen": 1639120896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025423269809428286, + "loss": 1.4206, + "theoretical_loss": 3.517377712809176, + "tokens_seen": 1639186432 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025422266800401204, + "loss": 1.3382, + "theoretical_loss": 3.517365622782375, + "tokens_seen": 1639251968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002542126379137412, + "loss": 1.2389, + "theoretical_loss": 3.5173535333742456, + "tokens_seen": 1639317504 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025420260782347046, + "loss": 1.4382, + "theoretical_loss": 3.5173414445847317, + "tokens_seen": 1639383040 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541925777331996, + "loss": 1.3971, + "theoretical_loss": 3.517329356413777, + "tokens_seen": 1639448576 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541825476429288, + "loss": 1.2898, + "theoretical_loss": 3.5173172688613255, + "tokens_seen": 1639514112 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025417251755265795, + "loss": 1.3396, + "theoretical_loss": 3.5173051819273193, + "tokens_seen": 1639579648 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541624874623872, + "loss": 1.3017, + "theoretical_loss": 3.517293095611704, + "tokens_seen": 1639645184 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541624874623872, + "loss": 1.4377, + "theoretical_loss": 3.5172810099144214, + "tokens_seen": 1639710720 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025415245737211637, + "loss": 1.3202, + "theoretical_loss": 3.5172689248354168, + "tokens_seen": 1639776256 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025414242728184555, + "loss": 1.3919, + "theoretical_loss": 3.5172568403746327, + "tokens_seen": 1639841792 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025413239719157473, + "loss": 1.3689, + "theoretical_loss": 3.517244756532013, + "tokens_seen": 1639907328 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541223671013039, + "loss": 1.5301, + "theoretical_loss": 3.517232673307502, + "tokens_seen": 1639972864 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541123370110331, + "loss": 1.3197, + "theoretical_loss": 3.5172205907010428, + "tokens_seen": 1640038400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002541023069207623, + "loss": 1.4426, + "theoretical_loss": 3.517208508712579, + "tokens_seen": 1640103936 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025409227683049145, + "loss": 1.3369, + "theoretical_loss": 3.5171964273420544, + "tokens_seen": 1640169472 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002540822467402207, + "loss": 1.39, + "theoretical_loss": 3.5171843465894135, + "tokens_seen": 1640235008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002540722166499498, + "loss": 1.3441, + "theoretical_loss": 3.5171722664545992, + "tokens_seen": 1640300544 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025406218655967905, + "loss": 1.2375, + "theoretical_loss": 3.517160186937555, + "tokens_seen": 1640366080 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025405215646940823, + "loss": 1.4008, + "theoretical_loss": 3.517148108038225, + "tokens_seen": 1640431616 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002540421263791374, + "loss": 1.2988, + "theoretical_loss": 3.5171360297565535, + "tokens_seen": 1640497152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002540320962888666, + "loss": 1.3682, + "theoretical_loss": 3.517123952092483, + "tokens_seen": 1640562688 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025402206619859583, + "loss": 1.1694, + "theoretical_loss": 3.5171118750459582, + "tokens_seen": 1640628224 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025401203610832495, + "loss": 1.3061, + "theoretical_loss": 3.517099798616923, + "tokens_seen": 1640693760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002540020060180542, + "loss": 1.442, + "theoretical_loss": 3.51708772280532, + "tokens_seen": 1640759296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539919759277833, + "loss": 1.2753, + "theoretical_loss": 3.5170756476110943, + "tokens_seen": 1640824832 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025398194583751255, + "loss": 1.3274, + "theoretical_loss": 3.517063573034189, + "tokens_seen": 1640890368 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025397191574724173, + "loss": 1.3999, + "theoretical_loss": 3.517051499074548, + "tokens_seen": 1640955904 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539618856569709, + "loss": 1.441, + "theoretical_loss": 3.517039425732115, + "tokens_seen": 1641021440 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539518555667001, + "loss": 1.3162, + "theoretical_loss": 3.5170273530068337, + "tokens_seen": 1641086976 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539418254764293, + "loss": 1.3672, + "theoretical_loss": 3.517015280898648, + "tokens_seen": 1641152512 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025393179538615846, + "loss": 1.3476, + "theoretical_loss": 3.5170032094075023, + "tokens_seen": 1641218048 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539217652958877, + "loss": 1.3551, + "theoretical_loss": 3.51699113853334, + "tokens_seen": 1641283584 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002539117352056168, + "loss": 1.3667, + "theoretical_loss": 3.516979068276105, + "tokens_seen": 1641349120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025390170511534606, + "loss": 1.2635, + "theoretical_loss": 3.5169669986357404, + "tokens_seen": 1641414656 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538916750250752, + "loss": 1.5488, + "theoretical_loss": 3.516954929612191, + "tokens_seen": 1641480192 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538816449348044, + "loss": 1.3676, + "theoretical_loss": 3.5169428612053997, + "tokens_seen": 1641545728 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538716148445336, + "loss": 1.4331, + "theoretical_loss": 3.5169307934153116, + "tokens_seen": 1641611264 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 599774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3520958423614502, + "objective/train/theoretical_loss": 3.5169187262418697, + "objective/train/tokens_used": 1662136800, + "theoretical_loss": 3.5169187262418697, + "tokens_seen": 1641676800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538615847542628, + "loss": 1.3422, + "theoretical_loss": 3.5169187262418697, + "tokens_seen": 1641676800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025385155466399196, + "loss": 1.3939, + "theoretical_loss": 3.5169066596850183, + "tokens_seen": 1641742336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538415245737212, + "loss": 1.4172, + "theoretical_loss": 3.516894593744701, + "tokens_seen": 1641807872 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538314944834503, + "loss": 1.4046, + "theoretical_loss": 3.516882528420862, + "tokens_seen": 1641873408 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025382146439317956, + "loss": 1.4043, + "theoretical_loss": 3.5168704637134445, + "tokens_seen": 1641938944 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538114343029087, + "loss": 1.46, + "theoretical_loss": 3.5168583996223934, + "tokens_seen": 1642004480 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002538014042126379, + "loss": 1.412, + "theoretical_loss": 3.516846336147652, + "tokens_seen": 1642070016 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002537913741223671, + "loss": 1.2078, + "theoretical_loss": 3.516834273289164, + "tokens_seen": 1642135552 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002537813440320963, + "loss": 1.348, + "theoretical_loss": 3.516822211046874, + "tokens_seen": 1642201088 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025377131394182546, + "loss": 1.3635, + "theoretical_loss": 3.516810149420725, + "tokens_seen": 1642266624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025376128385155465, + "loss": 1.4329, + "theoretical_loss": 3.5167980884106624, + "tokens_seen": 1642332160 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002537512537612839, + "loss": 1.3251, + "theoretical_loss": 3.5167860280166288, + "tokens_seen": 1642397696 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025374122367101306, + "loss": 1.4018, + "theoretical_loss": 3.516773968238569, + "tokens_seen": 1642463232 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025373119358074224, + "loss": 1.3855, + "theoretical_loss": 3.5167619090764264, + "tokens_seen": 1642528768 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002537211634904714, + "loss": 1.3843, + "theoretical_loss": 3.516749850530145, + "tokens_seen": 1642594304 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025371113340020066, + "loss": 1.3666, + "theoretical_loss": 3.5167377925996686, + "tokens_seen": 1642659840 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002537011033099298, + "loss": 1.4673, + "theoretical_loss": 3.516725735284942, + "tokens_seen": 1642725376 + }, + { + "epoch": 0.5, + "learning_rate": 0.000253691073219659, + "loss": 1.3673, + "theoretical_loss": 3.5167136785859086, + "tokens_seen": 1642790912 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025368104312938815, + "loss": 1.4395, + "theoretical_loss": 3.5167016225025125, + "tokens_seen": 1642856448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002536710130391174, + "loss": 1.3607, + "theoretical_loss": 3.516689567034698, + "tokens_seen": 1642921984 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025366098294884657, + "loss": 1.4124, + "theoretical_loss": 3.516677512182408, + "tokens_seen": 1642987520 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025365095285857575, + "loss": 1.4309, + "theoretical_loss": 3.5166654579455887, + "tokens_seen": 1643053056 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025364092276830493, + "loss": 1.4703, + "theoretical_loss": 3.516653404324182, + "tokens_seen": 1643118592 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002536308926780341, + "loss": 1.3122, + "theoretical_loss": 3.5166413513181327, + "tokens_seen": 1643184128 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002536208625877633, + "loss": 1.3444, + "theoretical_loss": 3.516629298927385, + "tokens_seen": 1643249664 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002536108324974925, + "loss": 1.4106, + "theoretical_loss": 3.516617247151883, + "tokens_seen": 1643315200 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025360080240722165, + "loss": 1.4384, + "theoretical_loss": 3.5166051959915703, + "tokens_seen": 1643380736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002535907723169509, + "loss": 1.4408, + "theoretical_loss": 3.5165931454463912, + "tokens_seen": 1643446272 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025358074222668, + "loss": 1.2795, + "theoretical_loss": 3.5165810955162904, + "tokens_seen": 1643511808 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025357071213640925, + "loss": 1.3702, + "theoretical_loss": 3.516569046201211, + "tokens_seen": 1643577344 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025356068204613843, + "loss": 1.4155, + "theoretical_loss": 3.5165569975010973, + "tokens_seen": 1643642880 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002535506519558676, + "loss": 1.3802, + "theoretical_loss": 3.516544949415894, + "tokens_seen": 1643708416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002535406218655968, + "loss": 1.2808, + "theoretical_loss": 3.5165329019455447, + "tokens_seen": 1643773952 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025353059177532603, + "loss": 1.3783, + "theoretical_loss": 3.5165208550899933, + "tokens_seen": 1643839488 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025352056168505516, + "loss": 1.3021, + "theoretical_loss": 3.516508808849184, + "tokens_seen": 1643905024 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002535105315947844, + "loss": 1.3818, + "theoretical_loss": 3.5164967632230617, + "tokens_seen": 1643970560 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002535005015045135, + "loss": 1.3134, + "theoretical_loss": 3.51648471821157, + "tokens_seen": 1644036096 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025349047141424275, + "loss": 1.3111, + "theoretical_loss": 3.5164726738146532, + "tokens_seen": 1644101632 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025348044132397193, + "loss": 1.2526, + "theoretical_loss": 3.5164606300322547, + "tokens_seen": 1644167168 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002534704112337011, + "loss": 1.3507, + "theoretical_loss": 3.5164485868643194, + "tokens_seen": 1644232704 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002534603811434303, + "loss": 1.3627, + "theoretical_loss": 3.5164365443107917, + "tokens_seen": 1644298240 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002534503510531595, + "loss": 1.4008, + "theoretical_loss": 3.5164245023716147, + "tokens_seen": 1644363776 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025344032096288866, + "loss": 1.3735, + "theoretical_loss": 3.5164124610467336, + "tokens_seen": 1644429312 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002534302908726179, + "loss": 1.4033, + "theoretical_loss": 3.516400420336092, + "tokens_seen": 1644494848 + }, + { + "epoch": 0.5, + "learning_rate": 0.000253420260782347, + "loss": 1.3908, + "theoretical_loss": 3.5163883802396345, + "tokens_seen": 1644560384 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025341023069207626, + "loss": 1.4353, + "theoretical_loss": 3.516376340757305, + "tokens_seen": 1644625920 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002534002006018054, + "loss": 1.3231, + "theoretical_loss": 3.516364301889048, + "tokens_seen": 1644691456 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533901705115346, + "loss": 1.5011, + "theoretical_loss": 3.516352263634807, + "tokens_seen": 1644756992 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533801404212638, + "loss": 1.4089, + "theoretical_loss": 3.516340225994527, + "tokens_seen": 1644822528 + }, + { + "epoch": 0.5, + "learning_rate": 0.000253370110330993, + "loss": 1.4204, + "theoretical_loss": 3.516328188968152, + "tokens_seen": 1644888064 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 601010, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.395542860031128, + "objective/train/theoretical_loss": 3.5163161525556266, + "objective/train/tokens_used": 1665413600, + "theoretical_loss": 3.5163161525556266, + "tokens_seen": 1644953600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025336008024072216, + "loss": 1.3342, + "theoretical_loss": 3.5163161525556266, + "tokens_seen": 1644953600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533500501504514, + "loss": 1.3589, + "theoretical_loss": 3.516304116756894, + "tokens_seen": 1645019136 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533400200601805, + "loss": 1.3123, + "theoretical_loss": 3.5162920815718994, + "tokens_seen": 1645084672 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025332998996990976, + "loss": 1.4397, + "theoretical_loss": 3.5162800470005866, + "tokens_seen": 1645150208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533199598796389, + "loss": 1.4968, + "theoretical_loss": 3.5162680130429003, + "tokens_seen": 1645215744 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002533099297893681, + "loss": 1.4416, + "theoretical_loss": 3.516255979698784, + "tokens_seen": 1645281280 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002532998996990973, + "loss": 1.3567, + "theoretical_loss": 3.5162439469681828, + "tokens_seen": 1645346816 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002532898696088265, + "loss": 1.4532, + "theoretical_loss": 3.5162319148510406, + "tokens_seen": 1645412352 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025327983951855566, + "loss": 1.4021, + "theoretical_loss": 3.5162198833473015, + "tokens_seen": 1645477888 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025326980942828485, + "loss": 1.4394, + "theoretical_loss": 3.51620785245691, + "tokens_seen": 1645543424 + }, + { + "epoch": 0.5, + "learning_rate": 0.000253259779338014, + "loss": 1.428, + "theoretical_loss": 3.5161958221798106, + "tokens_seen": 1645608960 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025324974924774326, + "loss": 1.2957, + "theoretical_loss": 3.516183792515948, + "tokens_seen": 1645674496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002532397191574724, + "loss": 1.3399, + "theoretical_loss": 3.516171763465265, + "tokens_seen": 1645740032 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002532296890672016, + "loss": 1.4079, + "theoretical_loss": 3.516159735027707, + "tokens_seen": 1645805568 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025321965897693075, + "loss": 1.3975, + "theoretical_loss": 3.516147707203219, + "tokens_seen": 1645871104 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025320962888666, + "loss": 1.3674, + "theoretical_loss": 3.516135679991744, + "tokens_seen": 1645936640 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025319959879638917, + "loss": 1.2961, + "theoretical_loss": 3.516123653393227, + "tokens_seen": 1646002176 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025318956870611835, + "loss": 1.4296, + "theoretical_loss": 3.5161116274076125, + "tokens_seen": 1646067712 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025317953861584753, + "loss": 1.4888, + "theoretical_loss": 3.5160996020348447, + "tokens_seen": 1646133248 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025316950852557677, + "loss": 1.382, + "theoretical_loss": 3.5160875772748676, + "tokens_seen": 1646198784 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002531594784353059, + "loss": 1.4412, + "theoretical_loss": 3.516075553127626, + "tokens_seen": 1646264320 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025314944834503513, + "loss": 1.274, + "theoretical_loss": 3.5160635295930645, + "tokens_seen": 1646329856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025313941825476425, + "loss": 1.3243, + "theoretical_loss": 3.5160515066711264, + "tokens_seen": 1646395392 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002531293881644935, + "loss": 1.4155, + "theoretical_loss": 3.516039484361758, + "tokens_seen": 1646460928 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025311935807422267, + "loss": 1.4047, + "theoretical_loss": 3.5160274626649017, + "tokens_seen": 1646526464 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025310932798395185, + "loss": 1.3647, + "theoretical_loss": 3.516015441580503, + "tokens_seen": 1646592000 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025309929789368103, + "loss": 1.3244, + "theoretical_loss": 3.5160034211085067, + "tokens_seen": 1646657536 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002530892678034102, + "loss": 1.4928, + "theoretical_loss": 3.515991401248856, + "tokens_seen": 1646723072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002530792377131394, + "loss": 1.2711, + "theoretical_loss": 3.515979382001496, + "tokens_seen": 1646788608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025306920762286863, + "loss": 1.3735, + "theoretical_loss": 3.5159673633663715, + "tokens_seen": 1646854144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025305917753259776, + "loss": 1.2941, + "theoretical_loss": 3.5159553453434267, + "tokens_seen": 1646919680 + }, + { + "epoch": 0.5, + "learning_rate": 0.000253049147442327, + "loss": 1.3517, + "theoretical_loss": 3.5159433279326056, + "tokens_seen": 1646985216 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002530391173520562, + "loss": 1.3232, + "theoretical_loss": 3.515931311133853, + "tokens_seen": 1647050752 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025302908726178536, + "loss": 1.3617, + "theoretical_loss": 3.515919294947114, + "tokens_seen": 1647116288 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025301905717151454, + "loss": 1.2661, + "theoretical_loss": 3.515907279372332, + "tokens_seen": 1647181824 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002530090270812437, + "loss": 1.3583, + "theoretical_loss": 3.5158952644094517, + "tokens_seen": 1647247360 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025299899699097295, + "loss": 1.3534, + "theoretical_loss": 3.515883250058418, + "tokens_seen": 1647312896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025298896690070213, + "loss": 1.3219, + "theoretical_loss": 3.5158712363191755, + "tokens_seen": 1647378432 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529789368104313, + "loss": 1.3865, + "theoretical_loss": 3.5158592231916685, + "tokens_seen": 1647443968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529689067201605, + "loss": 1.4216, + "theoretical_loss": 3.5158472106758416, + "tokens_seen": 1647509504 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529588766298897, + "loss": 1.3797, + "theoretical_loss": 3.515835198771639, + "tokens_seen": 1647575040 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025294884653961886, + "loss": 1.4127, + "theoretical_loss": 3.5158231874790054, + "tokens_seen": 1647640576 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529388164493481, + "loss": 1.2673, + "theoretical_loss": 3.5158111767978855, + "tokens_seen": 1647706112 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529287863590772, + "loss": 1.4363, + "theoretical_loss": 3.5157991667282236, + "tokens_seen": 1647771648 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025291875626880646, + "loss": 1.2617, + "theoretical_loss": 3.5157871572699646, + "tokens_seen": 1647837184 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002529087261785356, + "loss": 1.334, + "theoretical_loss": 3.515775148423053, + "tokens_seen": 1647902720 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528986960882648, + "loss": 1.3805, + "theoretical_loss": 3.515763140187433, + "tokens_seen": 1647968256 + }, + { + "epoch": 0.5, + "learning_rate": 0.000252888665997994, + "loss": 1.3532, + "theoretical_loss": 3.5157511325630493, + "tokens_seen": 1648033792 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528786359077232, + "loss": 1.3943, + "theoretical_loss": 3.5157391255498465, + "tokens_seen": 1648099328 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025286860581745236, + "loss": 1.3592, + "theoretical_loss": 3.5157271191477695, + "tokens_seen": 1648164864 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 602144, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4235153198242188, + "objective/train/theoretical_loss": 3.5157151133567623, + "objective/train/tokens_used": 1668690400, + "theoretical_loss": 3.5157151133567623, + "tokens_seen": 1648230400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528585757271816, + "loss": 1.3922, + "theoretical_loss": 3.5157151133567623, + "tokens_seen": 1648230400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528485456369107, + "loss": 1.4969, + "theoretical_loss": 3.5157031081767705, + "tokens_seen": 1648295936 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025283851554663996, + "loss": 1.3341, + "theoretical_loss": 3.515691103607738, + "tokens_seen": 1648361472 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528284854563691, + "loss": 1.4048, + "theoretical_loss": 3.515679099649609, + "tokens_seen": 1648427008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528184553660983, + "loss": 1.4063, + "theoretical_loss": 3.515667096302329, + "tokens_seen": 1648492544 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002528084252758275, + "loss": 1.332, + "theoretical_loss": 3.515655093565842, + "tokens_seen": 1648558080 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002527983951855567, + "loss": 1.4406, + "theoretical_loss": 3.5156430914400927, + "tokens_seen": 1648623616 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025278836509528586, + "loss": 1.4612, + "theoretical_loss": 3.5156310899250265, + "tokens_seen": 1648689152 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025277833500501505, + "loss": 1.3856, + "theoretical_loss": 3.5156190890205874, + "tokens_seen": 1648754688 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002527683049147442, + "loss": 1.2911, + "theoretical_loss": 3.5156070887267195, + "tokens_seen": 1648820224 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025275827482447346, + "loss": 1.3495, + "theoretical_loss": 3.515595089043369, + "tokens_seen": 1648885760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002527482447342026, + "loss": 1.2585, + "theoretical_loss": 3.5155830899704794, + "tokens_seen": 1648951296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002527382146439318, + "loss": 1.3679, + "theoretical_loss": 3.5155710915079954, + "tokens_seen": 1649016832 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025272818455366095, + "loss": 1.2711, + "theoretical_loss": 3.5155590936558623, + "tokens_seen": 1649082368 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002527181544633902, + "loss": 1.4442, + "theoretical_loss": 3.5155470964140245, + "tokens_seen": 1649147904 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025270812437311937, + "loss": 1.3219, + "theoretical_loss": 3.5155350997824266, + "tokens_seen": 1649213440 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025269809428284855, + "loss": 1.3308, + "theoretical_loss": 3.5155231037610135, + "tokens_seen": 1649278976 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025268806419257773, + "loss": 1.4121, + "theoretical_loss": 3.5155111083497297, + "tokens_seen": 1649344512 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025267803410230697, + "loss": 1.2819, + "theoretical_loss": 3.51549911354852, + "tokens_seen": 1649410048 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002526680040120361, + "loss": 1.4635, + "theoretical_loss": 3.5154871193573296, + "tokens_seen": 1649475584 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025265797392176533, + "loss": 1.2846, + "theoretical_loss": 3.5154751257761023, + "tokens_seen": 1649541120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025264794383149445, + "loss": 1.4292, + "theoretical_loss": 3.5154631328047836, + "tokens_seen": 1649606656 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002526379137412237, + "loss": 1.4185, + "theoretical_loss": 3.515451140443318, + "tokens_seen": 1649672192 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025262788365095287, + "loss": 1.4316, + "theoretical_loss": 3.51543914869165, + "tokens_seen": 1649737728 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025261785356068205, + "loss": 1.3704, + "theoretical_loss": 3.515427157549725, + "tokens_seen": 1649803264 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025260782347041123, + "loss": 1.3161, + "theoretical_loss": 3.5154151670174874, + "tokens_seen": 1649868800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525977933801404, + "loss": 1.3746, + "theoretical_loss": 3.515403177094882, + "tokens_seen": 1649934336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525877632898696, + "loss": 1.3378, + "theoretical_loss": 3.5153911877818533, + "tokens_seen": 1649999872 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025257773319959883, + "loss": 1.4057, + "theoretical_loss": 3.515379199078347, + "tokens_seen": 1650065408 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025256770310932796, + "loss": 1.3709, + "theoretical_loss": 3.515367210984307, + "tokens_seen": 1650130944 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525576730190572, + "loss": 1.2382, + "theoretical_loss": 3.515355223499678, + "tokens_seen": 1650196480 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525476429287864, + "loss": 1.3341, + "theoretical_loss": 3.515343236624406, + "tokens_seen": 1650262016 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025253761283851556, + "loss": 1.4548, + "theoretical_loss": 3.5153312503584346, + "tokens_seen": 1650327552 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025252758274824474, + "loss": 1.4804, + "theoretical_loss": 3.5153192647017093, + "tokens_seen": 1650393088 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525175526579739, + "loss": 1.39, + "theoretical_loss": 3.5153072796541744, + "tokens_seen": 1650458624 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002525075225677031, + "loss": 1.3435, + "theoretical_loss": 3.5152952952157754, + "tokens_seen": 1650524160 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025249749247743233, + "loss": 1.3937, + "theoretical_loss": 3.515283311386457, + "tokens_seen": 1650589696 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025248746238716146, + "loss": 1.4215, + "theoretical_loss": 3.515271328166164, + "tokens_seen": 1650655232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002524774322968907, + "loss": 1.3711, + "theoretical_loss": 3.515259345554841, + "tokens_seen": 1650720768 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002524674022066198, + "loss": 1.3701, + "theoretical_loss": 3.515247363552433, + "tokens_seen": 1650786304 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025245737211634906, + "loss": 1.3916, + "theoretical_loss": 3.515235382158885, + "tokens_seen": 1650851840 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025244734202607824, + "loss": 1.3442, + "theoretical_loss": 3.5152234013741417, + "tokens_seen": 1650917376 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002524373119358074, + "loss": 1.3067, + "theoretical_loss": 3.5152114211981487, + "tokens_seen": 1650982912 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002524272818455366, + "loss": 1.4334, + "theoretical_loss": 3.51519944163085, + "tokens_seen": 1651048448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002524172517552658, + "loss": 1.2827, + "theoretical_loss": 3.5151874626721904, + "tokens_seen": 1651113984 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025240722166499496, + "loss": 1.3895, + "theoretical_loss": 3.515175484322116, + "tokens_seen": 1651179520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523971915747242, + "loss": 1.4036, + "theoretical_loss": 3.5151635065805706, + "tokens_seen": 1651245056 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523871614844533, + "loss": 1.2742, + "theoretical_loss": 3.5151515294474995, + "tokens_seen": 1651310592 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025237713139418256, + "loss": 1.2462, + "theoretical_loss": 3.515139552922848, + "tokens_seen": 1651376128 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025236710130391174, + "loss": 1.3584, + "theoretical_loss": 3.5151275770065604, + "tokens_seen": 1651441664 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 603335, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2819124460220337, + "objective/train/theoretical_loss": 3.5151156016985823, + "objective/train/tokens_used": 1671967200, + "theoretical_loss": 3.5151156016985823, + "tokens_seen": 1651507200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523570712136409, + "loss": 1.3177, + "theoretical_loss": 3.5151156016985823, + "tokens_seen": 1651507200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523470411233701, + "loss": 1.4093, + "theoretical_loss": 3.5151036269988585, + "tokens_seen": 1651572736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523370110330993, + "loss": 1.4945, + "theoretical_loss": 3.5150916529073335, + "tokens_seen": 1651638272 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025232698094282847, + "loss": 1.3489, + "theoretical_loss": 3.5150796794239527, + "tokens_seen": 1651703808 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002523169508525577, + "loss": 1.3466, + "theoretical_loss": 3.5150677065486615, + "tokens_seen": 1651769344 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025230692076228683, + "loss": 1.4723, + "theoretical_loss": 3.515055734281404, + "tokens_seen": 1651834880 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025229689067201606, + "loss": 1.3533, + "theoretical_loss": 3.515043762622125, + "tokens_seen": 1651900416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002522868605817452, + "loss": 1.3139, + "theoretical_loss": 3.5150317915707707, + "tokens_seen": 1651965952 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025227683049147443, + "loss": 1.3009, + "theoretical_loss": 3.5150198211272854, + "tokens_seen": 1652031488 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002522668004012036, + "loss": 1.4133, + "theoretical_loss": 3.5150078512916143, + "tokens_seen": 1652097024 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002522567703109328, + "loss": 1.3471, + "theoretical_loss": 3.514995882063702, + "tokens_seen": 1652162560 + }, + { + "epoch": 0.5, + "learning_rate": 0.000252246740220662, + "loss": 1.3425, + "theoretical_loss": 3.5149839134434946, + "tokens_seen": 1652228096 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025223671013039115, + "loss": 1.2818, + "theoretical_loss": 3.5149719454309363, + "tokens_seen": 1652293632 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002522266800401204, + "loss": 1.4075, + "theoretical_loss": 3.5149599780259715, + "tokens_seen": 1652359168 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025221664994984957, + "loss": 1.5639, + "theoretical_loss": 3.514948011228547, + "tokens_seen": 1652424704 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025220661985957875, + "loss": 1.3753, + "theoretical_loss": 3.5149360450386062, + "tokens_seen": 1652490240 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025219658976930793, + "loss": 1.4621, + "theoretical_loss": 3.514924079456095, + "tokens_seen": 1652555776 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025218655967903717, + "loss": 1.5157, + "theoretical_loss": 3.5149121144809587, + "tokens_seen": 1652621312 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002521765295887663, + "loss": 1.3468, + "theoretical_loss": 3.5149001501131423, + "tokens_seen": 1652686848 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025216649949849553, + "loss": 1.3391, + "theoretical_loss": 3.5148881863525903, + "tokens_seen": 1652752384 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025215646940822465, + "loss": 1.3836, + "theoretical_loss": 3.5148762231992476, + "tokens_seen": 1652817920 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002521464393179539, + "loss": 1.4059, + "theoretical_loss": 3.5148642606530602, + "tokens_seen": 1652883456 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025213640922768307, + "loss": 1.3906, + "theoretical_loss": 3.514852298713973, + "tokens_seen": 1652948992 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025212637913741225, + "loss": 1.3577, + "theoretical_loss": 3.514840337381931, + "tokens_seen": 1653014528 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025211634904714143, + "loss": 1.4132, + "theoretical_loss": 3.5148283766568795, + "tokens_seen": 1653080064 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002521063189568706, + "loss": 1.4161, + "theoretical_loss": 3.514816416538763, + "tokens_seen": 1653145600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002520962888665998, + "loss": 1.318, + "theoretical_loss": 3.5148044570275276, + "tokens_seen": 1653211136 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025208625877632903, + "loss": 1.3779, + "theoretical_loss": 3.5147924981231173, + "tokens_seen": 1653276672 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025207622868605816, + "loss": 1.2997, + "theoretical_loss": 3.5147805398254786, + "tokens_seen": 1653342208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002520661985957874, + "loss": 1.2972, + "theoretical_loss": 3.514768582134555, + "tokens_seen": 1653407744 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002520561685055166, + "loss": 1.3538, + "theoretical_loss": 3.5147566250502935, + "tokens_seen": 1653473280 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025204613841524576, + "loss": 1.3626, + "theoretical_loss": 3.514744668572638, + "tokens_seen": 1653538816 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025203610832497494, + "loss": 1.4505, + "theoretical_loss": 3.514732712701534, + "tokens_seen": 1653604352 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002520260782347041, + "loss": 1.2681, + "theoretical_loss": 3.514720757436927, + "tokens_seen": 1653669888 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002520160481444333, + "loss": 1.2899, + "theoretical_loss": 3.5147088027787623, + "tokens_seen": 1653735424 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025200601805416253, + "loss": 1.4876, + "theoretical_loss": 3.5146968487269845, + "tokens_seen": 1653800960 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025199598796389166, + "loss": 1.3989, + "theoretical_loss": 3.5146848952815386, + "tokens_seen": 1653866496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002519859578736209, + "loss": 1.3689, + "theoretical_loss": 3.5146729424423713, + "tokens_seen": 1653932032 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025197592778335, + "loss": 1.4696, + "theoretical_loss": 3.514660990209426, + "tokens_seen": 1653997568 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025196589769307926, + "loss": 1.3786, + "theoretical_loss": 3.5146490385826494, + "tokens_seen": 1654063104 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025195586760280844, + "loss": 1.3127, + "theoretical_loss": 3.5146370875619857, + "tokens_seen": 1654128640 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002519458375125376, + "loss": 1.1616, + "theoretical_loss": 3.5146251371473807, + "tokens_seen": 1654194176 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002519358074222668, + "loss": 1.4035, + "theoretical_loss": 3.51461318733878, + "tokens_seen": 1654259712 + }, + { + "epoch": 0.5, + "learning_rate": 0.000251925777331996, + "loss": 1.4002, + "theoretical_loss": 3.5146012381361276, + "tokens_seen": 1654325248 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025191574724172516, + "loss": 1.4077, + "theoretical_loss": 3.51458928953937, + "tokens_seen": 1654390784 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002519057171514544, + "loss": 1.3067, + "theoretical_loss": 3.514577341548452, + "tokens_seen": 1654456320 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002518956870611835, + "loss": 1.2714, + "theoretical_loss": 3.5145653941633186, + "tokens_seen": 1654521856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025188565697091276, + "loss": 1.3757, + "theoretical_loss": 3.514553447383916, + "tokens_seen": 1654587392 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025187562688064194, + "loss": 1.3883, + "theoretical_loss": 3.5145415012101884, + "tokens_seen": 1654652928 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002518655967903711, + "loss": 1.403, + "theoretical_loss": 3.514529555642082, + "tokens_seen": 1654718464 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 604403, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4604017734527588, + "objective/train/theoretical_loss": 3.5145176106795417, + "objective/train/tokens_used": 1675244000, + "theoretical_loss": 3.5145176106795417, + "tokens_seen": 1654784000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002518555667001003, + "loss": 1.4569, + "theoretical_loss": 3.5145176106795417, + "tokens_seen": 1654784000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002518455366098295, + "loss": 1.4665, + "theoretical_loss": 3.514505666322513, + "tokens_seen": 1654849536 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025183550651955867, + "loss": 1.3361, + "theoretical_loss": 3.5144937225709407, + "tokens_seen": 1654915072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002518254764292879, + "loss": 1.3591, + "theoretical_loss": 3.514481779424771, + "tokens_seen": 1654980608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025181544633901703, + "loss": 1.2791, + "theoretical_loss": 3.5144698368839484, + "tokens_seen": 1655046144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025180541624874627, + "loss": 1.3223, + "theoretical_loss": 3.5144578949484186, + "tokens_seen": 1655111680 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002517953861584754, + "loss": 1.371, + "theoretical_loss": 3.514445953618127, + "tokens_seen": 1655177216 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025178535606820463, + "loss": 1.3334, + "theoretical_loss": 3.5144340128930196, + "tokens_seen": 1655242752 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002517753259779338, + "loss": 1.381, + "theoretical_loss": 3.51442207277304, + "tokens_seen": 1655308288 + }, + { + "epoch": 0.5, + "learning_rate": 0.000251765295887663, + "loss": 1.3274, + "theoretical_loss": 3.5144101332581354, + "tokens_seen": 1655373824 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025175526579739217, + "loss": 1.3057, + "theoretical_loss": 3.5143981943482503, + "tokens_seen": 1655439360 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025174523570712135, + "loss": 1.3832, + "theoretical_loss": 3.5143862560433305, + "tokens_seen": 1655504896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025173520561685053, + "loss": 1.4412, + "theoretical_loss": 3.5143743183433207, + "tokens_seen": 1655570432 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025172517552657977, + "loss": 1.3094, + "theoretical_loss": 3.514362381248167, + "tokens_seen": 1655635968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002517151454363089, + "loss": 1.4969, + "theoretical_loss": 3.514350444757815, + "tokens_seen": 1655701504 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025170511534603813, + "loss": 1.3395, + "theoretical_loss": 3.5143385088722088, + "tokens_seen": 1655767040 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002516950852557673, + "loss": 1.3274, + "theoretical_loss": 3.5143265735912954, + "tokens_seen": 1655832576 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002516850551654965, + "loss": 1.3274, + "theoretical_loss": 3.5143146389150193, + "tokens_seen": 1655898112 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002516750250752257, + "loss": 1.3627, + "theoretical_loss": 3.5143027048433266, + "tokens_seen": 1655963648 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025166499498495486, + "loss": 1.3342, + "theoretical_loss": 3.514290771376162, + "tokens_seen": 1656029184 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025165496489468404, + "loss": 1.4256, + "theoretical_loss": 3.514278838513471, + "tokens_seen": 1656094720 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025164493480441327, + "loss": 1.3252, + "theoretical_loss": 3.5142669062552, + "tokens_seen": 1656160256 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002516349047141424, + "loss": 1.2709, + "theoretical_loss": 3.514254974601293, + "tokens_seen": 1656225792 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025162487462387163, + "loss": 1.4551, + "theoretical_loss": 3.5142430435516974, + "tokens_seen": 1656291328 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025161484453360076, + "loss": 1.3978, + "theoretical_loss": 3.514231113106357, + "tokens_seen": 1656356864 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025160481444333, + "loss": 1.3963, + "theoretical_loss": 3.514219183265218, + "tokens_seen": 1656422400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002515947843530592, + "loss": 1.3446, + "theoretical_loss": 3.514207254028226, + "tokens_seen": 1656487936 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025158475426278836, + "loss": 1.3952, + "theoretical_loss": 3.5141953253953258, + "tokens_seen": 1656553472 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025157472417251754, + "loss": 1.3341, + "theoretical_loss": 3.5141833973664633, + "tokens_seen": 1656619008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002515646940822468, + "loss": 1.341, + "theoretical_loss": 3.5141714699415845, + "tokens_seen": 1656684544 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002515546639919759, + "loss": 1.3464, + "theoretical_loss": 3.5141595431206345, + "tokens_seen": 1656750080 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025154463390170514, + "loss": 1.3366, + "theoretical_loss": 3.514147616903559, + "tokens_seen": 1656815616 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025153460381143426, + "loss": 1.3575, + "theoretical_loss": 3.5141356912903032, + "tokens_seen": 1656881152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002515245737211635, + "loss": 1.3887, + "theoretical_loss": 3.5141237662808127, + "tokens_seen": 1656946688 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002515145436308927, + "loss": 1.2782, + "theoretical_loss": 3.5141118418750334, + "tokens_seen": 1657012224 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025150451354062186, + "loss": 1.4391, + "theoretical_loss": 3.514099918072911, + "tokens_seen": 1657077760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514944834503511, + "loss": 1.2788, + "theoretical_loss": 3.514087994874391, + "tokens_seen": 1657143296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514844533600802, + "loss": 1.2396, + "theoretical_loss": 3.5140760722794178, + "tokens_seen": 1657208832 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025147442326980946, + "loss": 1.3319, + "theoretical_loss": 3.5140641502879384, + "tokens_seen": 1657274368 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025146439317953864, + "loss": 1.4077, + "theoretical_loss": 3.5140522288998977, + "tokens_seen": 1657339904 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514543630892678, + "loss": 1.3397, + "theoretical_loss": 3.514040308115242, + "tokens_seen": 1657405440 + }, + { + "epoch": 0.5, + "learning_rate": 0.000251444332998997, + "loss": 1.245, + "theoretical_loss": 3.514028387933916, + "tokens_seen": 1657470976 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514343029087262, + "loss": 1.3166, + "theoretical_loss": 3.514016468355866, + "tokens_seen": 1657536512 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025142427281845536, + "loss": 1.3207, + "theoretical_loss": 3.5140045493810366, + "tokens_seen": 1657602048 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514142427281846, + "loss": 1.3858, + "theoretical_loss": 3.5139926310093745, + "tokens_seen": 1657667584 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002514042126379137, + "loss": 1.3118, + "theoretical_loss": 3.5139807132408256, + "tokens_seen": 1657733120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025139418254764296, + "loss": 1.2726, + "theoretical_loss": 3.513968796075334, + "tokens_seen": 1657798656 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025138415245737214, + "loss": 1.4558, + "theoretical_loss": 3.513956879512847, + "tokens_seen": 1657864192 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002513741223671013, + "loss": 1.3107, + "theoretical_loss": 3.513944963553309, + "tokens_seen": 1657929728 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002513640922768305, + "loss": 1.3503, + "theoretical_loss": 3.513933048196667, + "tokens_seen": 1657995264 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 605625, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.340224027633667, + "objective/train/theoretical_loss": 3.5139211334428646, + "objective/train/tokens_used": 1678520800, + "theoretical_loss": 3.5139211334428646, + "tokens_seen": 1658060800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002513540621865597, + "loss": 1.2981, + "theoretical_loss": 3.5139211334428646, + "tokens_seen": 1658060800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025134403209628887, + "loss": 1.2845, + "theoretical_loss": 3.5139092192918495, + "tokens_seen": 1658126336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002513340020060181, + "loss": 1.4518, + "theoretical_loss": 3.5138973057435665, + "tokens_seen": 1658191872 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025132397191574723, + "loss": 1.3234, + "theoretical_loss": 3.5138853927979614, + "tokens_seen": 1658257408 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025131394182547647, + "loss": 1.2285, + "theoretical_loss": 3.51387348045498, + "tokens_seen": 1658322944 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002513039117352056, + "loss": 1.3809, + "theoretical_loss": 3.513861568714568, + "tokens_seen": 1658388480 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025129388164493483, + "loss": 1.4028, + "theoretical_loss": 3.513849657576671, + "tokens_seen": 1658454016 + }, + { + "epoch": 0.5, + "learning_rate": 0.000251283851554664, + "loss": 1.3812, + "theoretical_loss": 3.513837747041234, + "tokens_seen": 1658519552 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002512738214643932, + "loss": 1.3349, + "theoretical_loss": 3.513825837108204, + "tokens_seen": 1658585088 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025126379137412237, + "loss": 1.4382, + "theoretical_loss": 3.513813927777526, + "tokens_seen": 1658650624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025125376128385155, + "loss": 1.4483, + "theoretical_loss": 3.5138020190491464, + "tokens_seen": 1658716160 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025124373119358073, + "loss": 1.2384, + "theoretical_loss": 3.5137901109230096, + "tokens_seen": 1658781696 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025123370110330997, + "loss": 1.3427, + "theoretical_loss": 3.513778203399063, + "tokens_seen": 1658847232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002512236710130391, + "loss": 1.3084, + "theoretical_loss": 3.513766296477251, + "tokens_seen": 1658912768 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025121364092276833, + "loss": 1.2524, + "theoretical_loss": 3.51375439015752, + "tokens_seen": 1658978304 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002512036108324975, + "loss": 1.3557, + "theoretical_loss": 3.5137424844398164, + "tokens_seen": 1659043840 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002511935807422267, + "loss": 1.3273, + "theoretical_loss": 3.5137305793240845, + "tokens_seen": 1659109376 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002511835506519559, + "loss": 1.3791, + "theoretical_loss": 3.5137186748102716, + "tokens_seen": 1659174912 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025117352056168506, + "loss": 1.3718, + "theoretical_loss": 3.513706770898322, + "tokens_seen": 1659240448 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025116349047141424, + "loss": 1.352, + "theoretical_loss": 3.5136948675881827, + "tokens_seen": 1659305984 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025115346038114347, + "loss": 1.3633, + "theoretical_loss": 3.5136829648797985, + "tokens_seen": 1659371520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002511434302908726, + "loss": 1.2592, + "theoretical_loss": 3.5136710627731165, + "tokens_seen": 1659437056 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025113340020060183, + "loss": 1.352, + "theoretical_loss": 3.5136591612680816, + "tokens_seen": 1659502592 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025112337011033096, + "loss": 1.3356, + "theoretical_loss": 3.51364726036464, + "tokens_seen": 1659568128 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002511133400200602, + "loss": 1.4659, + "theoretical_loss": 3.513635360062737, + "tokens_seen": 1659633664 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002511033099297894, + "loss": 1.3137, + "theoretical_loss": 3.513623460362319, + "tokens_seen": 1659699200 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025109327983951856, + "loss": 1.3663, + "theoretical_loss": 3.5136115612633314, + "tokens_seen": 1659764736 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025108324974924774, + "loss": 1.4105, + "theoretical_loss": 3.5135996627657207, + "tokens_seen": 1659830272 + }, + { + "epoch": 0.5, + "learning_rate": 0.000251073219658977, + "loss": 1.2829, + "theoretical_loss": 3.513587764869432, + "tokens_seen": 1659895808 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002510631895687061, + "loss": 1.373, + "theoretical_loss": 3.5135758675744118, + "tokens_seen": 1659961344 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025105315947843534, + "loss": 1.3827, + "theoretical_loss": 3.5135639708806057, + "tokens_seen": 1660026880 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025104312938816446, + "loss": 1.3917, + "theoretical_loss": 3.5135520747879596, + "tokens_seen": 1660092416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002510330992978937, + "loss": 1.2757, + "theoretical_loss": 3.51354017929642, + "tokens_seen": 1660157952 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002510230692076229, + "loss": 1.3217, + "theoretical_loss": 3.5135282844059317, + "tokens_seen": 1660223488 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025101303911735206, + "loss": 1.3192, + "theoretical_loss": 3.5135163901164406, + "tokens_seen": 1660289024 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025100300902708124, + "loss": 1.3887, + "theoretical_loss": 3.5135044964278936, + "tokens_seen": 1660354560 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002509929789368104, + "loss": 1.442, + "theoretical_loss": 3.5134926033402367, + "tokens_seen": 1660420096 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002509829488465396, + "loss": 1.328, + "theoretical_loss": 3.5134807108534147, + "tokens_seen": 1660485632 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025097291875626884, + "loss": 1.3253, + "theoretical_loss": 3.513468818967374, + "tokens_seen": 1660551168 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025096288866599797, + "loss": 1.4256, + "theoretical_loss": 3.5134569276820606, + "tokens_seen": 1660616704 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002509528585757272, + "loss": 1.2763, + "theoretical_loss": 3.513445036997421, + "tokens_seen": 1660682240 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025094282848545633, + "loss": 1.4657, + "theoretical_loss": 3.5134331469134006, + "tokens_seen": 1660747776 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025093279839518556, + "loss": 1.2239, + "theoretical_loss": 3.513421257429945, + "tokens_seen": 1660813312 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025092276830491475, + "loss": 1.4466, + "theoretical_loss": 3.5134093685470007, + "tokens_seen": 1660878848 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002509127382146439, + "loss": 1.34, + "theoretical_loss": 3.513397480264514, + "tokens_seen": 1660944384 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002509027081243731, + "loss": 1.3663, + "theoretical_loss": 3.5133855925824298, + "tokens_seen": 1661009920 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025089267803410234, + "loss": 1.419, + "theoretical_loss": 3.513373705500695, + "tokens_seen": 1661075456 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025088264794383147, + "loss": 1.4371, + "theoretical_loss": 3.5133618190192553, + "tokens_seen": 1661140992 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002508726178535607, + "loss": 1.3258, + "theoretical_loss": 3.513349933138057, + "tokens_seen": 1661206528 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025086258776328983, + "loss": 1.3124, + "theoretical_loss": 3.5133380478570455, + "tokens_seen": 1661272064 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 606795, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4034759998321533, + "objective/train/theoretical_loss": 3.513326163176167, + "objective/train/tokens_used": 1681797600, + "theoretical_loss": 3.513326163176167, + "tokens_seen": 1661337600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025085255767301907, + "loss": 1.3486, + "theoretical_loss": 3.513326163176167, + "tokens_seen": 1661337600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025084252758274825, + "loss": 1.2915, + "theoretical_loss": 3.5133142790953684, + "tokens_seen": 1661403136 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025083249749247743, + "loss": 1.2036, + "theoretical_loss": 3.513302395614595, + "tokens_seen": 1661468672 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002508224674022066, + "loss": 1.3618, + "theoretical_loss": 3.513290512733792, + "tokens_seen": 1661534208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002508124373119358, + "loss": 1.3852, + "theoretical_loss": 3.513278630452907, + "tokens_seen": 1661599744 + }, + { + "epoch": 0.5, + "learning_rate": 0.000250802407221665, + "loss": 1.3269, + "theoretical_loss": 3.513266748771885, + "tokens_seen": 1661665280 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002507923771313942, + "loss": 1.3003, + "theoretical_loss": 3.5132548676906725, + "tokens_seen": 1661730816 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025078234704112334, + "loss": 1.215, + "theoretical_loss": 3.5132429872092157, + "tokens_seen": 1661796352 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025077231695085257, + "loss": 1.3163, + "theoretical_loss": 3.51323110732746, + "tokens_seen": 1661861888 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002507622868605817, + "loss": 1.2523, + "theoretical_loss": 3.5132192280453527, + "tokens_seen": 1661927424 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025075225677031093, + "loss": 1.361, + "theoretical_loss": 3.5132073493628386, + "tokens_seen": 1661992960 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025074222668004017, + "loss": 1.3597, + "theoretical_loss": 3.5131954712798645, + "tokens_seen": 1662058496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002507321965897693, + "loss": 1.3439, + "theoretical_loss": 3.513183593796376, + "tokens_seen": 1662124032 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025072216649949853, + "loss": 1.3583, + "theoretical_loss": 3.5131717169123196, + "tokens_seen": 1662189568 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002507121364092277, + "loss": 1.3883, + "theoretical_loss": 3.5131598406276416, + "tokens_seen": 1662255104 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002507021063189569, + "loss": 1.3577, + "theoretical_loss": 3.513147964942288, + "tokens_seen": 1662320640 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002506920762286861, + "loss": 1.3855, + "theoretical_loss": 3.5131360898562045, + "tokens_seen": 1662386176 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025068204613841526, + "loss": 1.4284, + "theoretical_loss": 3.513124215369338, + "tokens_seen": 1662451712 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025067201604814444, + "loss": 1.3478, + "theoretical_loss": 3.5131123414816336, + "tokens_seen": 1662517248 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025066198595787367, + "loss": 1.3938, + "theoretical_loss": 3.513100468193038, + "tokens_seen": 1662582784 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002506519558676028, + "loss": 1.3975, + "theoretical_loss": 3.513088595503498, + "tokens_seen": 1662648320 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025064192577733203, + "loss": 1.3164, + "theoretical_loss": 3.5130767234129587, + "tokens_seen": 1662713856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025063189568706116, + "loss": 1.371, + "theoretical_loss": 3.5130648519213663, + "tokens_seen": 1662779392 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002506218655967904, + "loss": 1.5393, + "theoretical_loss": 3.513052981028668, + "tokens_seen": 1662844928 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002506118355065196, + "loss": 1.299, + "theoretical_loss": 3.513041110734809, + "tokens_seen": 1662910464 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025060180541624876, + "loss": 1.2582, + "theoretical_loss": 3.513029241039737, + "tokens_seen": 1662976000 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025059177532597794, + "loss": 1.3222, + "theoretical_loss": 3.5130173719433957, + "tokens_seen": 1663041536 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002505817452357072, + "loss": 1.3498, + "theoretical_loss": 3.5130055034457337, + "tokens_seen": 1663107072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002505717151454363, + "loss": 1.3871, + "theoretical_loss": 3.5129936355466955, + "tokens_seen": 1663172608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025056168505516554, + "loss": 1.3375, + "theoretical_loss": 3.5129817682462283, + "tokens_seen": 1663238144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025055165496489466, + "loss": 1.4611, + "theoretical_loss": 3.512969901544278, + "tokens_seen": 1663303680 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002505416248746239, + "loss": 1.3733, + "theoretical_loss": 3.512958035440791, + "tokens_seen": 1663369216 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002505315947843531, + "loss": 1.4081, + "theoretical_loss": 3.512946169935713, + "tokens_seen": 1663434752 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025052156469408226, + "loss": 1.3649, + "theoretical_loss": 3.512934305028991, + "tokens_seen": 1663500288 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025051153460381144, + "loss": 1.2797, + "theoretical_loss": 3.512922440720571, + "tokens_seen": 1663565824 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002505015045135406, + "loss": 1.3605, + "theoretical_loss": 3.512910577010399, + "tokens_seen": 1663631360 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002504914744232698, + "loss": 1.3223, + "theoretical_loss": 3.5128987138984216, + "tokens_seen": 1663696896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025048144433299904, + "loss": 1.2535, + "theoretical_loss": 3.512886851384585, + "tokens_seen": 1663762432 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025047141424272817, + "loss": 1.4, + "theoretical_loss": 3.512874989468835, + "tokens_seen": 1663827968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002504613841524574, + "loss": 1.3957, + "theoretical_loss": 3.5128631281511185, + "tokens_seen": 1663893504 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025045135406218653, + "loss": 1.3236, + "theoretical_loss": 3.512851267431382, + "tokens_seen": 1663959040 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025044132397191576, + "loss": 1.3225, + "theoretical_loss": 3.5128394073095706, + "tokens_seen": 1664024576 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025043129388164495, + "loss": 1.3084, + "theoretical_loss": 3.512827547785632, + "tokens_seen": 1664090112 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002504212637913741, + "loss": 1.3686, + "theoretical_loss": 3.512815688859512, + "tokens_seen": 1664155648 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002504112337011033, + "loss": 1.1508, + "theoretical_loss": 3.5128038305311566, + "tokens_seen": 1664221184 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025040120361083254, + "loss": 1.3924, + "theoretical_loss": 3.512791972800512, + "tokens_seen": 1664286720 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025039117352056167, + "loss": 1.4459, + "theoretical_loss": 3.5127801156675256, + "tokens_seen": 1664352256 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002503811434302909, + "loss": 1.4005, + "theoretical_loss": 3.5127682591321427, + "tokens_seen": 1664417792 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025037111334002003, + "loss": 1.3891, + "theoretical_loss": 3.5127564031943104, + "tokens_seen": 1664483328 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025036108324974927, + "loss": 1.4516, + "theoretical_loss": 3.5127445478539743, + "tokens_seen": 1664548864 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 608106, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2460435628890991, + "objective/train/theoretical_loss": 3.512732693111081, + "objective/train/tokens_used": 1685074400, + "theoretical_loss": 3.512732693111081, + "tokens_seen": 1664614400 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025035105315947845, + "loss": 1.3187, + "theoretical_loss": 3.512732693111081, + "tokens_seen": 1664614400 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025034102306920763, + "loss": 1.4033, + "theoretical_loss": 3.5127208389655777, + "tokens_seen": 1664679936 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002503309929789368, + "loss": 1.3892, + "theoretical_loss": 3.51270898541741, + "tokens_seen": 1664745472 + }, + { + "epoch": 0.5, + "learning_rate": 0.000250320962888666, + "loss": 1.3064, + "theoretical_loss": 3.5126971324665237, + "tokens_seen": 1664811008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002503109327983952, + "loss": 1.2571, + "theoretical_loss": 3.512685280112866, + "tokens_seen": 1664876544 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002503009027081244, + "loss": 1.3712, + "theoretical_loss": 3.5126734283563836, + "tokens_seen": 1664942080 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025029087261785354, + "loss": 1.4182, + "theoretical_loss": 3.512661577197022, + "tokens_seen": 1665007616 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025028084252758277, + "loss": 1.3316, + "theoretical_loss": 3.5126497266347285, + "tokens_seen": 1665073152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502708124373119, + "loss": 1.2318, + "theoretical_loss": 3.512637876669449, + "tokens_seen": 1665138688 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025026078234704113, + "loss": 1.4516, + "theoretical_loss": 3.5126260273011303, + "tokens_seen": 1665204224 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502507522567703, + "loss": 1.3294, + "theoretical_loss": 3.5126141785297182, + "tokens_seen": 1665269760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502407221664995, + "loss": 1.3476, + "theoretical_loss": 3.5126023303551595, + "tokens_seen": 1665335296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502306920762287, + "loss": 1.2634, + "theoretical_loss": 3.5125904827774006, + "tokens_seen": 1665400832 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502206619859579, + "loss": 1.4762, + "theoretical_loss": 3.5125786357963884, + "tokens_seen": 1665466368 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025021063189568704, + "loss": 1.3398, + "theoretical_loss": 3.5125667894120687, + "tokens_seen": 1665531904 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002502006018054163, + "loss": 1.3537, + "theoretical_loss": 3.5125549436243886, + "tokens_seen": 1665597440 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002501905717151454, + "loss": 1.4065, + "theoretical_loss": 3.512543098433294, + "tokens_seen": 1665662976 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025018054162487464, + "loss": 1.4306, + "theoretical_loss": 3.5125312538387314, + "tokens_seen": 1665728512 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002501705115346038, + "loss": 1.4101, + "theoretical_loss": 3.512519409840648, + "tokens_seen": 1665794048 + }, + { + "epoch": 0.5, + "learning_rate": 0.000250160481444333, + "loss": 1.2963, + "theoretical_loss": 3.5125075664389893, + "tokens_seen": 1665859584 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002501504513540622, + "loss": 1.3226, + "theoretical_loss": 3.5124957236337027, + "tokens_seen": 1665925120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025014042126379136, + "loss": 1.3522, + "theoretical_loss": 3.512483881424734, + "tokens_seen": 1665990656 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025013039117352054, + "loss": 1.2788, + "theoretical_loss": 3.5124720398120304, + "tokens_seen": 1666056192 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002501203610832498, + "loss": 1.3018, + "theoretical_loss": 3.5124601987955377, + "tokens_seen": 1666121728 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002501103309929789, + "loss": 1.3884, + "theoretical_loss": 3.5124483583752033, + "tokens_seen": 1666187264 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025010030090270814, + "loss": 1.353, + "theoretical_loss": 3.5124365185509725, + "tokens_seen": 1666252800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00025009027081243727, + "loss": 1.3906, + "theoretical_loss": 3.512424679322793, + "tokens_seen": 1666318336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002500802407221665, + "loss": 1.3926, + "theoretical_loss": 3.5124128406906108, + "tokens_seen": 1666383872 + }, + { + "epoch": 0.5, + "learning_rate": 0.0002500702106318957, + "loss": 1.3088, + "theoretical_loss": 3.512401002654373, + "tokens_seen": 1666449408 + }, + { + "epoch": 0.51, + "learning_rate": 0.00025006018054162486, + "loss": 1.3856, + "theoretical_loss": 3.5123891652140253, + "tokens_seen": 1666514944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00025005015045135405, + "loss": 1.4165, + "theoretical_loss": 3.5123773283695146, + "tokens_seen": 1666580480 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002500401203610833, + "loss": 1.4567, + "theoretical_loss": 3.512365492120788, + "tokens_seen": 1666646016 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002500300902708124, + "loss": 1.3288, + "theoretical_loss": 3.5123536564677913, + "tokens_seen": 1666711552 + }, + { + "epoch": 0.51, + "learning_rate": 0.00025002006018054164, + "loss": 1.4321, + "theoretical_loss": 3.5123418214104722, + "tokens_seen": 1666777088 + }, + { + "epoch": 0.51, + "learning_rate": 0.00025001003009027077, + "loss": 1.4363, + "theoretical_loss": 3.5123299869487763, + "tokens_seen": 1666842624 + }, + { + "epoch": 0.51, + "learning_rate": 0.00025, + "loss": 1.4892, + "theoretical_loss": 3.51231815308265, + "tokens_seen": 1666908160 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002499899699097292, + "loss": 1.3486, + "theoretical_loss": 3.5123063198120406, + "tokens_seen": 1666973696 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024997993981945837, + "loss": 1.356, + "theoretical_loss": 3.5122944871368946, + "tokens_seen": 1667039232 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024996990972918755, + "loss": 1.2421, + "theoretical_loss": 3.5122826550571586, + "tokens_seen": 1667104768 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024995987963891673, + "loss": 1.4235, + "theoretical_loss": 3.5122708235727798, + "tokens_seen": 1667170304 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024994984954864597, + "loss": 1.3591, + "theoretical_loss": 3.5122589926837033, + "tokens_seen": 1667235840 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024993981945837515, + "loss": 1.3861, + "theoretical_loss": 3.512247162389877, + "tokens_seen": 1667301376 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024992978936810433, + "loss": 1.2666, + "theoretical_loss": 3.5122353326912474, + "tokens_seen": 1667366912 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002499197592778335, + "loss": 1.2442, + "theoretical_loss": 3.512223503587761, + "tokens_seen": 1667432448 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002499097291875627, + "loss": 1.4768, + "theoretical_loss": 3.512211675079364, + "tokens_seen": 1667497984 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024989969909729187, + "loss": 1.266, + "theoretical_loss": 3.512199847166004, + "tokens_seen": 1667563520 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024988966900702105, + "loss": 1.3648, + "theoretical_loss": 3.512188019847627, + "tokens_seen": 1667629056 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024987963891675023, + "loss": 1.4459, + "theoretical_loss": 3.5121761931241804, + "tokens_seen": 1667694592 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002498696088264794, + "loss": 1.4004, + "theoretical_loss": 3.5121643669956097, + "tokens_seen": 1667760128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024985957873620865, + "loss": 1.384, + "theoretical_loss": 3.512152541461863, + "tokens_seen": 1667825664 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 609331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4771723747253418, + "objective/train/theoretical_loss": 3.512140716522886, + "objective/train/tokens_used": 1688351200, + "theoretical_loss": 3.512140716522886, + "tokens_seen": 1667891200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024984954864593783, + "loss": 1.4244, + "theoretical_loss": 3.512140716522886, + "tokens_seen": 1667891200 + }, + { + "epoch": 0.51, + "learning_rate": 0.000249839518555667, + "loss": 1.309, + "theoretical_loss": 3.512128892178626, + "tokens_seen": 1667956736 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002498294884653962, + "loss": 1.3356, + "theoretical_loss": 3.5121170684290286, + "tokens_seen": 1668022272 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002498194583751254, + "loss": 1.2857, + "theoretical_loss": 3.5121052452740424, + "tokens_seen": 1668087808 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024980942828485455, + "loss": 1.3376, + "theoretical_loss": 3.5120934227136127, + "tokens_seen": 1668153344 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024979939819458374, + "loss": 1.1767, + "theoretical_loss": 3.512081600747687, + "tokens_seen": 1668218880 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497893681043129, + "loss": 1.3475, + "theoretical_loss": 3.5120697793762115, + "tokens_seen": 1668284416 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497793380140421, + "loss": 1.2944, + "theoretical_loss": 3.512057958599133, + "tokens_seen": 1668349952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024976930792377133, + "loss": 1.3232, + "theoretical_loss": 3.512046138416399, + "tokens_seen": 1668415488 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497592778335005, + "loss": 1.3515, + "theoretical_loss": 3.512034318827955, + "tokens_seen": 1668481024 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497492477432297, + "loss": 1.4161, + "theoretical_loss": 3.512022499833749, + "tokens_seen": 1668546560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497392176529589, + "loss": 1.2583, + "theoretical_loss": 3.5120106814337273, + "tokens_seen": 1668612096 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024972918756268806, + "loss": 1.3881, + "theoretical_loss": 3.511998863627837, + "tokens_seen": 1668677632 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024971915747241724, + "loss": 1.376, + "theoretical_loss": 3.511987046416024, + "tokens_seen": 1668743168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002497091273821464, + "loss": 1.2576, + "theoretical_loss": 3.511975229798236, + "tokens_seen": 1668808704 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002496990972918756, + "loss": 1.3349, + "theoretical_loss": 3.5119634137744193, + "tokens_seen": 1668874240 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002496890672016048, + "loss": 1.2268, + "theoretical_loss": 3.5119515983445213, + "tokens_seen": 1668939776 + }, + { + "epoch": 0.51, + "learning_rate": 0.000249679037111334, + "loss": 1.3968, + "theoretical_loss": 3.5119397835084882, + "tokens_seen": 1669005312 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002496690070210632, + "loss": 1.4543, + "theoretical_loss": 3.5119279692662673, + "tokens_seen": 1669070848 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002496589769307924, + "loss": 1.3454, + "theoretical_loss": 3.5119161556178047, + "tokens_seen": 1669136384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024964894684052156, + "loss": 1.3096, + "theoretical_loss": 3.511904342563048, + "tokens_seen": 1669201920 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024963891675025074, + "loss": 1.4015, + "theoretical_loss": 3.511892530101944, + "tokens_seen": 1669267456 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002496288866599799, + "loss": 1.2834, + "theoretical_loss": 3.5118807182344396, + "tokens_seen": 1669332992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024961885656970916, + "loss": 1.5139, + "theoretical_loss": 3.511868906960481, + "tokens_seen": 1669398528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024960882647943834, + "loss": 1.4259, + "theoretical_loss": 3.511857096280016, + "tokens_seen": 1669464064 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002495987963891675, + "loss": 1.4205, + "theoretical_loss": 3.5118452861929907, + "tokens_seen": 1669529600 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002495887662988967, + "loss": 1.3635, + "theoretical_loss": 3.5118334766993526, + "tokens_seen": 1669595136 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002495787362086259, + "loss": 1.3316, + "theoretical_loss": 3.5118216677990484, + "tokens_seen": 1669660672 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024956870611835506, + "loss": 1.3161, + "theoretical_loss": 3.5118098594920246, + "tokens_seen": 1669726208 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024955867602808425, + "loss": 1.4242, + "theoretical_loss": 3.5117980517782286, + "tokens_seen": 1669791744 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002495486459378135, + "loss": 1.4176, + "theoretical_loss": 3.511786244657607, + "tokens_seen": 1669857280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024953861584754266, + "loss": 1.3885, + "theoretical_loss": 3.5117744381301064, + "tokens_seen": 1669922816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024952858575727184, + "loss": 1.1929, + "theoretical_loss": 3.5117626321956745, + "tokens_seen": 1669988352 + }, + { + "epoch": 0.51, + "learning_rate": 0.000249518555667001, + "loss": 1.298, + "theoretical_loss": 3.5117508268542585, + "tokens_seen": 1670053888 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002495085255767302, + "loss": 1.4086, + "theoretical_loss": 3.5117390221058042, + "tokens_seen": 1670119424 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002494984954864594, + "loss": 1.4463, + "theoretical_loss": 3.511727217950259, + "tokens_seen": 1670184960 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024948846539618857, + "loss": 1.3707, + "theoretical_loss": 3.5117154143875706, + "tokens_seen": 1670250496 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024947843530591775, + "loss": 1.3195, + "theoretical_loss": 3.511703611417685, + "tokens_seen": 1670316032 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024946840521564693, + "loss": 1.4287, + "theoretical_loss": 3.5116918090405496, + "tokens_seen": 1670381568 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024945837512537617, + "loss": 1.346, + "theoretical_loss": 3.5116800072561114, + "tokens_seen": 1670447104 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024944834503510535, + "loss": 1.4062, + "theoretical_loss": 3.511668206064317, + "tokens_seen": 1670512640 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024943831494483453, + "loss": 1.3026, + "theoretical_loss": 3.5116564054651134, + "tokens_seen": 1670578176 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002494282848545637, + "loss": 1.3571, + "theoretical_loss": 3.5116446054584483, + "tokens_seen": 1670643712 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002494182547642929, + "loss": 1.4049, + "theoretical_loss": 3.5116328060442683, + "tokens_seen": 1670709248 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024940822467402207, + "loss": 1.3029, + "theoretical_loss": 3.5116210072225202, + "tokens_seen": 1670774784 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024939819458375125, + "loss": 1.3479, + "theoretical_loss": 3.511609208993151, + "tokens_seen": 1670840320 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024938816449348043, + "loss": 1.4792, + "theoretical_loss": 3.5115974113561084, + "tokens_seen": 1670905856 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002493781344032096, + "loss": 1.3879, + "theoretical_loss": 3.511585614311339, + "tokens_seen": 1670971392 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024936810431293885, + "loss": 1.5227, + "theoretical_loss": 3.5115738178587894, + "tokens_seen": 1671036928 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024935807422266803, + "loss": 1.5433, + "theoretical_loss": 3.511562021998407, + "tokens_seen": 1671102464 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 610761, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3009065389633179, + "objective/train/theoretical_loss": 3.5115502267301393, + "objective/train/tokens_used": 1691628000, + "theoretical_loss": 3.5115502267301393, + "tokens_seen": 1671168000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002493480441323972, + "loss": 1.3171, + "theoretical_loss": 3.5115502267301393, + "tokens_seen": 1671168000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002493380140421264, + "loss": 1.3737, + "theoretical_loss": 3.5115384320539325, + "tokens_seen": 1671233536 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002493279839518556, + "loss": 1.2783, + "theoretical_loss": 3.5115266379697343, + "tokens_seen": 1671299072 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024931795386158476, + "loss": 1.2901, + "theoretical_loss": 3.5115148444774915, + "tokens_seen": 1671364608 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024930792377131394, + "loss": 1.1631, + "theoretical_loss": 3.5115030515771517, + "tokens_seen": 1671430144 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492978936810431, + "loss": 1.4616, + "theoretical_loss": 3.511491259268661, + "tokens_seen": 1671495680 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492878635907723, + "loss": 1.2645, + "theoretical_loss": 3.5114794675519674, + "tokens_seen": 1671561216 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024927783350050153, + "loss": 1.2097, + "theoretical_loss": 3.5114676764270176, + "tokens_seen": 1671626752 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492678034102307, + "loss": 1.3919, + "theoretical_loss": 3.5114558858937586, + "tokens_seen": 1671692288 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492577733199599, + "loss": 1.3777, + "theoretical_loss": 3.511444095952138, + "tokens_seen": 1671757824 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492477432296891, + "loss": 1.3137, + "theoretical_loss": 3.511432306602102, + "tokens_seen": 1671823360 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024923771313941826, + "loss": 1.3394, + "theoretical_loss": 3.5114205178435984, + "tokens_seen": 1671888896 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024922768304914744, + "loss": 1.401, + "theoretical_loss": 3.511408729676575, + "tokens_seen": 1671954432 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492176529588766, + "loss": 1.372, + "theoretical_loss": 3.5113969421009776, + "tokens_seen": 1672019968 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002492076228686058, + "loss": 1.337, + "theoretical_loss": 3.511385155116754, + "tokens_seen": 1672085504 + }, + { + "epoch": 0.51, + "learning_rate": 0.000249197592778335, + "loss": 1.2978, + "theoretical_loss": 3.5113733687238513, + "tokens_seen": 1672151040 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491875626880642, + "loss": 1.2674, + "theoretical_loss": 3.511361582922217, + "tokens_seen": 1672216576 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491775325977934, + "loss": 1.271, + "theoretical_loss": 3.5113497977117976, + "tokens_seen": 1672282112 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491675025075226, + "loss": 1.3859, + "theoretical_loss": 3.5113380130925407, + "tokens_seen": 1672347648 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024915747241725176, + "loss": 1.3111, + "theoretical_loss": 3.5113262290643936, + "tokens_seen": 1672413184 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024914744232698094, + "loss": 1.2658, + "theoretical_loss": 3.5113144456273027, + "tokens_seen": 1672478720 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491374122367101, + "loss": 1.3839, + "theoretical_loss": 3.5113026627812163, + "tokens_seen": 1672544256 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491273821464393, + "loss": 1.2671, + "theoretical_loss": 3.5112908805260803, + "tokens_seen": 1672609792 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002491173520561685, + "loss": 1.3569, + "theoretical_loss": 3.5112790988618436, + "tokens_seen": 1672675328 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024910732196589767, + "loss": 1.3052, + "theoretical_loss": 3.511267317788452, + "tokens_seen": 1672740864 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002490972918756269, + "loss": 1.3766, + "theoretical_loss": 3.5112555373058534, + "tokens_seen": 1672806400 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002490872617853561, + "loss": 1.3357, + "theoretical_loss": 3.511243757413995, + "tokens_seen": 1672871936 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024907723169508526, + "loss": 1.4089, + "theoretical_loss": 3.511231978112823, + "tokens_seen": 1672937472 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024906720160481445, + "loss": 1.2827, + "theoretical_loss": 3.5112201994022865, + "tokens_seen": 1673003008 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002490571715145436, + "loss": 1.3596, + "theoretical_loss": 3.5112084212823316, + "tokens_seen": 1673068544 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002490471414242728, + "loss": 1.1673, + "theoretical_loss": 3.5111966437529056, + "tokens_seen": 1673134080 + }, + { + "epoch": 0.51, + "learning_rate": 0.000249037111334002, + "loss": 1.3981, + "theoretical_loss": 3.511184866813956, + "tokens_seen": 1673199616 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024902708124373117, + "loss": 1.4242, + "theoretical_loss": 3.5111730904654297, + "tokens_seen": 1673265152 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024901705115346035, + "loss": 1.3163, + "theoretical_loss": 3.5111613147072744, + "tokens_seen": 1673330688 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002490070210631896, + "loss": 1.3787, + "theoretical_loss": 3.511149539539437, + "tokens_seen": 1673396224 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024899699097291877, + "loss": 1.4478, + "theoretical_loss": 3.5111377649618656, + "tokens_seen": 1673461760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024898696088264795, + "loss": 1.3693, + "theoretical_loss": 3.5111259909745067, + "tokens_seen": 1673527296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024897693079237713, + "loss": 1.3723, + "theoretical_loss": 3.5111142175773074, + "tokens_seen": 1673592832 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002489669007021063, + "loss": 1.3518, + "theoretical_loss": 3.5111024447702155, + "tokens_seen": 1673658368 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002489568706118355, + "loss": 1.332, + "theoretical_loss": 3.5110906725531787, + "tokens_seen": 1673723904 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002489468405215647, + "loss": 1.3407, + "theoretical_loss": 3.5110789009261434, + "tokens_seen": 1673789440 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024893681043129385, + "loss": 1.4008, + "theoretical_loss": 3.5110671298890574, + "tokens_seen": 1673854976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024892678034102304, + "loss": 1.2513, + "theoretical_loss": 3.511055359441868, + "tokens_seen": 1673920512 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024891675025075227, + "loss": 1.3447, + "theoretical_loss": 3.5110435895845233, + "tokens_seen": 1673986048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024890672016048145, + "loss": 1.3892, + "theoretical_loss": 3.511031820316969, + "tokens_seen": 1674051584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024889669007021063, + "loss": 1.4218, + "theoretical_loss": 3.511020051639154, + "tokens_seen": 1674117120 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002488866599799398, + "loss": 1.3077, + "theoretical_loss": 3.511008283551025, + "tokens_seen": 1674182656 + }, + { + "epoch": 0.51, + "learning_rate": 0.000248876629889669, + "loss": 1.3309, + "theoretical_loss": 3.510996516052529, + "tokens_seen": 1674248192 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024886659979939823, + "loss": 1.3587, + "theoretical_loss": 3.5109847491436144, + "tokens_seen": 1674313728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002488565697091274, + "loss": 1.3489, + "theoretical_loss": 3.5109729828242275, + "tokens_seen": 1674379264 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 611357, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.326997995376587, + "objective/train/theoretical_loss": 3.5109612170943163, + "objective/train/tokens_used": 1694904800, + "theoretical_loss": 3.5109612170943163, + "tokens_seen": 1674444800 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002488465396188566, + "loss": 1.2752, + "theoretical_loss": 3.5109612170943163, + "tokens_seen": 1674444800 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002488365095285858, + "loss": 1.3865, + "theoretical_loss": 3.5109494519538282, + "tokens_seen": 1674510336 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024882647943831496, + "loss": 1.1848, + "theoretical_loss": 3.5109376874027105, + "tokens_seen": 1674575872 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024881644934804414, + "loss": 1.3732, + "theoretical_loss": 3.5109259234409107, + "tokens_seen": 1674641408 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002488064192577733, + "loss": 1.3256, + "theoretical_loss": 3.510914160068376, + "tokens_seen": 1674706944 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487963891675025, + "loss": 1.2968, + "theoretical_loss": 3.5109023972850535, + "tokens_seen": 1674772480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024878635907723173, + "loss": 1.3255, + "theoretical_loss": 3.5108906350908917, + "tokens_seen": 1674838016 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487763289869609, + "loss": 1.3931, + "theoretical_loss": 3.510878873485837, + "tokens_seen": 1674903552 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487662988966901, + "loss": 1.2856, + "theoretical_loss": 3.510867112469837, + "tokens_seen": 1674969088 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487562688064193, + "loss": 1.3036, + "theoretical_loss": 3.51085535204284, + "tokens_seen": 1675034624 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024874623871614846, + "loss": 1.3678, + "theoretical_loss": 3.5108435922047927, + "tokens_seen": 1675100160 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024873620862587764, + "loss": 1.28, + "theoretical_loss": 3.5108318329556427, + "tokens_seen": 1675165696 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487261785356068, + "loss": 1.3905, + "theoretical_loss": 3.5108200742953377, + "tokens_seen": 1675231232 + }, + { + "epoch": 0.51, + "learning_rate": 0.000248716148445336, + "loss": 1.3425, + "theoretical_loss": 3.510808316223825, + "tokens_seen": 1675296768 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002487061183550652, + "loss": 1.3239, + "theoretical_loss": 3.5107965587410517, + "tokens_seen": 1675362304 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486960882647944, + "loss": 1.3054, + "theoretical_loss": 3.510784801846966, + "tokens_seen": 1675427840 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486860581745236, + "loss": 1.3447, + "theoretical_loss": 3.5107730455415145, + "tokens_seen": 1675493376 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486760280842528, + "loss": 1.3603, + "theoretical_loss": 3.5107612898246456, + "tokens_seen": 1675558912 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024866599799398196, + "loss": 1.2801, + "theoretical_loss": 3.5107495346963065, + "tokens_seen": 1675624448 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024865596790371114, + "loss": 1.3895, + "theoretical_loss": 3.510737780156445, + "tokens_seen": 1675689984 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486459378134403, + "loss": 1.2294, + "theoretical_loss": 3.510726026205008, + "tokens_seen": 1675755520 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486359077231695, + "loss": 1.3017, + "theoretical_loss": 3.5107142728419434, + "tokens_seen": 1675821056 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486258776328987, + "loss": 1.2937, + "theoretical_loss": 3.5107025200671984, + "tokens_seen": 1675886592 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024861584754262787, + "loss": 1.2326, + "theoretical_loss": 3.510690767880721, + "tokens_seen": 1675952128 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002486058174523571, + "loss": 1.3742, + "theoretical_loss": 3.510679016282459, + "tokens_seen": 1676017664 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002485957873620863, + "loss": 1.4058, + "theoretical_loss": 3.5106672652723594, + "tokens_seen": 1676083200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024858575727181546, + "loss": 1.3435, + "theoretical_loss": 3.51065551485037, + "tokens_seen": 1676148736 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024857572718154465, + "loss": 1.375, + "theoretical_loss": 3.5106437650164377, + "tokens_seen": 1676214272 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002485656970912738, + "loss": 1.3233, + "theoretical_loss": 3.5106320157705113, + "tokens_seen": 1676279808 + }, + { + "epoch": 0.51, + "learning_rate": 0.000248555667001003, + "loss": 1.3565, + "theoretical_loss": 3.5106202671125377, + "tokens_seen": 1676345344 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002485456369107322, + "loss": 1.2487, + "theoretical_loss": 3.5106085190424645, + "tokens_seen": 1676410880 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024853560682046137, + "loss": 1.3045, + "theoretical_loss": 3.5105967715602393, + "tokens_seen": 1676476416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024852557673019055, + "loss": 1.4098, + "theoretical_loss": 3.5105850246658097, + "tokens_seen": 1676541952 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002485155466399198, + "loss": 1.4574, + "theoretical_loss": 3.5105732783591237, + "tokens_seen": 1676607488 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024850551654964897, + "loss": 1.4106, + "theoretical_loss": 3.5105615326401285, + "tokens_seen": 1676673024 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024849548645937815, + "loss": 1.3016, + "theoretical_loss": 3.5105497875087717, + "tokens_seen": 1676738560 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024848545636910733, + "loss": 1.2032, + "theoretical_loss": 3.510538042965001, + "tokens_seen": 1676804096 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002484754262788365, + "loss": 1.3688, + "theoretical_loss": 3.510526299008764, + "tokens_seen": 1676869632 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002484653961885657, + "loss": 1.29, + "theoretical_loss": 3.5105145556400092, + "tokens_seen": 1676935168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002484553660982949, + "loss": 1.3434, + "theoretical_loss": 3.5105028128586824, + "tokens_seen": 1677000704 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024844533600802405, + "loss": 1.2903, + "theoretical_loss": 3.5104910706647336, + "tokens_seen": 1677066240 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024843530591775324, + "loss": 1.3188, + "theoretical_loss": 3.5104793290581084, + "tokens_seen": 1677131776 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024842527582748247, + "loss": 1.3723, + "theoretical_loss": 3.5104675880387552, + "tokens_seen": 1677197312 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024841524573721165, + "loss": 1.3325, + "theoretical_loss": 3.510455847606622, + "tokens_seen": 1677262848 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024840521564694083, + "loss": 1.2298, + "theoretical_loss": 3.5104441077616566, + "tokens_seen": 1677328384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024839518555667, + "loss": 1.3238, + "theoretical_loss": 3.510432368503806, + "tokens_seen": 1677393920 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002483851554663992, + "loss": 1.2682, + "theoretical_loss": 3.5104206298330185, + "tokens_seen": 1677459456 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002483751253761284, + "loss": 1.3649, + "theoretical_loss": 3.5104088917492415, + "tokens_seen": 1677524992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024836509528585756, + "loss": 1.3608, + "theoretical_loss": 3.510397154252423, + "tokens_seen": 1677590528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024835506519558674, + "loss": 1.272, + "theoretical_loss": 3.51038541734251, + "tokens_seen": 1677656064 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 612648, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.265733242034912, + "objective/train/theoretical_loss": 3.5103736810194515, + "objective/train/tokens_used": 1698181600, + "theoretical_loss": 3.5103736810194515, + "tokens_seen": 1677721600 + }, + { + "epoch": 0.51, + "learning_rate": 0.000248345035105316, + "loss": 1.3035, + "theoretical_loss": 3.5103736810194515, + "tokens_seen": 1677721600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024833500501504516, + "loss": 1.3127, + "theoretical_loss": 3.5103619452831936, + "tokens_seen": 1677787136 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024832497492477434, + "loss": 1.3456, + "theoretical_loss": 3.5103502101336854, + "tokens_seen": 1677852672 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002483149448345035, + "loss": 1.4063, + "theoretical_loss": 3.510338475570874, + "tokens_seen": 1677918208 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002483049147442327, + "loss": 1.3554, + "theoretical_loss": 3.510326741594708, + "tokens_seen": 1677983744 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002482948846539619, + "loss": 1.3765, + "theoretical_loss": 3.5103150082051338, + "tokens_seen": 1678049280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024828485456369106, + "loss": 1.3652, + "theoretical_loss": 3.5103032754021, + "tokens_seen": 1678114816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024827482447342024, + "loss": 1.4146, + "theoretical_loss": 3.5102915431855544, + "tokens_seen": 1678180352 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002482647943831494, + "loss": 1.1959, + "theoretical_loss": 3.5102798115554448, + "tokens_seen": 1678245888 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024825476429287866, + "loss": 1.3159, + "theoretical_loss": 3.510268080511718, + "tokens_seen": 1678311424 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024824473420260784, + "loss": 1.4064, + "theoretical_loss": 3.5102563500543233, + "tokens_seen": 1678376960 + }, + { + "epoch": 0.51, + "learning_rate": 0.000248234704112337, + "loss": 1.4492, + "theoretical_loss": 3.510244620183207, + "tokens_seen": 1678442496 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002482246740220662, + "loss": 1.2989, + "theoretical_loss": 3.510232890898319, + "tokens_seen": 1678508032 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002482146439317954, + "loss": 1.3357, + "theoretical_loss": 3.510221162199605, + "tokens_seen": 1678573568 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024820461384152456, + "loss": 1.3158, + "theoretical_loss": 3.5102094340870136, + "tokens_seen": 1678639104 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024819458375125375, + "loss": 1.2768, + "theoretical_loss": 3.5101977065604926, + "tokens_seen": 1678704640 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481845536609829, + "loss": 1.3932, + "theoretical_loss": 3.5101859796199903, + "tokens_seen": 1678770176 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481745235707121, + "loss": 1.3341, + "theoretical_loss": 3.510174253265454, + "tokens_seen": 1678835712 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024816449348044134, + "loss": 1.4073, + "theoretical_loss": 3.5101625274968318, + "tokens_seen": 1678901248 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481544633901705, + "loss": 1.2373, + "theoretical_loss": 3.510150802314071, + "tokens_seen": 1678966784 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481444332998997, + "loss": 1.3135, + "theoretical_loss": 3.5101390777171204, + "tokens_seen": 1679032320 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481344032096289, + "loss": 1.3452, + "theoretical_loss": 3.5101273537059274, + "tokens_seen": 1679097856 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024812437311935807, + "loss": 1.2182, + "theoretical_loss": 3.5101156302804393, + "tokens_seen": 1679163392 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481143430290873, + "loss": 1.3909, + "theoretical_loss": 3.5101039074406053, + "tokens_seen": 1679228928 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002481043129388165, + "loss": 1.3053, + "theoretical_loss": 3.5100921851863722, + "tokens_seen": 1679294464 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024809428284854566, + "loss": 1.3701, + "theoretical_loss": 3.5100804635176877, + "tokens_seen": 1679360000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024808425275827485, + "loss": 1.3853, + "theoretical_loss": 3.510068742434501, + "tokens_seen": 1679425536 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024807422266800403, + "loss": 1.3596, + "theoretical_loss": 3.510057021936759, + "tokens_seen": 1679491072 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002480641925777332, + "loss": 1.302, + "theoretical_loss": 3.5100453020244093, + "tokens_seen": 1679556608 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002480541624874624, + "loss": 1.3946, + "theoretical_loss": 3.510033582697401, + "tokens_seen": 1679622144 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024804413239719157, + "loss": 1.3403, + "theoretical_loss": 3.5100218639556813, + "tokens_seen": 1679687680 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024803410230692075, + "loss": 1.3887, + "theoretical_loss": 3.510010145799198, + "tokens_seen": 1679753216 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024802407221665, + "loss": 1.2575, + "theoretical_loss": 3.5099984282278993, + "tokens_seen": 1679818752 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024801404212637917, + "loss": 1.222, + "theoretical_loss": 3.5099867112417336, + "tokens_seen": 1679884288 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024800401203610835, + "loss": 1.5053, + "theoretical_loss": 3.509974994840648, + "tokens_seen": 1679949824 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024799398194583753, + "loss": 1.4598, + "theoretical_loss": 3.5099632790245905, + "tokens_seen": 1680015360 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002479839518555667, + "loss": 1.3908, + "theoretical_loss": 3.5099515637935097, + "tokens_seen": 1680080896 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002479739217652959, + "loss": 1.3631, + "theoretical_loss": 3.5099398491473535, + "tokens_seen": 1680146432 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002479638916750251, + "loss": 1.2894, + "theoretical_loss": 3.5099281350860694, + "tokens_seen": 1680211968 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024795386158475425, + "loss": 1.3228, + "theoretical_loss": 3.5099164216096055, + "tokens_seen": 1680277504 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024794383149448344, + "loss": 1.3686, + "theoretical_loss": 3.5099047087179107, + "tokens_seen": 1680343040 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024793380140421267, + "loss": 1.3983, + "theoretical_loss": 3.5098929964109313, + "tokens_seen": 1680408576 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024792377131394185, + "loss": 1.2973, + "theoretical_loss": 3.5098812846886167, + "tokens_seen": 1680474112 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024791374122367103, + "loss": 1.4109, + "theoretical_loss": 3.5098695735509144, + "tokens_seen": 1680539648 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002479037111334002, + "loss": 1.4068, + "theoretical_loss": 3.509857862997773, + "tokens_seen": 1680605184 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478936810431294, + "loss": 1.3924, + "theoretical_loss": 3.509846153029139, + "tokens_seen": 1680670720 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478836509528586, + "loss": 1.3856, + "theoretical_loss": 3.509834443644962, + "tokens_seen": 1680736256 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024787362086258776, + "loss": 1.4927, + "theoretical_loss": 3.50982273484519, + "tokens_seen": 1680801792 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024786359077231694, + "loss": 1.4346, + "theoretical_loss": 3.50981102662977, + "tokens_seen": 1680867328 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478535606820462, + "loss": 1.3664, + "theoretical_loss": 3.5097993189986507, + "tokens_seen": 1680932864 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 613922, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3088704347610474, + "objective/train/theoretical_loss": 3.50978761195178, + "objective/train/tokens_used": 1701458400, + "theoretical_loss": 3.50978761195178, + "tokens_seen": 1680998400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024784353059177536, + "loss": 1.3561, + "theoretical_loss": 3.50978761195178, + "tokens_seen": 1680998400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024783350050150454, + "loss": 1.3765, + "theoretical_loss": 3.5097759054891062, + "tokens_seen": 1681063936 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478234704112337, + "loss": 1.4741, + "theoretical_loss": 3.5097641996105766, + "tokens_seen": 1681129472 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478134403209629, + "loss": 1.4095, + "theoretical_loss": 3.5097524943161407, + "tokens_seen": 1681195008 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002478034102306921, + "loss": 1.2945, + "theoretical_loss": 3.5097407896057455, + "tokens_seen": 1681260544 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024779338014042126, + "loss": 1.4271, + "theoretical_loss": 3.5097290854793393, + "tokens_seen": 1681326080 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024778335005015044, + "loss": 1.3796, + "theoretical_loss": 3.50971738193687, + "tokens_seen": 1681391616 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002477733199598796, + "loss": 1.3584, + "theoretical_loss": 3.5097056789782863, + "tokens_seen": 1681457152 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024776328986960886, + "loss": 1.4263, + "theoretical_loss": 3.509693976603536, + "tokens_seen": 1681522688 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024775325977933804, + "loss": 1.4071, + "theoretical_loss": 3.509682274812567, + "tokens_seen": 1681588224 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002477432296890672, + "loss": 1.3971, + "theoretical_loss": 3.509670573605328, + "tokens_seen": 1681653760 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002477331995987964, + "loss": 1.3612, + "theoretical_loss": 3.509658872981767, + "tokens_seen": 1681719296 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002477231695085256, + "loss": 1.3825, + "theoretical_loss": 3.509647172941831, + "tokens_seen": 1681784832 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024771313941825476, + "loss": 1.3378, + "theoretical_loss": 3.50963547348547, + "tokens_seen": 1681850368 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024770310932798395, + "loss": 1.3624, + "theoretical_loss": 3.5096237746126304, + "tokens_seen": 1681915904 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476930792377131, + "loss": 1.3614, + "theoretical_loss": 3.509612076323261, + "tokens_seen": 1681981440 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476830491474423, + "loss": 1.3048, + "theoretical_loss": 3.5096003786173116, + "tokens_seen": 1682046976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024767301905717154, + "loss": 1.344, + "theoretical_loss": 3.5095886814947277, + "tokens_seen": 1682112512 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476629889669007, + "loss": 1.3482, + "theoretical_loss": 3.509576984955459, + "tokens_seen": 1682178048 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476529588766299, + "loss": 1.4698, + "theoretical_loss": 3.509565288999453, + "tokens_seen": 1682243584 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476429287863591, + "loss": 1.281, + "theoretical_loss": 3.5095535936266593, + "tokens_seen": 1682309120 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024763289869608827, + "loss": 1.334, + "theoretical_loss": 3.509541898837024, + "tokens_seen": 1682374656 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024762286860581745, + "loss": 1.3266, + "theoretical_loss": 3.5095302046304973, + "tokens_seen": 1682440192 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024761283851554663, + "loss": 1.2987, + "theoretical_loss": 3.509518511007026, + "tokens_seen": 1682505728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002476028084252758, + "loss": 1.3726, + "theoretical_loss": 3.5095068179665585, + "tokens_seen": 1682571264 + }, + { + "epoch": 0.51, + "learning_rate": 0.000247592778335005, + "loss": 1.3147, + "theoretical_loss": 3.5094951255090443, + "tokens_seen": 1682636800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024758274824473423, + "loss": 1.3602, + "theoretical_loss": 3.50948343363443, + "tokens_seen": 1682702336 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002475727181544634, + "loss": 1.3336, + "theoretical_loss": 3.5094717423426642, + "tokens_seen": 1682767872 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002475626880641926, + "loss": 1.3415, + "theoretical_loss": 3.5094600516336962, + "tokens_seen": 1682833408 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024755265797392177, + "loss": 1.2795, + "theoretical_loss": 3.5094483615074736, + "tokens_seen": 1682898944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024754262788365095, + "loss": 1.4268, + "theoretical_loss": 3.509436671963944, + "tokens_seen": 1682964480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024753259779338013, + "loss": 1.2426, + "theoretical_loss": 3.5094249830030564, + "tokens_seen": 1683030016 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002475225677031093, + "loss": 1.4138, + "theoretical_loss": 3.509413294624759, + "tokens_seen": 1683095552 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002475125376128385, + "loss": 1.3762, + "theoretical_loss": 3.509401606829, + "tokens_seen": 1683161088 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002475025075225677, + "loss": 1.4119, + "theoretical_loss": 3.509389919615728, + "tokens_seen": 1683226624 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474924774322969, + "loss": 1.3737, + "theoretical_loss": 3.509378232984891, + "tokens_seen": 1683292160 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474824473420261, + "loss": 1.3491, + "theoretical_loss": 3.5093665469364366, + "tokens_seen": 1683357696 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474724172517553, + "loss": 1.3659, + "theoretical_loss": 3.509354861470314, + "tokens_seen": 1683423232 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024746238716148445, + "loss": 1.3846, + "theoretical_loss": 3.509343176586472, + "tokens_seen": 1683488768 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024745235707121364, + "loss": 1.3001, + "theoretical_loss": 3.5093314922848577, + "tokens_seen": 1683554304 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474423269809428, + "loss": 1.4276, + "theoretical_loss": 3.5093198085654196, + "tokens_seen": 1683619840 + }, + { + "epoch": 0.51, + "learning_rate": 0.000247432296890672, + "loss": 1.3428, + "theoretical_loss": 3.509308125428107, + "tokens_seen": 1683685376 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474222668004012, + "loss": 1.3822, + "theoretical_loss": 3.5092964428728672, + "tokens_seen": 1683750912 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024741223671013036, + "loss": 1.2765, + "theoretical_loss": 3.509284760899649, + "tokens_seen": 1683816448 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002474022066198596, + "loss": 1.3746, + "theoretical_loss": 3.509273079508401, + "tokens_seen": 1683881984 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002473921765295888, + "loss": 1.5062, + "theoretical_loss": 3.5092613986990715, + "tokens_seen": 1683947520 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024738214643931796, + "loss": 1.429, + "theoretical_loss": 3.509249718471608, + "tokens_seen": 1684013056 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024737211634904714, + "loss": 1.431, + "theoretical_loss": 3.50923803882596, + "tokens_seen": 1684078592 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002473620862587764, + "loss": 1.2408, + "theoretical_loss": 3.509226359762075, + "tokens_seen": 1684144128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024735205616850556, + "loss": 1.4027, + "theoretical_loss": 3.5092146812799023, + "tokens_seen": 1684209664 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 615159, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2585194110870361, + "objective/train/theoretical_loss": 3.5092030033793895, + "objective/train/tokens_used": 1704735200, + "theoretical_loss": 3.5092030033793895, + "tokens_seen": 1684275200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024734202607823474, + "loss": 1.3053, + "theoretical_loss": 3.5092030033793895, + "tokens_seen": 1684275200 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002473319959879639, + "loss": 1.3532, + "theoretical_loss": 3.5091913260604857, + "tokens_seen": 1684340736 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002473219658976931, + "loss": 1.2922, + "theoretical_loss": 3.509179649323138, + "tokens_seen": 1684406272 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002473119358074223, + "loss": 1.3686, + "theoretical_loss": 3.5091679731672967, + "tokens_seen": 1684471808 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024730190571715146, + "loss": 1.3559, + "theoretical_loss": 3.5091562975929085, + "tokens_seen": 1684537344 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024729187562688064, + "loss": 1.3799, + "theoretical_loss": 3.5091446225999228, + "tokens_seen": 1684602880 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002472818455366098, + "loss": 1.3532, + "theoretical_loss": 3.509132948188287, + "tokens_seen": 1684668416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024727181544633906, + "loss": 1.2662, + "theoretical_loss": 3.5091212743579514, + "tokens_seen": 1684733952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024726178535606824, + "loss": 1.309, + "theoretical_loss": 3.5091096011088627, + "tokens_seen": 1684799488 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002472517552657974, + "loss": 1.3902, + "theoretical_loss": 3.5090979284409705, + "tokens_seen": 1684865024 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002472417251755266, + "loss": 1.3674, + "theoretical_loss": 3.5090862563542227, + "tokens_seen": 1684930560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002472316950852558, + "loss": 1.308, + "theoretical_loss": 3.509074584848568, + "tokens_seen": 1684996096 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024722166499498496, + "loss": 1.2274, + "theoretical_loss": 3.5090629139239535, + "tokens_seen": 1685061632 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024721163490471415, + "loss": 1.3324, + "theoretical_loss": 3.50905124358033, + "tokens_seen": 1685127168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002472016048144433, + "loss": 1.3459, + "theoretical_loss": 3.5090395738176445, + "tokens_seen": 1685192704 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002471915747241725, + "loss": 1.2409, + "theoretical_loss": 3.509027904635846, + "tokens_seen": 1685258240 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024718154463390174, + "loss": 1.4536, + "theoretical_loss": 3.509016236034883, + "tokens_seen": 1685323776 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002471715145436309, + "loss": 1.4113, + "theoretical_loss": 3.5090045680147037, + "tokens_seen": 1685389312 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002471614844533601, + "loss": 1.3135, + "theoretical_loss": 3.5089929005752563, + "tokens_seen": 1685454848 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002471514543630893, + "loss": 1.4568, + "theoretical_loss": 3.50898123371649, + "tokens_seen": 1685520384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024714142427281847, + "loss": 1.3973, + "theoretical_loss": 3.5089695674383536, + "tokens_seen": 1685585920 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024713139418254765, + "loss": 1.4024, + "theoretical_loss": 3.5089579017407946, + "tokens_seen": 1685651456 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024712136409227683, + "loss": 1.3994, + "theoretical_loss": 3.5089462366237623, + "tokens_seen": 1685716992 + }, + { + "epoch": 0.51, + "learning_rate": 0.000247111334002006, + "loss": 1.353, + "theoretical_loss": 3.508934572087205, + "tokens_seen": 1685782528 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002471013039117352, + "loss": 1.2824, + "theoretical_loss": 3.5089229081310713, + "tokens_seen": 1685848064 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024709127382146443, + "loss": 1.3263, + "theoretical_loss": 3.5089112447553097, + "tokens_seen": 1685913600 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470812437311936, + "loss": 1.382, + "theoretical_loss": 3.508899581959869, + "tokens_seen": 1685979136 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470712136409228, + "loss": 1.3207, + "theoretical_loss": 3.5088879197446974, + "tokens_seen": 1686044672 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024706118355065197, + "loss": 1.3373, + "theoretical_loss": 3.5088762581097432, + "tokens_seen": 1686110208 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024705115346038115, + "loss": 1.4344, + "theoretical_loss": 3.5088645970549557, + "tokens_seen": 1686175744 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024704112337011033, + "loss": 1.3943, + "theoretical_loss": 3.5088529365802836, + "tokens_seen": 1686241280 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470310932798395, + "loss": 1.3367, + "theoretical_loss": 3.5088412766856747, + "tokens_seen": 1686306816 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470210631895687, + "loss": 1.3128, + "theoretical_loss": 3.5088296173710782, + "tokens_seen": 1686372352 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470110330992979, + "loss": 1.328, + "theoretical_loss": 3.508817958636442, + "tokens_seen": 1686437888 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002470010030090271, + "loss": 1.2499, + "theoretical_loss": 3.508806300481716, + "tokens_seen": 1686503424 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002469909729187563, + "loss": 1.3098, + "theoretical_loss": 3.5087946429068477, + "tokens_seen": 1686568960 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002469809428284855, + "loss": 1.4099, + "theoretical_loss": 3.5087829859117856, + "tokens_seen": 1686634496 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024697091273821466, + "loss": 1.2747, + "theoretical_loss": 3.5087713294964793, + "tokens_seen": 1686700032 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024696088264794384, + "loss": 1.3606, + "theoretical_loss": 3.508759673660877, + "tokens_seen": 1686765568 + }, + { + "epoch": 0.51, + "learning_rate": 0.000246950852557673, + "loss": 1.3935, + "theoretical_loss": 3.508748018404927, + "tokens_seen": 1686831104 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002469408224674022, + "loss": 1.4027, + "theoretical_loss": 3.5087363637285787, + "tokens_seen": 1686896640 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002469307923771314, + "loss": 1.4615, + "theoretical_loss": 3.50872470963178, + "tokens_seen": 1686962176 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024692076228686056, + "loss": 1.5587, + "theoretical_loss": 3.5087130561144795, + "tokens_seen": 1687027712 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002469107321965898, + "loss": 1.471, + "theoretical_loss": 3.5087014031766266, + "tokens_seen": 1687093248 + }, + { + "epoch": 0.51, + "learning_rate": 0.000246900702106319, + "loss": 1.2799, + "theoretical_loss": 3.50868975081817, + "tokens_seen": 1687158784 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024689067201604816, + "loss": 1.4519, + "theoretical_loss": 3.5086780990390576, + "tokens_seen": 1687224320 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024688064192577734, + "loss": 1.3691, + "theoretical_loss": 3.5086664478392384, + "tokens_seen": 1687289856 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002468706118355065, + "loss": 1.5214, + "theoretical_loss": 3.5086547972186617, + "tokens_seen": 1687355392 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002468605817452357, + "loss": 1.3828, + "theoretical_loss": 3.5086431471772754, + "tokens_seen": 1687420928 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002468505516549649, + "loss": 1.3613, + "theoretical_loss": 3.5086314977150286, + "tokens_seen": 1687486464 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 616363, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4138387441635132, + "objective/train/theoretical_loss": 3.50861984883187, + "objective/train/tokens_used": 1708012000, + "theoretical_loss": 3.50861984883187, + "tokens_seen": 1687552000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024684052156469406, + "loss": 1.389, + "theoretical_loss": 3.50861984883187, + "tokens_seen": 1687552000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024683049147442325, + "loss": 1.3951, + "theoretical_loss": 3.508608200527749, + "tokens_seen": 1687617536 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002468204613841525, + "loss": 1.469, + "theoretical_loss": 3.5085965528026124, + "tokens_seen": 1687683072 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024681043129388166, + "loss": 1.2842, + "theoretical_loss": 3.5085849056564107, + "tokens_seen": 1687748608 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024680040120361084, + "loss": 1.3568, + "theoretical_loss": 3.508573259089092, + "tokens_seen": 1687814144 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024679037111334, + "loss": 1.4414, + "theoretical_loss": 3.508561613100605, + "tokens_seen": 1687879680 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002467803410230692, + "loss": 1.3269, + "theoretical_loss": 3.5085499676908993, + "tokens_seen": 1687945216 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002467703109327984, + "loss": 1.3375, + "theoretical_loss": 3.5085383228599225, + "tokens_seen": 1688010752 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024676028084252757, + "loss": 1.367, + "theoretical_loss": 3.5085266786076237, + "tokens_seen": 1688076288 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024675025075225675, + "loss": 1.3842, + "theoretical_loss": 3.5085150349339522, + "tokens_seen": 1688141824 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024674022066198593, + "loss": 1.3527, + "theoretical_loss": 3.508503391838856, + "tokens_seen": 1688207360 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024673019057171516, + "loss": 1.2578, + "theoretical_loss": 3.5084917493222845, + "tokens_seen": 1688272896 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024672016048144435, + "loss": 1.4414, + "theoretical_loss": 3.5084801073841865, + "tokens_seen": 1688338432 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002467101303911735, + "loss": 1.3236, + "theoretical_loss": 3.5084684660245107, + "tokens_seen": 1688403968 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002467001003009027, + "loss": 1.2562, + "theoretical_loss": 3.5084568252432056, + "tokens_seen": 1688469504 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002466900702106319, + "loss": 1.2448, + "theoretical_loss": 3.5084451850402205, + "tokens_seen": 1688535040 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024668004012036107, + "loss": 1.3393, + "theoretical_loss": 3.5084335454155036, + "tokens_seen": 1688600576 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024667001003009025, + "loss": 1.4707, + "theoretical_loss": 3.508421906369004, + "tokens_seen": 1688666112 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024665997993981943, + "loss": 1.3908, + "theoretical_loss": 3.508410267900671, + "tokens_seen": 1688731648 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002466499498495486, + "loss": 1.2334, + "theoretical_loss": 3.5083986300104533, + "tokens_seen": 1688797184 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024663991975927785, + "loss": 1.4001, + "theoretical_loss": 3.508386992698299, + "tokens_seen": 1688862720 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024662988966900703, + "loss": 1.1865, + "theoretical_loss": 3.5083753559641577, + "tokens_seen": 1688928256 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002466198595787362, + "loss": 1.3252, + "theoretical_loss": 3.508363719807978, + "tokens_seen": 1688993792 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002466098294884654, + "loss": 1.3559, + "theoretical_loss": 3.508352084229709, + "tokens_seen": 1689059328 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024659979939819463, + "loss": 1.346, + "theoretical_loss": 3.5083404492292996, + "tokens_seen": 1689124864 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002465897693079238, + "loss": 1.1906, + "theoretical_loss": 3.508328814806698, + "tokens_seen": 1689190400 + }, + { + "epoch": 0.51, + "learning_rate": 0.000246579739217653, + "loss": 1.3209, + "theoretical_loss": 3.5083171809618534, + "tokens_seen": 1689255936 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024656970912738217, + "loss": 1.4444, + "theoretical_loss": 3.5083055476947154, + "tokens_seen": 1689321472 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024655967903711135, + "loss": 1.3729, + "theoretical_loss": 3.508293915005232, + "tokens_seen": 1689387008 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024654964894684053, + "loss": 1.5384, + "theoretical_loss": 3.5082822828933526, + "tokens_seen": 1689452544 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002465396188565697, + "loss": 1.4049, + "theoretical_loss": 3.5082706513590263, + "tokens_seen": 1689518080 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002465295887662989, + "loss": 1.2085, + "theoretical_loss": 3.5082590204022015, + "tokens_seen": 1689583616 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002465195586760281, + "loss": 1.4742, + "theoretical_loss": 3.5082473900228273, + "tokens_seen": 1689649152 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002465095285857573, + "loss": 1.4461, + "theoretical_loss": 3.5082357602208525, + "tokens_seen": 1689714688 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464994984954865, + "loss": 1.3797, + "theoretical_loss": 3.5082241309962265, + "tokens_seen": 1689780224 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464894684052157, + "loss": 1.2313, + "theoretical_loss": 3.508212502348898, + "tokens_seen": 1689845760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024647943831494486, + "loss": 1.4153, + "theoretical_loss": 3.508200874278816, + "tokens_seen": 1689911296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024646940822467404, + "loss": 1.3963, + "theoretical_loss": 3.5081892467859292, + "tokens_seen": 1689976832 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464593781344032, + "loss": 1.4256, + "theoretical_loss": 3.5081776198701866, + "tokens_seen": 1690042368 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464493480441324, + "loss": 1.4033, + "theoretical_loss": 3.5081659935315375, + "tokens_seen": 1690107904 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464393179538616, + "loss": 1.2935, + "theoretical_loss": 3.5081543677699307, + "tokens_seen": 1690173440 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024642928786359076, + "loss": 1.3129, + "theoretical_loss": 3.5081427425853153, + "tokens_seen": 1690238976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024641925777332, + "loss": 1.3527, + "theoretical_loss": 3.5081311179776398, + "tokens_seen": 1690304512 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002464092276830492, + "loss": 1.3774, + "theoretical_loss": 3.508119493946854, + "tokens_seen": 1690370048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024639919759277836, + "loss": 1.3383, + "theoretical_loss": 3.5081078704929065, + "tokens_seen": 1690435584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024638916750250754, + "loss": 1.4107, + "theoretical_loss": 3.5080962476157462, + "tokens_seen": 1690501120 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002463791374122367, + "loss": 1.3304, + "theoretical_loss": 3.5080846253153224, + "tokens_seen": 1690566656 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002463691073219659, + "loss": 1.3799, + "theoretical_loss": 3.508073003591584, + "tokens_seen": 1690632192 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002463590772316951, + "loss": 1.403, + "theoretical_loss": 3.5080613824444793, + "tokens_seen": 1690697728 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024634904714142426, + "loss": 1.2545, + "theoretical_loss": 3.508049761873959, + "tokens_seen": 1690763264 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 617433, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.380690336227417, + "objective/train/theoretical_loss": 3.5080381418799704, + "objective/train/tokens_used": 1711288800, + "theoretical_loss": 3.5080381418799704, + "tokens_seen": 1690828800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024633901705115345, + "loss": 1.3584, + "theoretical_loss": 3.5080381418799704, + "tokens_seen": 1690828800 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002463289869608827, + "loss": 1.3843, + "theoretical_loss": 3.5080265224624636, + "tokens_seen": 1690894336 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024631895687061186, + "loss": 1.3268, + "theoretical_loss": 3.5080149036213877, + "tokens_seen": 1690959872 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024630892678034104, + "loss": 1.2941, + "theoretical_loss": 3.508003285356691, + "tokens_seen": 1691025408 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002462988966900702, + "loss": 1.396, + "theoretical_loss": 3.5079916676683234, + "tokens_seen": 1691090944 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002462888665997994, + "loss": 1.418, + "theoretical_loss": 3.507980050556233, + "tokens_seen": 1691156480 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002462788365095286, + "loss": 1.3715, + "theoretical_loss": 3.50796843402037, + "tokens_seen": 1691222016 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024626880641925777, + "loss": 1.3479, + "theoretical_loss": 3.507956818060683, + "tokens_seen": 1691287552 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024625877632898695, + "loss": 1.3351, + "theoretical_loss": 3.5079452026771207, + "tokens_seen": 1691353088 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024624874623871613, + "loss": 1.4005, + "theoretical_loss": 3.507933587869633, + "tokens_seen": 1691418624 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024623871614844536, + "loss": 1.3632, + "theoretical_loss": 3.5079219736381684, + "tokens_seen": 1691484160 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024622868605817455, + "loss": 1.3704, + "theoretical_loss": 3.507910359982676, + "tokens_seen": 1691549696 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002462186559679037, + "loss": 1.371, + "theoretical_loss": 3.5078987469031055, + "tokens_seen": 1691615232 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002462086258776329, + "loss": 1.3351, + "theoretical_loss": 3.507887134399405, + "tokens_seen": 1691680768 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002461985957873621, + "loss": 1.4649, + "theoretical_loss": 3.507875522471525, + "tokens_seen": 1691746304 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024618856569709127, + "loss": 1.253, + "theoretical_loss": 3.5078639111194136, + "tokens_seen": 1691811840 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024617853560682045, + "loss": 1.34, + "theoretical_loss": 3.5078523003430204, + "tokens_seen": 1691877376 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024616850551654963, + "loss": 1.4018, + "theoretical_loss": 3.5078406901422943, + "tokens_seen": 1691942912 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002461584754262788, + "loss": 1.3781, + "theoretical_loss": 3.5078290805171846, + "tokens_seen": 1692008448 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024614844533600805, + "loss": 1.3586, + "theoretical_loss": 3.5078174714676407, + "tokens_seen": 1692073984 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024613841524573723, + "loss": 1.4277, + "theoretical_loss": 3.507805862993611, + "tokens_seen": 1692139520 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002461283851554664, + "loss": 1.4526, + "theoretical_loss": 3.507794255095046, + "tokens_seen": 1692205056 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002461183550651956, + "loss": 1.3678, + "theoretical_loss": 3.5077826477718936, + "tokens_seen": 1692270592 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002461083249749248, + "loss": 1.3536, + "theoretical_loss": 3.5077710410241036, + "tokens_seen": 1692336128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024609829488465395, + "loss": 1.3764, + "theoretical_loss": 3.507759434851625, + "tokens_seen": 1692401664 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024608826479438314, + "loss": 1.3485, + "theoretical_loss": 3.5077478292544075, + "tokens_seen": 1692467200 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002460782347041123, + "loss": 1.3448, + "theoretical_loss": 3.5077362242323993, + "tokens_seen": 1692532736 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002460682046138415, + "loss": 1.3108, + "theoretical_loss": 3.5077246197855505, + "tokens_seen": 1692598272 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024605817452357073, + "loss": 1.2553, + "theoretical_loss": 3.5077130159138106, + "tokens_seen": 1692663808 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002460481444332999, + "loss": 1.4035, + "theoretical_loss": 3.507701412617128, + "tokens_seen": 1692729344 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002460381143430291, + "loss": 1.3589, + "theoretical_loss": 3.507689809895452, + "tokens_seen": 1692794880 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002460280842527583, + "loss": 1.3485, + "theoretical_loss": 3.5076782077487323, + "tokens_seen": 1692860416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024601805416248746, + "loss": 1.3427, + "theoretical_loss": 3.507666606176918, + "tokens_seen": 1692925952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024600802407221664, + "loss": 1.3749, + "theoretical_loss": 3.5076550051799575, + "tokens_seen": 1692991488 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459979939819458, + "loss": 1.4418, + "theoretical_loss": 3.507643404757802, + "tokens_seen": 1693057024 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245987963891675, + "loss": 1.4488, + "theoretical_loss": 3.5076318049103987, + "tokens_seen": 1693122560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459779338014042, + "loss": 1.3501, + "theoretical_loss": 3.507620205637698, + "tokens_seen": 1693188096 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459679037111334, + "loss": 1.4649, + "theoretical_loss": 3.50760860693965, + "tokens_seen": 1693253632 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459578736208626, + "loss": 1.3274, + "theoretical_loss": 3.507597008816202, + "tokens_seen": 1693319168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459478435305918, + "loss": 1.3831, + "theoretical_loss": 3.5075854112673044, + "tokens_seen": 1693384704 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024593781344032096, + "loss": 1.273, + "theoretical_loss": 3.5075738142929067, + "tokens_seen": 1693450240 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024592778335005014, + "loss": 1.2879, + "theoretical_loss": 3.507562217892958, + "tokens_seen": 1693515776 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459177532597793, + "loss": 1.2381, + "theoretical_loss": 3.5075506220674066, + "tokens_seen": 1693581312 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002459077231695085, + "loss": 1.3411, + "theoretical_loss": 3.507539026816204, + "tokens_seen": 1693646848 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002458976930792377, + "loss": 1.362, + "theoretical_loss": 3.5075274321392973, + "tokens_seen": 1693712384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024588766298896687, + "loss": 1.3877, + "theoretical_loss": 3.507515838036637, + "tokens_seen": 1693777920 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002458776328986961, + "loss": 1.2485, + "theoretical_loss": 3.5075042445081728, + "tokens_seen": 1693843456 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002458676028084253, + "loss": 1.3479, + "theoretical_loss": 3.5074926515538527, + "tokens_seen": 1693908992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024585757271815446, + "loss": 1.2669, + "theoretical_loss": 3.507481059173627, + "tokens_seen": 1693974528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024584754262788365, + "loss": 1.3736, + "theoretical_loss": 3.5074694673674456, + "tokens_seen": 1694040064 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 618718, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4511775970458984, + "objective/train/theoretical_loss": 3.507457876135257, + "objective/train/tokens_used": 1714565600, + "theoretical_loss": 3.507457876135257, + "tokens_seen": 1694105600 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002458375125376129, + "loss": 1.4514, + "theoretical_loss": 3.507457876135257, + "tokens_seen": 1694105600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024582748244734206, + "loss": 1.4053, + "theoretical_loss": 3.5074462854770103, + "tokens_seen": 1694171136 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024581745235707124, + "loss": 1.3289, + "theoretical_loss": 3.5074346953926554, + "tokens_seen": 1694236672 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002458074222668004, + "loss": 1.3751, + "theoretical_loss": 3.507423105882142, + "tokens_seen": 1694302208 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002457973921765296, + "loss": 1.2681, + "theoretical_loss": 3.5074115169454188, + "tokens_seen": 1694367744 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002457873620862588, + "loss": 1.3241, + "theoretical_loss": 3.5073999285824353, + "tokens_seen": 1694433280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024577733199598797, + "loss": 1.2367, + "theoretical_loss": 3.507388340793142, + "tokens_seen": 1694498816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024576730190571715, + "loss": 1.311, + "theoretical_loss": 3.5073767535774865, + "tokens_seen": 1694564352 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024575727181544633, + "loss": 1.4114, + "theoretical_loss": 3.5073651669354193, + "tokens_seen": 1694629888 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024574724172517556, + "loss": 1.369, + "theoretical_loss": 3.50735358086689, + "tokens_seen": 1694695424 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024573721163490475, + "loss": 1.2613, + "theoretical_loss": 3.5073419953718474, + "tokens_seen": 1694760960 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024572718154463393, + "loss": 1.3193, + "theoretical_loss": 3.507330410450242, + "tokens_seen": 1694826496 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002457171514543631, + "loss": 1.364, + "theoretical_loss": 3.507318826102022, + "tokens_seen": 1694892032 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002457071213640923, + "loss": 1.3624, + "theoretical_loss": 3.507307242327137, + "tokens_seen": 1694957568 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024569709127382147, + "loss": 1.3927, + "theoretical_loss": 3.5072956591255373, + "tokens_seen": 1695023104 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024568706118355065, + "loss": 1.3058, + "theoretical_loss": 3.5072840764971716, + "tokens_seen": 1695088640 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024567703109327983, + "loss": 1.3637, + "theoretical_loss": 3.50727249444199, + "tokens_seen": 1695154176 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245667001003009, + "loss": 1.2978, + "theoretical_loss": 3.507260912959941, + "tokens_seen": 1695219712 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024565697091273825, + "loss": 1.3175, + "theoretical_loss": 3.507249332050975, + "tokens_seen": 1695285248 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024564694082246743, + "loss": 1.3339, + "theoretical_loss": 3.5072377517150413, + "tokens_seen": 1695350784 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002456369107321966, + "loss": 1.2477, + "theoretical_loss": 3.507226171952089, + "tokens_seen": 1695416320 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002456268806419258, + "loss": 1.3136, + "theoretical_loss": 3.507214592762068, + "tokens_seen": 1695481856 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245616850551655, + "loss": 1.3241, + "theoretical_loss": 3.5072030141449284, + "tokens_seen": 1695547392 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024560682046138415, + "loss": 1.1637, + "theoretical_loss": 3.507191436100618, + "tokens_seen": 1695612928 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024559679037111334, + "loss": 1.4099, + "theoretical_loss": 3.5071798586290877, + "tokens_seen": 1695678464 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002455867602808425, + "loss": 1.4232, + "theoretical_loss": 3.507168281730287, + "tokens_seen": 1695744000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002455767301905717, + "loss": 1.3186, + "theoretical_loss": 3.5071567054041646, + "tokens_seen": 1695809536 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024556670010030093, + "loss": 1.2322, + "theoretical_loss": 3.507145129650671, + "tokens_seen": 1695875072 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002455566700100301, + "loss": 1.4849, + "theoretical_loss": 3.507133554469755, + "tokens_seen": 1695940608 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002455466399197593, + "loss": 1.3144, + "theoretical_loss": 3.507121979861367, + "tokens_seen": 1696006144 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002455366098294885, + "loss": 1.4557, + "theoretical_loss": 3.507110405825455, + "tokens_seen": 1696071680 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024552657973921766, + "loss": 1.3244, + "theoretical_loss": 3.50709883236197, + "tokens_seen": 1696137216 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024551654964894684, + "loss": 1.262, + "theoretical_loss": 3.507087259470861, + "tokens_seen": 1696202752 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245506519558676, + "loss": 1.3467, + "theoretical_loss": 3.5070756871520783, + "tokens_seen": 1696268288 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454964894684052, + "loss": 1.3468, + "theoretical_loss": 3.50706411540557, + "tokens_seen": 1696333824 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454864593781344, + "loss": 1.3546, + "theoretical_loss": 3.5070525442312874, + "tokens_seen": 1696399360 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454764292878636, + "loss": 1.3305, + "theoretical_loss": 3.507040973629179, + "tokens_seen": 1696464896 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454663991975928, + "loss": 1.2876, + "theoretical_loss": 3.507029403599195, + "tokens_seen": 1696530432 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245456369107322, + "loss": 1.3695, + "theoretical_loss": 3.5070178341412843, + "tokens_seen": 1696595968 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024544633901705116, + "loss": 1.3925, + "theoretical_loss": 3.5070062652553977, + "tokens_seen": 1696661504 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024543630892678034, + "loss": 1.3215, + "theoretical_loss": 3.506994696941483, + "tokens_seen": 1696727040 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454262788365095, + "loss": 1.3822, + "theoretical_loss": 3.5069831291994915, + "tokens_seen": 1696792576 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454162487462387, + "loss": 1.2597, + "theoretical_loss": 3.506971562029372, + "tokens_seen": 1696858112 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002454062186559679, + "loss": 1.2778, + "theoretical_loss": 3.5069599954310746, + "tokens_seen": 1696923648 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024539618856569707, + "loss": 1.4464, + "theoretical_loss": 3.5069484294045483, + "tokens_seen": 1696989184 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002453861584754263, + "loss": 1.3961, + "theoretical_loss": 3.5069368639497434, + "tokens_seen": 1697054720 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002453761283851555, + "loss": 1.4282, + "theoretical_loss": 3.5069252990666095, + "tokens_seen": 1697120256 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024536609829488466, + "loss": 1.3172, + "theoretical_loss": 3.506913734755096, + "tokens_seen": 1697185792 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024535606820461385, + "loss": 1.2968, + "theoretical_loss": 3.5069021710151524, + "tokens_seen": 1697251328 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245346038114343, + "loss": 1.4324, + "theoretical_loss": 3.506890607846729, + "tokens_seen": 1697316864 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 619912, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.303410530090332, + "objective/train/theoretical_loss": 3.506879045249775, + "objective/train/tokens_used": 1717842400, + "theoretical_loss": 3.506879045249775, + "tokens_seen": 1697382400 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002453360080240722, + "loss": 1.3895, + "theoretical_loss": 3.506879045249775, + "tokens_seen": 1697382400 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002453259779338014, + "loss": 1.3807, + "theoretical_loss": 3.5068674832242404, + "tokens_seen": 1697447936 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024531594784353057, + "loss": 1.246, + "theoretical_loss": 3.5068559217700748, + "tokens_seen": 1697513472 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024530591775325975, + "loss": 1.4142, + "theoretical_loss": 3.5068443608872277, + "tokens_seen": 1697579008 + }, + { + "epoch": 0.51, + "learning_rate": 0.000245295887662989, + "loss": 1.2278, + "theoretical_loss": 3.5068328005756486, + "tokens_seen": 1697644544 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024528585757271817, + "loss": 1.3894, + "theoretical_loss": 3.506821240835288, + "tokens_seen": 1697710080 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024527582748244735, + "loss": 1.4176, + "theoretical_loss": 3.5068096816660956, + "tokens_seen": 1697775616 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024526579739217653, + "loss": 1.4001, + "theoretical_loss": 3.5067981230680196, + "tokens_seen": 1697841152 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002452557673019057, + "loss": 1.2688, + "theoretical_loss": 3.506786565041012, + "tokens_seen": 1697906688 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002452457372116349, + "loss": 1.3736, + "theoretical_loss": 3.506775007585021, + "tokens_seen": 1697972224 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002452357071213641, + "loss": 1.3271, + "theoretical_loss": 3.5067634506999967, + "tokens_seen": 1698037760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024522567703109325, + "loss": 1.4198, + "theoretical_loss": 3.5067518943858893, + "tokens_seen": 1698103296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024521564694082244, + "loss": 1.4026, + "theoretical_loss": 3.506740338642648, + "tokens_seen": 1698168832 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024520561685055167, + "loss": 1.1659, + "theoretical_loss": 3.5067287834702223, + "tokens_seen": 1698234368 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024519558676028085, + "loss": 1.3317, + "theoretical_loss": 3.506717228868563, + "tokens_seen": 1698299904 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024518555667001003, + "loss": 1.3644, + "theoretical_loss": 3.506705674837619, + "tokens_seen": 1698365440 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002451755265797392, + "loss": 1.4125, + "theoretical_loss": 3.5066941213773406, + "tokens_seen": 1698430976 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002451654964894684, + "loss": 1.409, + "theoretical_loss": 3.5066825684876775, + "tokens_seen": 1698496512 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002451554663991976, + "loss": 1.333, + "theoretical_loss": 3.5066710161685792, + "tokens_seen": 1698562048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024514543630892676, + "loss": 1.2897, + "theoretical_loss": 3.5066594644199958, + "tokens_seen": 1698627584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024513540621865594, + "loss": 1.3213, + "theoretical_loss": 3.506647913241877, + "tokens_seen": 1698693120 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002451253761283852, + "loss": 1.3406, + "theoretical_loss": 3.5066363626341728, + "tokens_seen": 1698758656 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024511534603811435, + "loss": 1.3669, + "theoretical_loss": 3.506624812596833, + "tokens_seen": 1698824192 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024510531594784354, + "loss": 1.3128, + "theoretical_loss": 3.5066132631298075, + "tokens_seen": 1698889728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450952858575727, + "loss": 1.3834, + "theoretical_loss": 3.506601714233045, + "tokens_seen": 1698955264 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450852557673019, + "loss": 1.271, + "theoretical_loss": 3.5065901659064975, + "tokens_seen": 1699020800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024507522567703113, + "loss": 1.3756, + "theoretical_loss": 3.5065786181501126, + "tokens_seen": 1699086336 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450651955867603, + "loss": 1.314, + "theoretical_loss": 3.5065670709638423, + "tokens_seen": 1699151872 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450551654964895, + "loss": 1.4478, + "theoretical_loss": 3.5065555243476347, + "tokens_seen": 1699217408 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450451354062187, + "loss": 1.3194, + "theoretical_loss": 3.5065439783014405, + "tokens_seen": 1699282944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024503510531594786, + "loss": 1.2874, + "theoretical_loss": 3.5065324328252094, + "tokens_seen": 1699348480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00024502507522567704, + "loss": 1.3566, + "theoretical_loss": 3.5065208879188914, + "tokens_seen": 1699414016 + }, + { + "epoch": 0.51, + "learning_rate": 0.0002450150451354062, + "loss": 1.3645, + "theoretical_loss": 3.506509343582436, + "tokens_seen": 1699479552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002450050150451354, + "loss": 1.3117, + "theoretical_loss": 3.506497799815794, + "tokens_seen": 1699545088 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449949849548646, + "loss": 1.317, + "theoretical_loss": 3.5064862566189143, + "tokens_seen": 1699610624 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449849548645938, + "loss": 1.2949, + "theoretical_loss": 3.5064747139917474, + "tokens_seen": 1699676160 + }, + { + "epoch": 0.52, + "learning_rate": 0.000244974924774323, + "loss": 1.4236, + "theoretical_loss": 3.506463171934243, + "tokens_seen": 1699741696 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449648946840522, + "loss": 1.3526, + "theoretical_loss": 3.5064516304463513, + "tokens_seen": 1699807232 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024495486459378136, + "loss": 1.2377, + "theoretical_loss": 3.506440089528022, + "tokens_seen": 1699872768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024494483450351054, + "loss": 1.2671, + "theoretical_loss": 3.5064285491792044, + "tokens_seen": 1699938304 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449348044132397, + "loss": 1.2657, + "theoretical_loss": 3.5064170093998497, + "tokens_seen": 1700003840 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449247743229689, + "loss": 1.412, + "theoretical_loss": 3.506405470189907, + "tokens_seen": 1700069376 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002449147442326981, + "loss": 1.3702, + "theoretical_loss": 3.5063939315493267, + "tokens_seen": 1700134912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024490471414242727, + "loss": 1.3316, + "theoretical_loss": 3.5063823934780585, + "tokens_seen": 1700200448 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448946840521565, + "loss": 1.387, + "theoretical_loss": 3.506370855976052, + "tokens_seen": 1700265984 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448846539618857, + "loss": 1.406, + "theoretical_loss": 3.5063593190432574, + "tokens_seen": 1700331520 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024487462387161486, + "loss": 1.3108, + "theoretical_loss": 3.506347782679626, + "tokens_seen": 1700397056 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024486459378134405, + "loss": 1.2605, + "theoretical_loss": 3.5063362468851054, + "tokens_seen": 1700462592 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448545636910732, + "loss": 1.3659, + "theoretical_loss": 3.5063247116596474, + "tokens_seen": 1700528128 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448445336008024, + "loss": 1.3561, + "theoretical_loss": 3.506313177003202, + "tokens_seen": 1700593664 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 621118, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4468097686767578, + "objective/train/theoretical_loss": 3.5063016429157177, + "objective/train/tokens_used": 1721119200, + "theoretical_loss": 3.5063016429157177, + "tokens_seen": 1700659200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448345035105316, + "loss": 1.3167, + "theoretical_loss": 3.5063016429157177, + "tokens_seen": 1700659200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024482447342026077, + "loss": 1.3496, + "theoretical_loss": 3.506290109397146, + "tokens_seen": 1700724736 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024481444332998995, + "loss": 1.3714, + "theoretical_loss": 3.5062785764474365, + "tokens_seen": 1700790272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002448044132397192, + "loss": 1.3513, + "theoretical_loss": 3.506267044066538, + "tokens_seen": 1700855808 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024479438314944837, + "loss": 1.3399, + "theoretical_loss": 3.506255512254403, + "tokens_seen": 1700921344 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024478435305917755, + "loss": 1.284, + "theoretical_loss": 3.5062439810109796, + "tokens_seen": 1700986880 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024477432296890673, + "loss": 1.2956, + "theoretical_loss": 3.5062324503362183, + "tokens_seen": 1701052416 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002447642928786359, + "loss": 1.2954, + "theoretical_loss": 3.5062209202300694, + "tokens_seen": 1701117952 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002447542627883651, + "loss": 1.4586, + "theoretical_loss": 3.506209390692483, + "tokens_seen": 1701183488 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002447442326980943, + "loss": 1.4271, + "theoretical_loss": 3.5061978617234084, + "tokens_seen": 1701249024 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024473420260782345, + "loss": 1.3037, + "theoretical_loss": 3.5061863333227965, + "tokens_seen": 1701314560 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024472417251755264, + "loss": 1.3315, + "theoretical_loss": 3.5061748054905975, + "tokens_seen": 1701380096 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024471414242728187, + "loss": 1.3768, + "theoretical_loss": 3.5061632782267607, + "tokens_seen": 1701445632 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024470411233701105, + "loss": 1.3162, + "theoretical_loss": 3.506151751531237, + "tokens_seen": 1701511168 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024469408224674023, + "loss": 1.2676, + "theoretical_loss": 3.5061402254039757, + "tokens_seen": 1701576704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002446840521564694, + "loss": 1.293, + "theoretical_loss": 3.5061286998449273, + "tokens_seen": 1701642240 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002446740220661986, + "loss": 1.2693, + "theoretical_loss": 3.506117174854042, + "tokens_seen": 1701707776 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002446639919759278, + "loss": 1.2875, + "theoretical_loss": 3.50610565043127, + "tokens_seen": 1701773312 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024465396188565696, + "loss": 1.3669, + "theoretical_loss": 3.506094126576561, + "tokens_seen": 1701838848 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024464393179538614, + "loss": 1.4296, + "theoretical_loss": 3.5060826032898653, + "tokens_seen": 1701904384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002446339017051154, + "loss": 1.3466, + "theoretical_loss": 3.5060710805711333, + "tokens_seen": 1701969920 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024462387161484456, + "loss": 1.3925, + "theoretical_loss": 3.506059558420314, + "tokens_seen": 1702035456 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024461384152457374, + "loss": 1.4402, + "theoretical_loss": 3.5060480368373597, + "tokens_seen": 1702100992 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002446038114343029, + "loss": 1.3639, + "theoretical_loss": 3.5060365158222186, + "tokens_seen": 1702166528 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445937813440321, + "loss": 1.3664, + "theoretical_loss": 3.5060249953748417, + "tokens_seen": 1702232064 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445837512537613, + "loss": 1.2893, + "theoretical_loss": 3.506013475495179, + "tokens_seen": 1702297600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024457372116349046, + "loss": 1.35, + "theoretical_loss": 3.506001956183181, + "tokens_seen": 1702363136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024456369107321964, + "loss": 1.3025, + "theoretical_loss": 3.5059904374387973, + "tokens_seen": 1702428672 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445536609829488, + "loss": 1.2893, + "theoretical_loss": 3.5059789192619784, + "tokens_seen": 1702494208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024454363089267806, + "loss": 1.3728, + "theoretical_loss": 3.5059674016526747, + "tokens_seen": 1702559744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024453360080240724, + "loss": 1.3547, + "theoretical_loss": 3.5059558846108354, + "tokens_seen": 1702625280 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445235707121364, + "loss": 1.323, + "theoretical_loss": 3.5059443681364124, + "tokens_seen": 1702690816 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445135406218656, + "loss": 1.3722, + "theoretical_loss": 3.5059328522293542, + "tokens_seen": 1702756352 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002445035105315948, + "loss": 1.359, + "theoretical_loss": 3.505921336889612, + "tokens_seen": 1702821888 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024449348044132396, + "loss": 1.3576, + "theoretical_loss": 3.505909822117136, + "tokens_seen": 1702887424 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024448345035105315, + "loss": 1.3246, + "theoretical_loss": 3.5058983079118757, + "tokens_seen": 1702952960 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002444734202607823, + "loss": 1.2321, + "theoretical_loss": 3.505886794273782, + "tokens_seen": 1703018496 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002444633901705115, + "loss": 1.3113, + "theoretical_loss": 3.505875281202805, + "tokens_seen": 1703084032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024445336008024074, + "loss": 1.2794, + "theoretical_loss": 3.5058637686988954, + "tokens_seen": 1703149568 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002444433299899699, + "loss": 1.3561, + "theoretical_loss": 3.5058522567620023, + "tokens_seen": 1703215104 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002444332998996991, + "loss": 1.2616, + "theoretical_loss": 3.505840745392077, + "tokens_seen": 1703280640 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002444232698094283, + "loss": 1.4436, + "theoretical_loss": 3.5058292345890694, + "tokens_seen": 1703346176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024441323971915747, + "loss": 1.2717, + "theoretical_loss": 3.505817724352929, + "tokens_seen": 1703411712 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024440320962888665, + "loss": 1.3361, + "theoretical_loss": 3.505806214683608, + "tokens_seen": 1703477248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024439317953861583, + "loss": 1.4258, + "theoretical_loss": 3.5057947055810548, + "tokens_seen": 1703542784 + }, + { + "epoch": 0.52, + "learning_rate": 0.000244383149448345, + "loss": 1.2832, + "theoretical_loss": 3.5057831970452207, + "tokens_seen": 1703608320 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002443731193580742, + "loss": 1.4556, + "theoretical_loss": 3.5057716890760555, + "tokens_seen": 1703673856 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002443630892678034, + "loss": 1.3605, + "theoretical_loss": 3.5057601816735096, + "tokens_seen": 1703739392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002443530591775326, + "loss": 1.2158, + "theoretical_loss": 3.505748674837534, + "tokens_seen": 1703804928 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002443430290872618, + "loss": 1.4813, + "theoretical_loss": 3.5057371685680776, + "tokens_seen": 1703870464 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 622235, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3637685775756836, + "objective/train/theoretical_loss": 3.505725662865092, + "objective/train/tokens_used": 1724396000, + "theoretical_loss": 3.505725662865092, + "tokens_seen": 1703936000 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024433299899699097, + "loss": 1.3508, + "theoretical_loss": 3.505725662865092, + "tokens_seen": 1703936000 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024432296890672015, + "loss": 1.2589, + "theoretical_loss": 3.5057141577285273, + "tokens_seen": 1704001536 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002443129388164494, + "loss": 1.3435, + "theoretical_loss": 3.505702653158333, + "tokens_seen": 1704067072 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024430290872617857, + "loss": 1.2695, + "theoretical_loss": 3.5056911491544605, + "tokens_seen": 1704132608 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024429287863590775, + "loss": 1.3737, + "theoretical_loss": 3.5056796457168593, + "tokens_seen": 1704198144 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024428284854563693, + "loss": 1.3381, + "theoretical_loss": 3.5056681428454803, + "tokens_seen": 1704263680 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002442728184553661, + "loss": 1.3393, + "theoretical_loss": 3.5056566405402743, + "tokens_seen": 1704329216 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002442627883650953, + "loss": 1.3291, + "theoretical_loss": 3.5056451388011904, + "tokens_seen": 1704394752 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002442527582748245, + "loss": 1.3216, + "theoretical_loss": 3.5056336376281796, + "tokens_seen": 1704460288 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024424272818455365, + "loss": 1.3983, + "theoretical_loss": 3.5056221370211924, + "tokens_seen": 1704525824 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024423269809428284, + "loss": 1.3415, + "theoretical_loss": 3.505610636980179, + "tokens_seen": 1704591360 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024422266800401207, + "loss": 1.325, + "theoretical_loss": 3.50559913750509, + "tokens_seen": 1704656896 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024421263791374125, + "loss": 1.2982, + "theoretical_loss": 3.505587638595876, + "tokens_seen": 1704722432 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024420260782347043, + "loss": 1.3069, + "theoretical_loss": 3.5055761402524865, + "tokens_seen": 1704787968 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002441925777331996, + "loss": 1.3393, + "theoretical_loss": 3.5055646424748725, + "tokens_seen": 1704853504 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002441825476429288, + "loss": 1.3943, + "theoretical_loss": 3.5055531452629847, + "tokens_seen": 1704919040 + }, + { + "epoch": 0.52, + "learning_rate": 0.000244172517552658, + "loss": 1.3083, + "theoretical_loss": 3.505541648616773, + "tokens_seen": 1704984576 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024416248746238716, + "loss": 1.33, + "theoretical_loss": 3.5055301525361884, + "tokens_seen": 1705050112 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024415245737211634, + "loss": 1.2872, + "theoretical_loss": 3.5055186570211805, + "tokens_seen": 1705115648 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002441424272818456, + "loss": 1.3282, + "theoretical_loss": 3.5055071620717007, + "tokens_seen": 1705181184 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024413239719157473, + "loss": 1.4241, + "theoretical_loss": 3.5054956676876987, + "tokens_seen": 1705246720 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002441223671013039, + "loss": 1.2679, + "theoretical_loss": 3.505484173869125, + "tokens_seen": 1705312256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024411233701103312, + "loss": 1.4571, + "theoretical_loss": 3.5054726806159304, + "tokens_seen": 1705377792 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002441023069207623, + "loss": 1.3494, + "theoretical_loss": 3.505461187928065, + "tokens_seen": 1705443328 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024409227683049148, + "loss": 1.3802, + "theoretical_loss": 3.50544969580548, + "tokens_seen": 1705508864 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024408224674022066, + "loss": 1.441, + "theoretical_loss": 3.505438204248125, + "tokens_seen": 1705574400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024407221664994987, + "loss": 1.4149, + "theoretical_loss": 3.505426713255951, + "tokens_seen": 1705639936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024406218655967905, + "loss": 1.3109, + "theoretical_loss": 3.5054152228289084, + "tokens_seen": 1705705472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024405215646940823, + "loss": 1.2205, + "theoretical_loss": 3.5054037329669474, + "tokens_seen": 1705771008 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002440421263791374, + "loss": 1.3464, + "theoretical_loss": 3.505392243670019, + "tokens_seen": 1705836544 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002440320962888666, + "loss": 1.5338, + "theoretical_loss": 3.5053807549380736, + "tokens_seen": 1705902080 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002440220661985958, + "loss": 1.4172, + "theoretical_loss": 3.5053692667710608, + "tokens_seen": 1705967616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024401203610832498, + "loss": 1.2773, + "theoretical_loss": 3.5053577791689325, + "tokens_seen": 1706033152 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024400200601805416, + "loss": 1.337, + "theoretical_loss": 3.5053462921316383, + "tokens_seen": 1706098688 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024399197592778335, + "loss": 1.4228, + "theoretical_loss": 3.505334805659129, + "tokens_seen": 1706164224 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024398194583751255, + "loss": 1.3569, + "theoretical_loss": 3.505323319751356, + "tokens_seen": 1706229760 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024397191574724173, + "loss": 1.2475, + "theoretical_loss": 3.505311834408268, + "tokens_seen": 1706295296 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024396188565697092, + "loss": 1.307, + "theoretical_loss": 3.505300349629817, + "tokens_seen": 1706360832 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002439518555667001, + "loss": 1.3753, + "theoretical_loss": 3.5052888654159533, + "tokens_seen": 1706426368 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024394182547642928, + "loss": 1.3131, + "theoretical_loss": 3.5052773817666267, + "tokens_seen": 1706491904 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024393179538615849, + "loss": 1.2875, + "theoretical_loss": 3.505265898681789, + "tokens_seen": 1706557440 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024392176529588767, + "loss": 1.4207, + "theoretical_loss": 3.5052544161613897, + "tokens_seen": 1706622976 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024391173520561685, + "loss": 1.4169, + "theoretical_loss": 3.50524293420538, + "tokens_seen": 1706688512 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024390170511534603, + "loss": 1.2935, + "theoretical_loss": 3.5052314528137103, + "tokens_seen": 1706754048 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024389167502507524, + "loss": 1.3376, + "theoretical_loss": 3.5052199719863317, + "tokens_seen": 1706819584 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024388164493480442, + "loss": 1.3643, + "theoretical_loss": 3.5052084917231934, + "tokens_seen": 1706885120 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002438716148445336, + "loss": 1.4034, + "theoretical_loss": 3.5051970120242473, + "tokens_seen": 1706950656 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024386158475426278, + "loss": 1.3238, + "theoretical_loss": 3.5051855328894437, + "tokens_seen": 1707016192 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024385155466399196, + "loss": 1.384, + "theoretical_loss": 3.5051740543187333, + "tokens_seen": 1707081728 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024384152457372117, + "loss": 1.2883, + "theoretical_loss": 3.5051625763120664, + "tokens_seen": 1707147264 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 623332, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4895689487457275, + "objective/train/theoretical_loss": 3.5051510988693937, + "objective/train/tokens_used": 1727672800, + "theoretical_loss": 3.5051510988693937, + "tokens_seen": 1707212800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024383149448345035, + "loss": 1.475, + "theoretical_loss": 3.5051510988693937, + "tokens_seen": 1707212800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024382146439317953, + "loss": 1.248, + "theoretical_loss": 3.505139621990666, + "tokens_seen": 1707278336 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024381143430290871, + "loss": 1.27, + "theoretical_loss": 3.505128145675834, + "tokens_seen": 1707343872 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024380140421263792, + "loss": 1.266, + "theoretical_loss": 3.505116669924848, + "tokens_seen": 1707409408 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002437913741223671, + "loss": 1.2891, + "theoretical_loss": 3.5051051947376592, + "tokens_seen": 1707474944 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024378134403209628, + "loss": 1.2788, + "theoretical_loss": 3.5050937201142176, + "tokens_seen": 1707540480 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024377131394182547, + "loss": 1.3115, + "theoretical_loss": 3.5050822460544744, + "tokens_seen": 1707606016 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024376128385155467, + "loss": 1.2763, + "theoretical_loss": 3.5050707725583803, + "tokens_seen": 1707671552 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024375125376128385, + "loss": 1.3862, + "theoretical_loss": 3.5050592996258856, + "tokens_seen": 1707737088 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024374122367101304, + "loss": 1.2546, + "theoretical_loss": 3.5050478272569414, + "tokens_seen": 1707802624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024373119358074222, + "loss": 1.3979, + "theoretical_loss": 3.505036355451498, + "tokens_seen": 1707868160 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002437211634904714, + "loss": 1.3597, + "theoretical_loss": 3.505024884209506, + "tokens_seen": 1707933696 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002437111334002006, + "loss": 1.4266, + "theoretical_loss": 3.5050134135309166, + "tokens_seen": 1707999232 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002437011033099298, + "loss": 1.3275, + "theoretical_loss": 3.50500194341568, + "tokens_seen": 1708064768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024369107321965897, + "loss": 1.301, + "theoretical_loss": 3.5049904738637476, + "tokens_seen": 1708130304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024368104312938815, + "loss": 1.191, + "theoretical_loss": 3.5049790048750697, + "tokens_seen": 1708195840 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024367101303911736, + "loss": 1.3495, + "theoretical_loss": 3.504967536449597, + "tokens_seen": 1708261376 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024366098294884654, + "loss": 1.3064, + "theoretical_loss": 3.50495606858728, + "tokens_seen": 1708326912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024365095285857572, + "loss": 1.3223, + "theoretical_loss": 3.50494460128807, + "tokens_seen": 1708392448 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002436409227683049, + "loss": 1.3281, + "theoretical_loss": 3.5049331345519175, + "tokens_seen": 1708457984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024363089267803408, + "loss": 1.3496, + "theoretical_loss": 3.504921668378773, + "tokens_seen": 1708523520 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002436208625877633, + "loss": 1.3614, + "theoretical_loss": 3.504910202768588, + "tokens_seen": 1708589056 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002436108324974925, + "loss": 1.3259, + "theoretical_loss": 3.5048987377213123, + "tokens_seen": 1708654592 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024360080240722168, + "loss": 1.305, + "theoretical_loss": 3.5048872732368976, + "tokens_seen": 1708720128 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024359077231695086, + "loss": 1.3281, + "theoretical_loss": 3.504875809315294, + "tokens_seen": 1708785664 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024358074222668007, + "loss": 1.4997, + "theoretical_loss": 3.5048643459564524, + "tokens_seen": 1708851200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024357071213640925, + "loss": 1.3856, + "theoretical_loss": 3.5048528831603236, + "tokens_seen": 1708916736 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024356068204613843, + "loss": 1.3605, + "theoretical_loss": 3.5048414209268586, + "tokens_seen": 1708982272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002435506519558676, + "loss": 1.2588, + "theoretical_loss": 3.504829959256009, + "tokens_seen": 1709047808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002435406218655968, + "loss": 1.3384, + "theoretical_loss": 3.5048184981477233, + "tokens_seen": 1709113344 + }, + { + "epoch": 0.52, + "learning_rate": 0.000243530591775326, + "loss": 1.3806, + "theoretical_loss": 3.504807037601955, + "tokens_seen": 1709178880 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024352056168505518, + "loss": 1.4095, + "theoretical_loss": 3.504795577618653, + "tokens_seen": 1709244416 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024351053159478436, + "loss": 1.3079, + "theoretical_loss": 3.504784118197769, + "tokens_seen": 1709309952 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024350050150451355, + "loss": 1.298, + "theoretical_loss": 3.5047726593392534, + "tokens_seen": 1709375488 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024349047141424275, + "loss": 1.5186, + "theoretical_loss": 3.504761201043057, + "tokens_seen": 1709441024 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024348044132397193, + "loss": 1.3813, + "theoretical_loss": 3.5047497433091315, + "tokens_seen": 1709506560 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024347041123370112, + "loss": 1.347, + "theoretical_loss": 3.504738286137427, + "tokens_seen": 1709572096 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002434603811434303, + "loss": 1.3527, + "theoretical_loss": 3.5047268295278946, + "tokens_seen": 1709637632 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024345035105315948, + "loss": 1.256, + "theoretical_loss": 3.504715373480485, + "tokens_seen": 1709703168 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024344032096288869, + "loss": 1.445, + "theoretical_loss": 3.5047039179951494, + "tokens_seen": 1709768704 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024343029087261787, + "loss": 1.4103, + "theoretical_loss": 3.504692463071838, + "tokens_seen": 1709834240 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024342026078234705, + "loss": 1.3353, + "theoretical_loss": 3.5046810087105027, + "tokens_seen": 1709899776 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024341023069207623, + "loss": 1.2842, + "theoretical_loss": 3.5046695549110938, + "tokens_seen": 1709965312 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024340020060180544, + "loss": 1.4278, + "theoretical_loss": 3.504658101673562, + "tokens_seen": 1710030848 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024339017051153462, + "loss": 1.2891, + "theoretical_loss": 3.5046466489978583, + "tokens_seen": 1710096384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002433801404212638, + "loss": 1.3455, + "theoretical_loss": 3.504635196883934, + "tokens_seen": 1710161920 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024337011033099298, + "loss": 1.2584, + "theoretical_loss": 3.50462374533174, + "tokens_seen": 1710227456 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024336008024072216, + "loss": 1.3992, + "theoretical_loss": 3.504612294341227, + "tokens_seen": 1710292992 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024335005015045137, + "loss": 1.3546, + "theoretical_loss": 3.5046008439123453, + "tokens_seen": 1710358528 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024334002006018055, + "loss": 1.3413, + "theoretical_loss": 3.5045893940450465, + "tokens_seen": 1710424064 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 624643, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4121699333190918, + "objective/train/theoretical_loss": 3.504577944739282, + "objective/train/tokens_used": 1730949600, + "theoretical_loss": 3.504577944739282, + "tokens_seen": 1710489600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024332998996990973, + "loss": 1.3551, + "theoretical_loss": 3.504577944739282, + "tokens_seen": 1710489600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024331995987963891, + "loss": 1.3365, + "theoretical_loss": 3.504566495995002, + "tokens_seen": 1710555136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024330992978936812, + "loss": 1.4065, + "theoretical_loss": 3.504555047812158, + "tokens_seen": 1710620672 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002432998996990973, + "loss": 1.2837, + "theoretical_loss": 3.5045436001907007, + "tokens_seen": 1710686208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024328986960882648, + "loss": 1.4559, + "theoretical_loss": 3.5045321531305804, + "tokens_seen": 1710751744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024327983951855567, + "loss": 1.2808, + "theoretical_loss": 3.5045207066317494, + "tokens_seen": 1710817280 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024326980942828487, + "loss": 1.4521, + "theoretical_loss": 3.5045092606941575, + "tokens_seen": 1710882816 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024325977933801405, + "loss": 1.1815, + "theoretical_loss": 3.5044978153177566, + "tokens_seen": 1710948352 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024324974924774324, + "loss": 1.4254, + "theoretical_loss": 3.504486370502497, + "tokens_seen": 1711013888 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024323971915747242, + "loss": 1.4069, + "theoretical_loss": 3.5044749262483297, + "tokens_seen": 1711079424 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002432296890672016, + "loss": 1.3711, + "theoretical_loss": 3.504463482555207, + "tokens_seen": 1711144960 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002432196589769308, + "loss": 1.3375, + "theoretical_loss": 3.504452039423078, + "tokens_seen": 1711210496 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024320962888666, + "loss": 1.3906, + "theoretical_loss": 3.5044405968518944, + "tokens_seen": 1711276032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024319959879638917, + "loss": 1.3975, + "theoretical_loss": 3.5044291548416076, + "tokens_seen": 1711341568 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024318956870611835, + "loss": 1.2709, + "theoretical_loss": 3.5044177133921686, + "tokens_seen": 1711407104 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024317953861584756, + "loss": 1.2825, + "theoretical_loss": 3.504406272503528, + "tokens_seen": 1711472640 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024316950852557674, + "loss": 1.1578, + "theoretical_loss": 3.5043948321756373, + "tokens_seen": 1711538176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024315947843530592, + "loss": 1.3116, + "theoretical_loss": 3.5043833924084473, + "tokens_seen": 1711603712 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002431494483450351, + "loss": 1.3372, + "theoretical_loss": 3.5043719532019093, + "tokens_seen": 1711669248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024313941825476428, + "loss": 1.4633, + "theoretical_loss": 3.504360514555974, + "tokens_seen": 1711734784 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002431293881644935, + "loss": 1.333, + "theoretical_loss": 3.504349076470593, + "tokens_seen": 1711800320 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024311935807422267, + "loss": 1.471, + "theoretical_loss": 3.5043376389457164, + "tokens_seen": 1711865856 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024310932798395185, + "loss": 1.2933, + "theoretical_loss": 3.504326201981296, + "tokens_seen": 1711931392 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024309929789368103, + "loss": 1.4039, + "theoretical_loss": 3.504314765577283, + "tokens_seen": 1711996928 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024308926780341024, + "loss": 1.4071, + "theoretical_loss": 3.504303329733628, + "tokens_seen": 1712062464 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024307923771313942, + "loss": 1.4773, + "theoretical_loss": 3.5042918944502826, + "tokens_seen": 1712128000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002430692076228686, + "loss": 1.3153, + "theoretical_loss": 3.5042804597271973, + "tokens_seen": 1712193536 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024305917753259779, + "loss": 1.3211, + "theoretical_loss": 3.504269025564324, + "tokens_seen": 1712259072 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024304914744232697, + "loss": 1.3295, + "theoretical_loss": 3.504257591961613, + "tokens_seen": 1712324608 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024303911735205618, + "loss": 1.2357, + "theoretical_loss": 3.504246158919016, + "tokens_seen": 1712390144 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024302908726178536, + "loss": 1.2985, + "theoretical_loss": 3.5042347264364837, + "tokens_seen": 1712455680 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024301905717151454, + "loss": 1.377, + "theoretical_loss": 3.5042232945139675, + "tokens_seen": 1712521216 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024300902708124372, + "loss": 1.4441, + "theoretical_loss": 3.5042118631514185, + "tokens_seen": 1712586752 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024299899699097293, + "loss": 1.4356, + "theoretical_loss": 3.504200432348788, + "tokens_seen": 1712652288 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002429889669007021, + "loss": 1.387, + "theoretical_loss": 3.5041890021060267, + "tokens_seen": 1712717824 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002429789368104313, + "loss": 1.4116, + "theoretical_loss": 3.504177572423086, + "tokens_seen": 1712783360 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024296890672016047, + "loss": 1.3199, + "theoretical_loss": 3.5041661432999174, + "tokens_seen": 1712848896 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024295887662988965, + "loss": 1.4019, + "theoretical_loss": 3.504154714736472, + "tokens_seen": 1712914432 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024294884653961886, + "loss": 1.4086, + "theoretical_loss": 3.5041432867327007, + "tokens_seen": 1712979968 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024293881644934804, + "loss": 1.3076, + "theoretical_loss": 3.504131859288554, + "tokens_seen": 1713045504 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024292878635907722, + "loss": 1.3491, + "theoretical_loss": 3.504120432403984, + "tokens_seen": 1713111040 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002429187562688064, + "loss": 1.3663, + "theoretical_loss": 3.5041090060789424, + "tokens_seen": 1713176576 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002429087261785356, + "loss": 1.3518, + "theoretical_loss": 3.504097580313379, + "tokens_seen": 1713242112 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002428986960882648, + "loss": 1.4574, + "theoretical_loss": 3.504086155107246, + "tokens_seen": 1713307648 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024288866599799397, + "loss": 1.3631, + "theoretical_loss": 3.504074730460494, + "tokens_seen": 1713373184 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024287863590772315, + "loss": 1.3839, + "theoretical_loss": 3.504063306373075, + "tokens_seen": 1713438720 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024286860581745234, + "loss": 1.333, + "theoretical_loss": 3.50405188284494, + "tokens_seen": 1713504256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024285857572718157, + "loss": 1.2829, + "theoretical_loss": 3.5040404598760393, + "tokens_seen": 1713569792 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024284854563691075, + "loss": 1.272, + "theoretical_loss": 3.504029037466325, + "tokens_seen": 1713635328 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024283851554663993, + "loss": 1.4924, + "theoretical_loss": 3.504017615615748, + "tokens_seen": 1713700864 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 625897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2701644897460938, + "objective/train/theoretical_loss": 3.50400619432426, + "objective/train/tokens_used": 1734226400, + "theoretical_loss": 3.50400619432426, + "tokens_seen": 1713766400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024282848545636911, + "loss": 1.2725, + "theoretical_loss": 3.50400619432426, + "tokens_seen": 1713766400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024281845536609832, + "loss": 1.313, + "theoretical_loss": 3.5039947735918124, + "tokens_seen": 1713831936 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002428084252758275, + "loss": 1.3435, + "theoretical_loss": 3.5039833534183553, + "tokens_seen": 1713897472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024279839518555668, + "loss": 1.4028, + "theoretical_loss": 3.5039719338038413, + "tokens_seen": 1713963008 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024278836509528587, + "loss": 1.232, + "theoretical_loss": 3.5039605147482202, + "tokens_seen": 1714028544 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024277833500501507, + "loss": 1.2816, + "theoretical_loss": 3.503949096251445, + "tokens_seen": 1714094080 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024276830491474426, + "loss": 1.4232, + "theoretical_loss": 3.5039376783134655, + "tokens_seen": 1714159616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024275827482447344, + "loss": 1.3162, + "theoretical_loss": 3.503926260934234, + "tokens_seen": 1714225152 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024274824473420262, + "loss": 1.2798, + "theoretical_loss": 3.5039148441137016, + "tokens_seen": 1714290688 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002427382146439318, + "loss": 1.3929, + "theoretical_loss": 3.503903427851819, + "tokens_seen": 1714356224 + }, + { + "epoch": 0.52, + "learning_rate": 0.000242728184553661, + "loss": 1.3373, + "theoretical_loss": 3.5038920121485386, + "tokens_seen": 1714421760 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002427181544633902, + "loss": 1.3752, + "theoretical_loss": 3.5038805970038105, + "tokens_seen": 1714487296 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024270812437311937, + "loss": 1.4091, + "theoretical_loss": 3.5038691824175867, + "tokens_seen": 1714552832 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024269809428284855, + "loss": 1.3339, + "theoretical_loss": 3.5038577683898184, + "tokens_seen": 1714618368 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024268806419257776, + "loss": 1.2681, + "theoretical_loss": 3.503846354920457, + "tokens_seen": 1714683904 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024267803410230694, + "loss": 1.3723, + "theoretical_loss": 3.503834942009454, + "tokens_seen": 1714749440 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024266800401203612, + "loss": 1.4661, + "theoretical_loss": 3.50382352965676, + "tokens_seen": 1714814976 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002426579739217653, + "loss": 1.4304, + "theoretical_loss": 3.5038121178623274, + "tokens_seen": 1714880512 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024264794383149448, + "loss": 1.2942, + "theoretical_loss": 3.503800706626107, + "tokens_seen": 1714946048 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002426379137412237, + "loss": 1.3529, + "theoretical_loss": 3.5037892959480494, + "tokens_seen": 1715011584 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024262788365095287, + "loss": 1.3908, + "theoretical_loss": 3.5037778858281077, + "tokens_seen": 1715077120 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024261785356068205, + "loss": 1.3972, + "theoretical_loss": 3.503766476266232, + "tokens_seen": 1715142656 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024260782347041123, + "loss": 1.4784, + "theoretical_loss": 3.503755067262374, + "tokens_seen": 1715208192 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024259779338014044, + "loss": 1.3771, + "theoretical_loss": 3.503743658816485, + "tokens_seen": 1715273728 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024258776328986962, + "loss": 1.5199, + "theoretical_loss": 3.503732250928517, + "tokens_seen": 1715339264 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002425777331995988, + "loss": 1.3593, + "theoretical_loss": 3.5037208435984204, + "tokens_seen": 1715404800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024256770310932799, + "loss": 1.3995, + "theoretical_loss": 3.5037094368261474, + "tokens_seen": 1715470336 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024255767301905717, + "loss": 1.3725, + "theoretical_loss": 3.503698030611649, + "tokens_seen": 1715535872 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024254764292878638, + "loss": 1.3105, + "theoretical_loss": 3.503686624954877, + "tokens_seen": 1715601408 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024253761283851556, + "loss": 1.3482, + "theoretical_loss": 3.503675219855782, + "tokens_seen": 1715666944 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024252758274824474, + "loss": 1.3685, + "theoretical_loss": 3.503663815314317, + "tokens_seen": 1715732480 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024251755265797392, + "loss": 1.3565, + "theoretical_loss": 3.5036524113304317, + "tokens_seen": 1715798016 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024250752256770313, + "loss": 1.3457, + "theoretical_loss": 3.5036410079040783, + "tokens_seen": 1715863552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002424974924774323, + "loss": 1.3398, + "theoretical_loss": 3.5036296050352087, + "tokens_seen": 1715929088 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002424874623871615, + "loss": 1.3629, + "theoretical_loss": 3.5036182027237732, + "tokens_seen": 1715994624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024247743229689067, + "loss": 1.3766, + "theoretical_loss": 3.5036068009697248, + "tokens_seen": 1716060160 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024246740220661985, + "loss": 1.2839, + "theoretical_loss": 3.5035953997730136, + "tokens_seen": 1716125696 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024245737211634906, + "loss": 1.3343, + "theoretical_loss": 3.5035839991335918, + "tokens_seen": 1716191232 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024244734202607824, + "loss": 1.2963, + "theoretical_loss": 3.5035725990514104, + "tokens_seen": 1716256768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024243731193580742, + "loss": 1.3312, + "theoretical_loss": 3.5035611995264215, + "tokens_seen": 1716322304 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002424272818455366, + "loss": 1.36, + "theoretical_loss": 3.503549800558576, + "tokens_seen": 1716387840 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002424172517552658, + "loss": 1.318, + "theoretical_loss": 3.503538402147826, + "tokens_seen": 1716453376 + }, + { + "epoch": 0.52, + "learning_rate": 0.000242407221664995, + "loss": 1.3795, + "theoretical_loss": 3.5035270042941224, + "tokens_seen": 1716518912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024239719157472417, + "loss": 1.3672, + "theoretical_loss": 3.503515606997417, + "tokens_seen": 1716584448 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024238716148445335, + "loss": 1.375, + "theoretical_loss": 3.5035042102576615, + "tokens_seen": 1716649984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024237713139418254, + "loss": 1.2756, + "theoretical_loss": 3.5034928140748067, + "tokens_seen": 1716715520 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024236710130391174, + "loss": 1.3637, + "theoretical_loss": 3.5034814184488052, + "tokens_seen": 1716781056 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024235707121364092, + "loss": 1.3748, + "theoretical_loss": 3.5034700233796077, + "tokens_seen": 1716846592 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002423470411233701, + "loss": 1.2484, + "theoretical_loss": 3.5034586288671656, + "tokens_seen": 1716912128 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002423370110330993, + "loss": 1.2931, + "theoretical_loss": 3.5034472349114316, + "tokens_seen": 1716977664 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 627098, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3342537879943848, + "objective/train/theoretical_loss": 3.5034358415123563, + "objective/train/tokens_used": 1737503200, + "theoretical_loss": 3.5034358415123563, + "tokens_seen": 1717043200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002423269809428285, + "loss": 1.2765, + "theoretical_loss": 3.5034358415123563, + "tokens_seen": 1717043200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024231695085255768, + "loss": 1.3813, + "theoretical_loss": 3.5034244486698913, + "tokens_seen": 1717108736 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024230692076228686, + "loss": 1.3554, + "theoretical_loss": 3.5034130563839887, + "tokens_seen": 1717174272 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024229689067201604, + "loss": 1.2712, + "theoretical_loss": 3.5034016646545996, + "tokens_seen": 1717239808 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024228686058174522, + "loss": 1.369, + "theoretical_loss": 3.503390273481676, + "tokens_seen": 1717305344 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024227683049147443, + "loss": 1.3676, + "theoretical_loss": 3.503378882865168, + "tokens_seen": 1717370880 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002422668004012036, + "loss": 1.3294, + "theoretical_loss": 3.5033674928050296, + "tokens_seen": 1717436416 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002422567703109328, + "loss": 1.3078, + "theoretical_loss": 3.5033561033012104, + "tokens_seen": 1717501952 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024224674022066197, + "loss": 1.3396, + "theoretical_loss": 3.5033447143536636, + "tokens_seen": 1717567488 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024223671013039118, + "loss": 1.3844, + "theoretical_loss": 3.5033333259623394, + "tokens_seen": 1717633024 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024222668004012036, + "loss": 1.3421, + "theoretical_loss": 3.50332193812719, + "tokens_seen": 1717698560 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024221664994984954, + "loss": 1.4199, + "theoretical_loss": 3.503310550848167, + "tokens_seen": 1717764096 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024220661985957872, + "loss": 1.4277, + "theoretical_loss": 3.503299164125222, + "tokens_seen": 1717829632 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024219658976930793, + "loss": 1.3898, + "theoretical_loss": 3.503287777958307, + "tokens_seen": 1717895168 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002421865596790371, + "loss": 1.3647, + "theoretical_loss": 3.503276392347373, + "tokens_seen": 1717960704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002421765295887663, + "loss": 1.2702, + "theoretical_loss": 3.503265007292372, + "tokens_seen": 1718026240 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024216649949849547, + "loss": 1.403, + "theoretical_loss": 3.5032536227932556, + "tokens_seen": 1718091776 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024215646940822466, + "loss": 1.3214, + "theoretical_loss": 3.5032422388499755, + "tokens_seen": 1718157312 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024214643931795386, + "loss": 1.3012, + "theoretical_loss": 3.503230855462483, + "tokens_seen": 1718222848 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024213640922768305, + "loss": 1.2697, + "theoretical_loss": 3.5032194726307306, + "tokens_seen": 1718288384 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024212637913741223, + "loss": 1.3394, + "theoretical_loss": 3.503208090354669, + "tokens_seen": 1718353920 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024211634904714143, + "loss": 1.3109, + "theoretical_loss": 3.5031967086342504, + "tokens_seen": 1718419456 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024210631895687064, + "loss": 1.272, + "theoretical_loss": 3.5031853274694265, + "tokens_seen": 1718484992 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024209628886659982, + "loss": 1.3742, + "theoretical_loss": 3.503173946860149, + "tokens_seen": 1718550528 + }, + { + "epoch": 0.52, + "learning_rate": 0.000242086258776329, + "loss": 1.3921, + "theoretical_loss": 3.5031625668063695, + "tokens_seen": 1718616064 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024207622868605819, + "loss": 1.3462, + "theoretical_loss": 3.5031511873080396, + "tokens_seen": 1718681600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024206619859578737, + "loss": 1.3404, + "theoretical_loss": 3.5031398083651113, + "tokens_seen": 1718747136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024205616850551658, + "loss": 1.2578, + "theoretical_loss": 3.503128429977536, + "tokens_seen": 1718812672 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024204613841524576, + "loss": 1.3413, + "theoretical_loss": 3.5031170521452655, + "tokens_seen": 1718878208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024203610832497494, + "loss": 1.3205, + "theoretical_loss": 3.5031056748682516, + "tokens_seen": 1718943744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024202607823470412, + "loss": 1.2806, + "theoretical_loss": 3.503094298146446, + "tokens_seen": 1719009280 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024201604814443333, + "loss": 1.278, + "theoretical_loss": 3.503082921979801, + "tokens_seen": 1719074816 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002420060180541625, + "loss": 1.4293, + "theoretical_loss": 3.5030715463682673, + "tokens_seen": 1719140352 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002419959879638917, + "loss": 1.4323, + "theoretical_loss": 3.5030601713117973, + "tokens_seen": 1719205888 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024198595787362087, + "loss": 1.3715, + "theoretical_loss": 3.5030487968103428, + "tokens_seen": 1719271424 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024197592778335005, + "loss": 1.3082, + "theoretical_loss": 3.503037422863855, + "tokens_seen": 1719336960 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024196589769307926, + "loss": 1.3639, + "theoretical_loss": 3.5030260494722865, + "tokens_seen": 1719402496 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024195586760280844, + "loss": 1.474, + "theoretical_loss": 3.5030146766355887, + "tokens_seen": 1719468032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024194583751253762, + "loss": 1.4185, + "theoretical_loss": 3.5030033043537125, + "tokens_seen": 1719533568 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002419358074222668, + "loss": 1.2688, + "theoretical_loss": 3.5029919326266112, + "tokens_seen": 1719599104 + }, + { + "epoch": 0.52, + "learning_rate": 0.000241925777331996, + "loss": 1.4031, + "theoretical_loss": 3.502980561454236, + "tokens_seen": 1719664640 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002419157472417252, + "loss": 1.4537, + "theoretical_loss": 3.502969190836538, + "tokens_seen": 1719730176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024190571715145437, + "loss": 1.4032, + "theoretical_loss": 3.50295782077347, + "tokens_seen": 1719795712 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024189568706118355, + "loss": 1.3652, + "theoretical_loss": 3.5029464512649833, + "tokens_seen": 1719861248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024188565697091274, + "loss": 1.3102, + "theoretical_loss": 3.50293508231103, + "tokens_seen": 1719926784 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024187562688064194, + "loss": 1.4616, + "theoretical_loss": 3.502923713911562, + "tokens_seen": 1719992320 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024186559679037113, + "loss": 1.3421, + "theoretical_loss": 3.5029123460665303, + "tokens_seen": 1720057856 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002418555667001003, + "loss": 1.3786, + "theoretical_loss": 3.502900978775888, + "tokens_seen": 1720123392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002418455366098295, + "loss": 1.3827, + "theoretical_loss": 3.502889612039586, + "tokens_seen": 1720188928 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002418355065195587, + "loss": 1.334, + "theoretical_loss": 3.502878245857576, + "tokens_seen": 1720254464 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 628251, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.53982675075531, + "objective/train/theoretical_loss": 3.502866880229811, + "objective/train/tokens_used": 1740780000, + "theoretical_loss": 3.502866880229811, + "tokens_seen": 1720320000 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024182547642928788, + "loss": 1.4191, + "theoretical_loss": 3.502866880229811, + "tokens_seen": 1720320000 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024181544633901706, + "loss": 1.3161, + "theoretical_loss": 3.5028555151562415, + "tokens_seen": 1720385536 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024180541624874624, + "loss": 1.3396, + "theoretical_loss": 3.5028441506368204, + "tokens_seen": 1720451072 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024179538615847542, + "loss": 1.3787, + "theoretical_loss": 3.502832786671499, + "tokens_seen": 1720516608 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024178535606820463, + "loss": 1.3766, + "theoretical_loss": 3.5028214232602295, + "tokens_seen": 1720582144 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002417753259779338, + "loss": 1.3769, + "theoretical_loss": 3.5028100604029637, + "tokens_seen": 1720647680 + }, + { + "epoch": 0.52, + "learning_rate": 0.000241765295887663, + "loss": 1.2747, + "theoretical_loss": 3.5027986980996535, + "tokens_seen": 1720713216 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024175526579739217, + "loss": 1.3779, + "theoretical_loss": 3.5027873363502504, + "tokens_seen": 1720778752 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024174523570712138, + "loss": 1.3352, + "theoretical_loss": 3.502775975154707, + "tokens_seen": 1720844288 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024173520561685056, + "loss": 1.3739, + "theoretical_loss": 3.5027646145129747, + "tokens_seen": 1720909824 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024172517552657974, + "loss": 1.3217, + "theoretical_loss": 3.5027532544250057, + "tokens_seen": 1720975360 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024171514543630892, + "loss": 1.4321, + "theoretical_loss": 3.502741894890751, + "tokens_seen": 1721040896 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024170511534603813, + "loss": 1.4082, + "theoretical_loss": 3.5027305359101644, + "tokens_seen": 1721106432 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002416950852557673, + "loss": 1.3639, + "theoretical_loss": 3.502719177483196, + "tokens_seen": 1721171968 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002416850551654965, + "loss": 1.3088, + "theoretical_loss": 3.502707819609799, + "tokens_seen": 1721237504 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024167502507522567, + "loss": 1.3232, + "theoretical_loss": 3.502696462289925, + "tokens_seen": 1721303040 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024166499498495486, + "loss": 1.3175, + "theoretical_loss": 3.502685105523525, + "tokens_seen": 1721368576 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024165496489468406, + "loss": 1.3814, + "theoretical_loss": 3.5026737493105524, + "tokens_seen": 1721434112 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024164493480441325, + "loss": 1.3612, + "theoretical_loss": 3.502662393650958, + "tokens_seen": 1721499648 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024163490471414243, + "loss": 1.3952, + "theoretical_loss": 3.5026510385446947, + "tokens_seen": 1721565184 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002416248746238716, + "loss": 1.3419, + "theoretical_loss": 3.502639683991714, + "tokens_seen": 1721630720 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024161484453360082, + "loss": 1.4162, + "theoretical_loss": 3.5026283299919676, + "tokens_seen": 1721696256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024160481444333, + "loss": 1.3314, + "theoretical_loss": 3.5026169765454083, + "tokens_seen": 1721761792 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024159478435305918, + "loss": 1.3224, + "theoretical_loss": 3.5026056236519874, + "tokens_seen": 1721827328 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024158475426278836, + "loss": 1.3041, + "theoretical_loss": 3.502594271311657, + "tokens_seen": 1721892864 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024157472417251754, + "loss": 1.4209, + "theoretical_loss": 3.502582919524369, + "tokens_seen": 1721958400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024156469408224675, + "loss": 1.3382, + "theoretical_loss": 3.502571568290076, + "tokens_seen": 1722023936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024155466399197593, + "loss": 1.3906, + "theoretical_loss": 3.50256021760873, + "tokens_seen": 1722089472 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002415446339017051, + "loss": 1.3103, + "theoretical_loss": 3.502548867480282, + "tokens_seen": 1722155008 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002415346038114343, + "loss": 1.3834, + "theoretical_loss": 3.5025375179046847, + "tokens_seen": 1722220544 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002415245737211635, + "loss": 1.283, + "theoretical_loss": 3.5025261688818903, + "tokens_seen": 1722286080 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024151454363089268, + "loss": 1.2866, + "theoretical_loss": 3.502514820411851, + "tokens_seen": 1722351616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024150451354062186, + "loss": 1.3272, + "theoretical_loss": 3.502503472494518, + "tokens_seen": 1722417152 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024149448345035104, + "loss": 1.4179, + "theoretical_loss": 3.5024921251298444, + "tokens_seen": 1722482688 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024148445336008022, + "loss": 1.319, + "theoretical_loss": 3.502480778317781, + "tokens_seen": 1722548224 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024147442326980943, + "loss": 1.4136, + "theoretical_loss": 3.502469432058281, + "tokens_seen": 1722613760 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024146439317953861, + "loss": 1.378, + "theoretical_loss": 3.5024580863512966, + "tokens_seen": 1722679296 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002414543630892678, + "loss": 1.3659, + "theoretical_loss": 3.5024467411967786, + "tokens_seen": 1722744832 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024144433299899698, + "loss": 1.3776, + "theoretical_loss": 3.50243539659468, + "tokens_seen": 1722810368 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024143430290872618, + "loss": 1.3137, + "theoretical_loss": 3.502424052544953, + "tokens_seen": 1722875904 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024142427281845537, + "loss": 1.3849, + "theoretical_loss": 3.502412709047549, + "tokens_seen": 1722941440 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024141424272818455, + "loss": 1.4101, + "theoretical_loss": 3.502401366102421, + "tokens_seen": 1723006976 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024140421263791373, + "loss": 1.3439, + "theoretical_loss": 3.5023900237095202, + "tokens_seen": 1723072512 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002413941825476429, + "loss": 1.3247, + "theoretical_loss": 3.502378681868799, + "tokens_seen": 1723138048 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024138415245737212, + "loss": 1.3495, + "theoretical_loss": 3.50236734058021, + "tokens_seen": 1723203584 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002413741223671013, + "loss": 1.3647, + "theoretical_loss": 3.5023559998437053, + "tokens_seen": 1723269120 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002413640922768305, + "loss": 1.4309, + "theoretical_loss": 3.5023446596592365, + "tokens_seen": 1723334656 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002413540621865597, + "loss": 1.3316, + "theoretical_loss": 3.5023333200267555, + "tokens_seen": 1723400192 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002413440320962889, + "loss": 1.3758, + "theoretical_loss": 3.502321980946215, + "tokens_seen": 1723465728 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024133400200601808, + "loss": 1.4281, + "theoretical_loss": 3.502310642417567, + "tokens_seen": 1723531264 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 629447, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4750317335128784, + "objective/train/theoretical_loss": 3.502299304440764, + "objective/train/tokens_used": 1744056800, + "theoretical_loss": 3.502299304440764, + "tokens_seen": 1723596800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024132397191574726, + "loss": 1.3907, + "theoretical_loss": 3.502299304440764, + "tokens_seen": 1723596800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024131394182547644, + "loss": 1.4119, + "theoretical_loss": 3.5022879670157576, + "tokens_seen": 1723662336 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024130391173520562, + "loss": 1.3786, + "theoretical_loss": 3.5022766301425, + "tokens_seen": 1723727872 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024129388164493483, + "loss": 1.2417, + "theoretical_loss": 3.502265293820944, + "tokens_seen": 1723793408 + }, + { + "epoch": 0.52, + "learning_rate": 0.000241283851554664, + "loss": 1.3219, + "theoretical_loss": 3.5022539580510417, + "tokens_seen": 1723858944 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002412738214643932, + "loss": 1.3726, + "theoretical_loss": 3.5022426228327443, + "tokens_seen": 1723924480 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024126379137412237, + "loss": 1.3466, + "theoretical_loss": 3.502231288166005, + "tokens_seen": 1723990016 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024125376128385158, + "loss": 1.4032, + "theoretical_loss": 3.5022199540507755, + "tokens_seen": 1724055552 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024124373119358076, + "loss": 1.3926, + "theoretical_loss": 3.502208620487008, + "tokens_seen": 1724121088 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024123370110330994, + "loss": 1.2919, + "theoretical_loss": 3.5021972874746554, + "tokens_seen": 1724186624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024122367101303912, + "loss": 1.3696, + "theoretical_loss": 3.502185955013669, + "tokens_seen": 1724252160 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024121364092276833, + "loss": 1.3005, + "theoretical_loss": 3.5021746231040014, + "tokens_seen": 1724317696 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002412036108324975, + "loss": 1.4053, + "theoretical_loss": 3.5021632917456045, + "tokens_seen": 1724383232 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002411935807422267, + "loss": 1.3081, + "theoretical_loss": 3.502151960938431, + "tokens_seen": 1724448768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024118355065195587, + "loss": 1.3935, + "theoretical_loss": 3.502140630682433, + "tokens_seen": 1724514304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024117352056168506, + "loss": 1.2996, + "theoretical_loss": 3.502129300977563, + "tokens_seen": 1724579840 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024116349047141426, + "loss": 1.3788, + "theoretical_loss": 3.5021179718237727, + "tokens_seen": 1724645376 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024115346038114345, + "loss": 1.3329, + "theoretical_loss": 3.5021066432210146, + "tokens_seen": 1724710912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024114343029087263, + "loss": 1.5195, + "theoretical_loss": 3.5020953151692416, + "tokens_seen": 1724776448 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002411334002006018, + "loss": 1.2427, + "theoretical_loss": 3.5020839876684047, + "tokens_seen": 1724841984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024112337011033102, + "loss": 1.3386, + "theoretical_loss": 3.502072660718457, + "tokens_seen": 1724907520 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002411133400200602, + "loss": 1.4814, + "theoretical_loss": 3.5020613343193507, + "tokens_seen": 1724973056 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024110330992978938, + "loss": 1.3627, + "theoretical_loss": 3.5020500084710378, + "tokens_seen": 1725038592 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024109327983951856, + "loss": 1.3749, + "theoretical_loss": 3.502038683173471, + "tokens_seen": 1725104128 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024108324974924774, + "loss": 1.3854, + "theoretical_loss": 3.502027358426602, + "tokens_seen": 1725169664 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024107321965897695, + "loss": 1.3056, + "theoretical_loss": 3.502016034230384, + "tokens_seen": 1725235200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024106318956870613, + "loss": 1.3771, + "theoretical_loss": 3.5020047105847687, + "tokens_seen": 1725300736 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002410531594784353, + "loss": 1.4687, + "theoretical_loss": 3.5019933874897085, + "tokens_seen": 1725366272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002410431293881645, + "loss": 1.4729, + "theoretical_loss": 3.5019820649451554, + "tokens_seen": 1725431808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002410330992978937, + "loss": 1.3074, + "theoretical_loss": 3.5019707429510625, + "tokens_seen": 1725497344 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024102306920762288, + "loss": 1.3779, + "theoretical_loss": 3.5019594215073813, + "tokens_seen": 1725562880 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024101303911735206, + "loss": 1.3523, + "theoretical_loss": 3.501948100614065, + "tokens_seen": 1725628416 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024100300902708124, + "loss": 1.3874, + "theoretical_loss": 3.5019367802710653, + "tokens_seen": 1725693952 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024099297893681042, + "loss": 1.2737, + "theoretical_loss": 3.5019254604783345, + "tokens_seen": 1725759488 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024098294884653963, + "loss": 1.3286, + "theoretical_loss": 3.5019141412358255, + "tokens_seen": 1725825024 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024097291875626881, + "loss": 1.5654, + "theoretical_loss": 3.5019028225434905, + "tokens_seen": 1725890560 + }, + { + "epoch": 0.52, + "learning_rate": 0.000240962888665998, + "loss": 1.3452, + "theoretical_loss": 3.5018915044012813, + "tokens_seen": 1725956096 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024095285857572718, + "loss": 1.3517, + "theoretical_loss": 3.5018801868091507, + "tokens_seen": 1726021632 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024094282848545638, + "loss": 1.3677, + "theoretical_loss": 3.5018688697670513, + "tokens_seen": 1726087168 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024093279839518557, + "loss": 1.4761, + "theoretical_loss": 3.501857553274935, + "tokens_seen": 1726152704 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024092276830491475, + "loss": 1.4268, + "theoretical_loss": 3.501846237332755, + "tokens_seen": 1726218240 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024091273821464393, + "loss": 1.3792, + "theoretical_loss": 3.5018349219404628, + "tokens_seen": 1726283776 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002409027081243731, + "loss": 1.3684, + "theoretical_loss": 3.501823607098011, + "tokens_seen": 1726349312 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024089267803410232, + "loss": 1.3701, + "theoretical_loss": 3.5018122928053526, + "tokens_seen": 1726414848 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002408826479438315, + "loss": 1.3026, + "theoretical_loss": 3.5018009790624394, + "tokens_seen": 1726480384 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024087261785356068, + "loss": 1.4556, + "theoretical_loss": 3.5017896658692234, + "tokens_seen": 1726545920 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024086258776328986, + "loss": 1.3651, + "theoretical_loss": 3.5017783532256583, + "tokens_seen": 1726611456 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024085255767301907, + "loss": 1.4045, + "theoretical_loss": 3.501767041131696, + "tokens_seen": 1726676992 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024084252758274825, + "loss": 1.4007, + "theoretical_loss": 3.5017557295872885, + "tokens_seen": 1726742528 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024083249749247743, + "loss": 1.3384, + "theoretical_loss": 3.5017444185923887, + "tokens_seen": 1726808064 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 630619, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4317607879638672, + "objective/train/theoretical_loss": 3.501733108146949, + "objective/train/tokens_used": 1747333600, + "theoretical_loss": 3.501733108146949, + "tokens_seen": 1726873600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002408224674022066, + "loss": 1.4525, + "theoretical_loss": 3.501733108146949, + "tokens_seen": 1726873600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002408124373119358, + "loss": 1.3807, + "theoretical_loss": 3.5017217982509212, + "tokens_seen": 1726939136 + }, + { + "epoch": 0.52, + "learning_rate": 0.000240802407221665, + "loss": 1.3026, + "theoretical_loss": 3.5017104889042585, + "tokens_seen": 1727004672 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024079237713139418, + "loss": 1.234, + "theoretical_loss": 3.5016991801069137, + "tokens_seen": 1727070208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024078234704112336, + "loss": 1.4708, + "theoretical_loss": 3.501687871858839, + "tokens_seen": 1727135744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024077231695085254, + "loss": 1.4406, + "theoretical_loss": 3.501676564159986, + "tokens_seen": 1727201280 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024076228686058175, + "loss": 1.4114, + "theoretical_loss": 3.501665257010308, + "tokens_seen": 1727266816 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024075225677031093, + "loss": 1.3048, + "theoretical_loss": 3.5016539504097572, + "tokens_seen": 1727332352 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024074222668004012, + "loss": 1.2722, + "theoretical_loss": 3.501642644358286, + "tokens_seen": 1727397888 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002407321965897693, + "loss": 1.3949, + "theoretical_loss": 3.501631338855848, + "tokens_seen": 1727463424 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024072216649949848, + "loss": 1.3576, + "theoretical_loss": 3.5016200339023946, + "tokens_seen": 1727528960 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024071213640922769, + "loss": 1.4203, + "theoretical_loss": 3.501608729497878, + "tokens_seen": 1727594496 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024070210631895687, + "loss": 1.3254, + "theoretical_loss": 3.5015974256422515, + "tokens_seen": 1727660032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024069207622868605, + "loss": 1.3179, + "theoretical_loss": 3.501586122335468, + "tokens_seen": 1727725568 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024068204613841523, + "loss": 1.3341, + "theoretical_loss": 3.501574819577479, + "tokens_seen": 1727791104 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024067201604814444, + "loss": 1.331, + "theoretical_loss": 3.5015635173682376, + "tokens_seen": 1727856640 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024066198595787362, + "loss": 1.3437, + "theoretical_loss": 3.5015522157076964, + "tokens_seen": 1727922176 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002406519558676028, + "loss": 1.2981, + "theoretical_loss": 3.5015409145958074, + "tokens_seen": 1727987712 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024064192577733198, + "loss": 1.3564, + "theoretical_loss": 3.5015296140325236, + "tokens_seen": 1728053248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024063189568706116, + "loss": 1.4753, + "theoretical_loss": 3.5015183140177975, + "tokens_seen": 1728118784 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024062186559679037, + "loss": 1.3464, + "theoretical_loss": 3.5015070145515823, + "tokens_seen": 1728184320 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024061183550651958, + "loss": 1.306, + "theoretical_loss": 3.5014957156338293, + "tokens_seen": 1728249856 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024060180541624876, + "loss": 1.2932, + "theoretical_loss": 3.501484417264492, + "tokens_seen": 1728315392 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024059177532597794, + "loss": 1.4067, + "theoretical_loss": 3.5014731194435225, + "tokens_seen": 1728380928 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024058174523570715, + "loss": 1.4285, + "theoretical_loss": 3.501461822170874, + "tokens_seen": 1728446464 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024057171514543633, + "loss": 1.3613, + "theoretical_loss": 3.5014505254464985, + "tokens_seen": 1728512000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002405616850551655, + "loss": 1.275, + "theoretical_loss": 3.501439229270349, + "tokens_seen": 1728577536 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002405516549648947, + "loss": 1.2204, + "theoretical_loss": 3.5014279336423777, + "tokens_seen": 1728643072 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002405416248746239, + "loss": 1.3869, + "theoretical_loss": 3.501416638562538, + "tokens_seen": 1728708608 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024053159478435308, + "loss": 1.4136, + "theoretical_loss": 3.501405344030781, + "tokens_seen": 1728774144 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024052156469408226, + "loss": 1.2267, + "theoretical_loss": 3.501394050047061, + "tokens_seen": 1728839680 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024051153460381144, + "loss": 1.3585, + "theoretical_loss": 3.5013827566113296, + "tokens_seen": 1728905216 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024050150451354062, + "loss": 1.2766, + "theoretical_loss": 3.5013714637235402, + "tokens_seen": 1728970752 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024049147442326983, + "loss": 1.3728, + "theoretical_loss": 3.501360171383645, + "tokens_seen": 1729036288 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024048144433299901, + "loss": 1.374, + "theoretical_loss": 3.5013488795915966, + "tokens_seen": 1729101824 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002404714142427282, + "loss": 1.3645, + "theoretical_loss": 3.5013375883473477, + "tokens_seen": 1729167360 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024046138415245738, + "loss": 1.2327, + "theoretical_loss": 3.501326297650851, + "tokens_seen": 1729232896 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024045135406218658, + "loss": 1.3114, + "theoretical_loss": 3.5013150075020594, + "tokens_seen": 1729298432 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024044132397191577, + "loss": 1.3137, + "theoretical_loss": 3.5013037179009245, + "tokens_seen": 1729363968 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024043129388164495, + "loss": 1.3529, + "theoretical_loss": 3.5012924288474006, + "tokens_seen": 1729429504 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024042126379137413, + "loss": 1.4115, + "theoretical_loss": 3.5012811403414394, + "tokens_seen": 1729495040 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002404112337011033, + "loss": 1.324, + "theoretical_loss": 3.501269852382994, + "tokens_seen": 1729560576 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024040120361083252, + "loss": 1.3938, + "theoretical_loss": 3.501258564972017, + "tokens_seen": 1729626112 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002403911735205617, + "loss": 1.3721, + "theoretical_loss": 3.501247278108461, + "tokens_seen": 1729691648 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024038114343029088, + "loss": 1.3765, + "theoretical_loss": 3.501235991792279, + "tokens_seen": 1729757184 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024037111334002006, + "loss": 1.3094, + "theoretical_loss": 3.5012247060234225, + "tokens_seen": 1729822720 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024036108324974927, + "loss": 1.4328, + "theoretical_loss": 3.5012134208018457, + "tokens_seen": 1729888256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024035105315947845, + "loss": 1.3023, + "theoretical_loss": 3.501202136127501, + "tokens_seen": 1729953792 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024034102306920763, + "loss": 1.194, + "theoretical_loss": 3.5011908520003407, + "tokens_seen": 1730019328 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002403309929789368, + "loss": 1.3472, + "theoretical_loss": 3.501179568420318, + "tokens_seen": 1730084864 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 631803, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4235368967056274, + "objective/train/theoretical_loss": 3.5011682853873856, + "objective/train/tokens_used": 1750610400, + "theoretical_loss": 3.5011682853873856, + "tokens_seen": 1730150400 + }, + { + "epoch": 0.52, + "learning_rate": 0.000240320962888666, + "loss": 1.3791, + "theoretical_loss": 3.5011682853873856, + "tokens_seen": 1730150400 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002403109327983952, + "loss": 1.4311, + "theoretical_loss": 3.501157002901495, + "tokens_seen": 1730215936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024030090270812438, + "loss": 1.4297, + "theoretical_loss": 3.501145720962601, + "tokens_seen": 1730281472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024029087261785356, + "loss": 1.3889, + "theoretical_loss": 3.5011344395706554, + "tokens_seen": 1730347008 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024028084252758274, + "loss": 1.4004, + "theoretical_loss": 3.501123158725611, + "tokens_seen": 1730412544 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024027081243731195, + "loss": 1.3792, + "theoretical_loss": 3.5011118784274204, + "tokens_seen": 1730478080 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024026078234704113, + "loss": 1.3573, + "theoretical_loss": 3.501100598676036, + "tokens_seen": 1730543616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024025075225677032, + "loss": 1.2524, + "theoretical_loss": 3.501089319471412, + "tokens_seen": 1730609152 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002402407221664995, + "loss": 1.4513, + "theoretical_loss": 3.5010780408135, + "tokens_seen": 1730674688 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024023069207622868, + "loss": 1.4127, + "theoretical_loss": 3.501066762702253, + "tokens_seen": 1730740224 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024022066198595789, + "loss": 1.2134, + "theoretical_loss": 3.501055485137624, + "tokens_seen": 1730805760 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024021063189568707, + "loss": 1.417, + "theoretical_loss": 3.5010442081195654, + "tokens_seen": 1730871296 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024020060180541625, + "loss": 1.4023, + "theoretical_loss": 3.5010329316480306, + "tokens_seen": 1730936832 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024019057171514543, + "loss": 1.339, + "theoretical_loss": 3.501021655722972, + "tokens_seen": 1731002368 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024018054162487464, + "loss": 1.3373, + "theoretical_loss": 3.501010380344343, + "tokens_seen": 1731067904 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024017051153460382, + "loss": 1.4103, + "theoretical_loss": 3.5009991055120957, + "tokens_seen": 1731133440 + }, + { + "epoch": 0.52, + "learning_rate": 0.000240160481444333, + "loss": 1.4945, + "theoretical_loss": 3.5009878312261833, + "tokens_seen": 1731198976 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024015045135406218, + "loss": 1.3522, + "theoretical_loss": 3.500976557486559, + "tokens_seen": 1731264512 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024014042126379136, + "loss": 1.3581, + "theoretical_loss": 3.5009652842931747, + "tokens_seen": 1731330048 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024013039117352057, + "loss": 1.3927, + "theoretical_loss": 3.5009540116459843, + "tokens_seen": 1731395584 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024012036108324975, + "loss": 1.4439, + "theoretical_loss": 3.5009427395449397, + "tokens_seen": 1731461120 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024011033099297893, + "loss": 1.3879, + "theoretical_loss": 3.5009314679899948, + "tokens_seen": 1731526656 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024010030090270811, + "loss": 1.3517, + "theoretical_loss": 3.5009201969811015, + "tokens_seen": 1731592192 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024009027081243732, + "loss": 1.2705, + "theoretical_loss": 3.500908926518213, + "tokens_seen": 1731657728 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002400802407221665, + "loss": 1.3098, + "theoretical_loss": 3.500897656601283, + "tokens_seen": 1731723264 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024007021063189568, + "loss": 1.3086, + "theoretical_loss": 3.5008863872302634, + "tokens_seen": 1731788800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024006018054162487, + "loss": 1.3839, + "theoretical_loss": 3.5008751184051072, + "tokens_seen": 1731854336 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024005015045135407, + "loss": 1.4612, + "theoretical_loss": 3.500863850125768, + "tokens_seen": 1731919872 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024004012036108325, + "loss": 1.4163, + "theoretical_loss": 3.5008525823921977, + "tokens_seen": 1731985408 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024003009027081244, + "loss": 1.3202, + "theoretical_loss": 3.5008413152043496, + "tokens_seen": 1732050944 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024002006018054162, + "loss": 1.3968, + "theoretical_loss": 3.500830048562177, + "tokens_seen": 1732116480 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002400100300902708, + "loss": 1.3383, + "theoretical_loss": 3.5008187824656325, + "tokens_seen": 1732182016 + }, + { + "epoch": 0.52, + "learning_rate": 0.00024, + "loss": 1.3516, + "theoretical_loss": 3.5008075169146693, + "tokens_seen": 1732247552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0002399899699097292, + "loss": 1.3509, + "theoretical_loss": 3.50079625190924, + "tokens_seen": 1732313088 + }, + { + "epoch": 0.52, + "learning_rate": 0.00023997993981945837, + "loss": 1.3151, + "theoretical_loss": 3.500784987449298, + "tokens_seen": 1732378624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00023996990972918755, + "loss": 1.2872, + "theoretical_loss": 3.5007737235347953, + "tokens_seen": 1732444160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023995987963891676, + "loss": 1.3409, + "theoretical_loss": 3.500762460165686, + "tokens_seen": 1732509696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023994984954864594, + "loss": 1.3092, + "theoretical_loss": 3.5007511973419225, + "tokens_seen": 1732575232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023993981945837512, + "loss": 1.3151, + "theoretical_loss": 3.5007399350634576, + "tokens_seen": 1732640768 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002399297893681043, + "loss": 1.2646, + "theoretical_loss": 3.5007286733302445, + "tokens_seen": 1732706304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023991975927783348, + "loss": 1.2513, + "theoretical_loss": 3.5007174121422366, + "tokens_seen": 1732771840 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002399097291875627, + "loss": 1.3308, + "theoretical_loss": 3.5007061514993865, + "tokens_seen": 1732837376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023989969909729187, + "loss": 1.4239, + "theoretical_loss": 3.500694891401647, + "tokens_seen": 1732902912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023988966900702105, + "loss": 1.3471, + "theoretical_loss": 3.500683631848971, + "tokens_seen": 1732968448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023987963891675023, + "loss": 1.4534, + "theoretical_loss": 3.5006723728413123, + "tokens_seen": 1733033984 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023986960882647944, + "loss": 1.4649, + "theoretical_loss": 3.500661114378623, + "tokens_seen": 1733099520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023985957873620865, + "loss": 1.3452, + "theoretical_loss": 3.500649856460856, + "tokens_seen": 1733165056 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023984954864593783, + "loss": 1.4726, + "theoretical_loss": 3.5006385990879663, + "tokens_seen": 1733230592 + }, + { + "epoch": 0.53, + "learning_rate": 0.000239839518555667, + "loss": 1.3574, + "theoretical_loss": 3.500627342259904, + "tokens_seen": 1733296128 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002398294884653962, + "loss": 1.3125, + "theoretical_loss": 3.5006160859766244, + "tokens_seen": 1733361664 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 632954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3418610095977783, + "objective/train/theoretical_loss": 3.5006048302380792, + "objective/train/tokens_used": 1753887200, + "theoretical_loss": 3.5006048302380792, + "tokens_seen": 1733427200 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002398194583751254, + "loss": 1.3471, + "theoretical_loss": 3.5006048302380792, + "tokens_seen": 1733427200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023980942828485458, + "loss": 1.3348, + "theoretical_loss": 3.5005935750442223, + "tokens_seen": 1733492736 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023979939819458376, + "loss": 1.3595, + "theoretical_loss": 3.5005823203950066, + "tokens_seen": 1733558272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023978936810431295, + "loss": 1.3141, + "theoretical_loss": 3.5005710662903846, + "tokens_seen": 1733623808 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023977933801404215, + "loss": 1.3297, + "theoretical_loss": 3.50055981273031, + "tokens_seen": 1733689344 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023976930792377133, + "loss": 1.3202, + "theoretical_loss": 3.500548559714735, + "tokens_seen": 1733754880 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023975927783350052, + "loss": 1.4291, + "theoretical_loss": 3.500537307243614, + "tokens_seen": 1733820416 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002397492477432297, + "loss": 1.2644, + "theoretical_loss": 3.5005260553168993, + "tokens_seen": 1733885952 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023973921765295888, + "loss": 1.3772, + "theoretical_loss": 3.5005148039345437, + "tokens_seen": 1733951488 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023972918756268809, + "loss": 1.3704, + "theoretical_loss": 3.500503553096501, + "tokens_seen": 1734017024 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023971915747241727, + "loss": 1.2668, + "theoretical_loss": 3.5004923028027237, + "tokens_seen": 1734082560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023970912738214645, + "loss": 1.4293, + "theoretical_loss": 3.500481053053165, + "tokens_seen": 1734148096 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023969909729187563, + "loss": 1.278, + "theoretical_loss": 3.5004698038477784, + "tokens_seen": 1734213632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023968906720160484, + "loss": 1.3645, + "theoretical_loss": 3.500458555186517, + "tokens_seen": 1734279168 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023967903711133402, + "loss": 1.4576, + "theoretical_loss": 3.500447307069333, + "tokens_seen": 1734344704 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002396690070210632, + "loss": 1.299, + "theoretical_loss": 3.500436059496181, + "tokens_seen": 1734410240 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023965897693079238, + "loss": 1.3345, + "theoretical_loss": 3.5004248124670125, + "tokens_seen": 1734475776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023964894684052156, + "loss": 1.3897, + "theoretical_loss": 3.5004135659817823, + "tokens_seen": 1734541312 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023963891675025077, + "loss": 1.3939, + "theoretical_loss": 3.500402320040442, + "tokens_seen": 1734606848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023962888665997995, + "loss": 1.3732, + "theoretical_loss": 3.5003910746429456, + "tokens_seen": 1734672384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023961885656970913, + "loss": 1.3131, + "theoretical_loss": 3.5003798297892468, + "tokens_seen": 1734737920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023960882647943831, + "loss": 1.3222, + "theoretical_loss": 3.500368585479297, + "tokens_seen": 1734803456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023959879638916752, + "loss": 1.4689, + "theoretical_loss": 3.500357341713051, + "tokens_seen": 1734868992 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002395887662988967, + "loss": 1.4797, + "theoretical_loss": 3.5003460984904615, + "tokens_seen": 1734934528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023957873620862588, + "loss": 1.3452, + "theoretical_loss": 3.5003348558114817, + "tokens_seen": 1735000064 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023956870611835507, + "loss": 1.3579, + "theoretical_loss": 3.5003236136760645, + "tokens_seen": 1735065600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023955867602808427, + "loss": 1.3847, + "theoretical_loss": 3.5003123720841636, + "tokens_seen": 1735131136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023954864593781345, + "loss": 1.3608, + "theoretical_loss": 3.500301131035731, + "tokens_seen": 1735196672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023953861584754264, + "loss": 1.3319, + "theoretical_loss": 3.5002898905307216, + "tokens_seen": 1735262208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023952858575727182, + "loss": 1.4284, + "theoretical_loss": 3.5002786505690873, + "tokens_seen": 1735327744 + }, + { + "epoch": 0.53, + "learning_rate": 0.000239518555667001, + "loss": 1.3835, + "theoretical_loss": 3.5002674111507823, + "tokens_seen": 1735393280 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002395085255767302, + "loss": 1.3826, + "theoretical_loss": 3.5002561722757592, + "tokens_seen": 1735458816 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002394984954864594, + "loss": 1.3913, + "theoretical_loss": 3.500244933943971, + "tokens_seen": 1735524352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023948846539618857, + "loss": 1.3674, + "theoretical_loss": 3.5002336961553713, + "tokens_seen": 1735589888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023947843530591775, + "loss": 1.3169, + "theoretical_loss": 3.5002224589099136, + "tokens_seen": 1735655424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023946840521564696, + "loss": 1.3161, + "theoretical_loss": 3.5002112222075508, + "tokens_seen": 1735720960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023945837512537614, + "loss": 1.3422, + "theoretical_loss": 3.5001999860482362, + "tokens_seen": 1735786496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023944834503510532, + "loss": 1.3704, + "theoretical_loss": 3.500188750431923, + "tokens_seen": 1735852032 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002394383149448345, + "loss": 1.3501, + "theoretical_loss": 3.5001775153585646, + "tokens_seen": 1735917568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023942828485456368, + "loss": 1.4517, + "theoretical_loss": 3.5001662808281138, + "tokens_seen": 1735983104 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002394182547642929, + "loss": 1.365, + "theoretical_loss": 3.5001550468405247, + "tokens_seen": 1736048640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023940822467402207, + "loss": 1.3041, + "theoretical_loss": 3.5001438133957494, + "tokens_seen": 1736114176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023939819458375125, + "loss": 1.3939, + "theoretical_loss": 3.5001325804937427, + "tokens_seen": 1736179712 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023938816449348043, + "loss": 1.4062, + "theoretical_loss": 3.500121348134457, + "tokens_seen": 1736245248 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023937813440320964, + "loss": 1.2585, + "theoretical_loss": 3.5001101163178454, + "tokens_seen": 1736310784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023936810431293882, + "loss": 1.3523, + "theoretical_loss": 3.5000988850438617, + "tokens_seen": 1736376320 + }, + { + "epoch": 0.53, + "learning_rate": 0.000239358074222668, + "loss": 1.4097, + "theoretical_loss": 3.500087654312459, + "tokens_seen": 1736441856 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023934804413239719, + "loss": 1.3061, + "theoretical_loss": 3.5000764241235904, + "tokens_seen": 1736507392 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023933801404212637, + "loss": 1.2965, + "theoretical_loss": 3.500065194477209, + "tokens_seen": 1736572928 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023932798395185557, + "loss": 1.3743, + "theoretical_loss": 3.5000539653732696, + "tokens_seen": 1736638464 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 634251, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3751026391983032, + "objective/train/theoretical_loss": 3.5000427368117233, + "objective/train/tokens_used": 1757164000, + "theoretical_loss": 3.5000427368117233, + "tokens_seen": 1736704000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023931795386158476, + "loss": 1.4464, + "theoretical_loss": 3.5000427368117233, + "tokens_seen": 1736704000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023930792377131394, + "loss": 1.3547, + "theoretical_loss": 3.5000315087925253, + "tokens_seen": 1736769536 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023929789368104312, + "loss": 1.4029, + "theoretical_loss": 3.5000202813156287, + "tokens_seen": 1736835072 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023928786359077233, + "loss": 1.3839, + "theoretical_loss": 3.5000090543809854, + "tokens_seen": 1736900608 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002392778335005015, + "loss": 1.4183, + "theoretical_loss": 3.4999978279885506, + "tokens_seen": 1736966144 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002392678034102307, + "loss": 1.3307, + "theoretical_loss": 3.4999866021382764, + "tokens_seen": 1737031680 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023925777331995987, + "loss": 1.4089, + "theoretical_loss": 3.4999753768301165, + "tokens_seen": 1737097216 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023924774322968905, + "loss": 1.3353, + "theoretical_loss": 3.499964152064025, + "tokens_seen": 1737162752 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023923771313941826, + "loss": 1.4075, + "theoretical_loss": 3.4999529278399537, + "tokens_seen": 1737228288 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023922768304914744, + "loss": 1.4057, + "theoretical_loss": 3.499941704157857, + "tokens_seen": 1737293824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023921765295887662, + "loss": 1.376, + "theoretical_loss": 3.4999304810176888, + "tokens_seen": 1737359360 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002392076228686058, + "loss": 1.3543, + "theoretical_loss": 3.499919258419401, + "tokens_seen": 1737424896 + }, + { + "epoch": 0.53, + "learning_rate": 0.000239197592778335, + "loss": 1.3098, + "theoretical_loss": 3.4999080363629487, + "tokens_seen": 1737490432 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002391875626880642, + "loss": 1.2176, + "theoretical_loss": 3.4998968148482845, + "tokens_seen": 1737555968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023917753259779337, + "loss": 1.2987, + "theoretical_loss": 3.4998855938753612, + "tokens_seen": 1737621504 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023916750250752255, + "loss": 1.3792, + "theoretical_loss": 3.499874373444133, + "tokens_seen": 1737687040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023915747241725174, + "loss": 1.4154, + "theoretical_loss": 3.4998631535545535, + "tokens_seen": 1737752576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023914744232698094, + "loss": 1.4261, + "theoretical_loss": 3.4998519342065753, + "tokens_seen": 1737818112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023913741223671012, + "loss": 1.3118, + "theoretical_loss": 3.4998407154001523, + "tokens_seen": 1737883648 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002391273821464393, + "loss": 1.265, + "theoretical_loss": 3.499829497135238, + "tokens_seen": 1737949184 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002391173520561685, + "loss": 1.3527, + "theoretical_loss": 3.4998182794117856, + "tokens_seen": 1738014720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023910732196589772, + "loss": 1.4188, + "theoretical_loss": 3.499807062229749, + "tokens_seen": 1738080256 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002390972918756269, + "loss": 1.3351, + "theoretical_loss": 3.4997958455890816, + "tokens_seen": 1738145792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023908726178535608, + "loss": 1.337, + "theoretical_loss": 3.4997846294897363, + "tokens_seen": 1738211328 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023907723169508527, + "loss": 1.3602, + "theoretical_loss": 3.499773413931667, + "tokens_seen": 1738276864 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023906720160481447, + "loss": 1.3834, + "theoretical_loss": 3.4997621989148273, + "tokens_seen": 1738342400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023905717151454365, + "loss": 1.2846, + "theoretical_loss": 3.49975098443917, + "tokens_seen": 1738407936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023904714142427284, + "loss": 1.3412, + "theoretical_loss": 3.4997397705046493, + "tokens_seen": 1738473472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023903711133400202, + "loss": 1.3449, + "theoretical_loss": 3.4997285571112178, + "tokens_seen": 1738539008 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002390270812437312, + "loss": 1.2908, + "theoretical_loss": 3.4997173442588307, + "tokens_seen": 1738604544 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002390170511534604, + "loss": 1.299, + "theoretical_loss": 3.4997061319474394, + "tokens_seen": 1738670080 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002390070210631896, + "loss": 1.2823, + "theoretical_loss": 3.4996949201769993, + "tokens_seen": 1738735616 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023899699097291877, + "loss": 1.3643, + "theoretical_loss": 3.4996837089474626, + "tokens_seen": 1738801152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023898696088264795, + "loss": 1.3145, + "theoretical_loss": 3.4996724982587835, + "tokens_seen": 1738866688 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023897693079237716, + "loss": 1.2694, + "theoretical_loss": 3.4996612881109153, + "tokens_seen": 1738932224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023896690070210634, + "loss": 1.3302, + "theoretical_loss": 3.4996500785038114, + "tokens_seen": 1738997760 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023895687061183552, + "loss": 1.3664, + "theoretical_loss": 3.499638869437425, + "tokens_seen": 1739063296 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002389468405215647, + "loss": 1.446, + "theoretical_loss": 3.4996276609117105, + "tokens_seen": 1739128832 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023893681043129388, + "loss": 1.4105, + "theoretical_loss": 3.499616452926621, + "tokens_seen": 1739194368 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002389267803410231, + "loss": 1.3532, + "theoretical_loss": 3.4996052454821105, + "tokens_seen": 1739259904 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023891675025075227, + "loss": 1.3619, + "theoretical_loss": 3.499594038578132, + "tokens_seen": 1739325440 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023890672016048145, + "loss": 1.3855, + "theoretical_loss": 3.4995828322146387, + "tokens_seen": 1739390976 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023889669007021063, + "loss": 1.4624, + "theoretical_loss": 3.4995716263915853, + "tokens_seen": 1739456512 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023888665997993984, + "loss": 1.3659, + "theoretical_loss": 3.4995604211089244, + "tokens_seen": 1739522048 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023887662988966902, + "loss": 1.3773, + "theoretical_loss": 3.49954921636661, + "tokens_seen": 1739587584 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002388665997993982, + "loss": 1.2898, + "theoretical_loss": 3.499538012164596, + "tokens_seen": 1739653120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023885656970912739, + "loss": 1.445, + "theoretical_loss": 3.499526808502835, + "tokens_seen": 1739718656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023884653961885657, + "loss": 1.3619, + "theoretical_loss": 3.4995156053812817, + "tokens_seen": 1739784192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023883650952858577, + "loss": 1.3424, + "theoretical_loss": 3.4995044027998894, + "tokens_seen": 1739849728 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023882647943831496, + "loss": 1.29, + "theoretical_loss": 3.499493200758611, + "tokens_seen": 1739915264 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 635535, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.308260202407837, + "objective/train/theoretical_loss": 3.499481999257401, + "objective/train/tokens_used": 1760440800, + "theoretical_loss": 3.499481999257401, + "tokens_seen": 1739980800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023881644934804414, + "loss": 1.3022, + "theoretical_loss": 3.499481999257401, + "tokens_seen": 1739980800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023880641925777332, + "loss": 1.3538, + "theoretical_loss": 3.4994707982962128, + "tokens_seen": 1740046336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023879638916750253, + "loss": 1.3092, + "theoretical_loss": 3.499459597875, + "tokens_seen": 1740111872 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002387863590772317, + "loss": 1.3673, + "theoretical_loss": 3.4994483979937154, + "tokens_seen": 1740177408 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002387763289869609, + "loss": 1.3191, + "theoretical_loss": 3.4994371986523145, + "tokens_seen": 1740242944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023876629889669007, + "loss": 1.3639, + "theoretical_loss": 3.499425999850749, + "tokens_seen": 1740308480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023875626880641925, + "loss": 1.2584, + "theoretical_loss": 3.4994148015889737, + "tokens_seen": 1740374016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023874623871614846, + "loss": 1.3147, + "theoretical_loss": 3.499403603866942, + "tokens_seen": 1740439552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023873620862587764, + "loss": 1.4315, + "theoretical_loss": 3.499392406684607, + "tokens_seen": 1740505088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023872617853560682, + "loss": 1.3552, + "theoretical_loss": 3.4993812100419235, + "tokens_seen": 1740570624 + }, + { + "epoch": 0.53, + "learning_rate": 0.000238716148445336, + "loss": 1.3388, + "theoretical_loss": 3.4993700139388446, + "tokens_seen": 1740636160 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002387061183550652, + "loss": 1.346, + "theoretical_loss": 3.4993588183753235, + "tokens_seen": 1740701696 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002386960882647944, + "loss": 1.3883, + "theoretical_loss": 3.4993476233513148, + "tokens_seen": 1740767232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023868605817452357, + "loss": 1.2937, + "theoretical_loss": 3.499336428866771, + "tokens_seen": 1740832768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023867602808425275, + "loss": 1.3817, + "theoretical_loss": 3.4993252349216473, + "tokens_seen": 1740898304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023866599799398194, + "loss": 1.3171, + "theoretical_loss": 3.4993140415158956, + "tokens_seen": 1740963840 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023865596790371114, + "loss": 1.299, + "theoretical_loss": 3.4993028486494717, + "tokens_seen": 1741029376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023864593781344032, + "loss": 1.3066, + "theoretical_loss": 3.4992916563223275, + "tokens_seen": 1741094912 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002386359077231695, + "loss": 1.3654, + "theoretical_loss": 3.4992804645344178, + "tokens_seen": 1741160448 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002386258776328987, + "loss": 1.2226, + "theoretical_loss": 3.499269273285696, + "tokens_seen": 1741225984 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002386158475426279, + "loss": 1.395, + "theoretical_loss": 3.499258082576116, + "tokens_seen": 1741291520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023860581745235708, + "loss": 1.292, + "theoretical_loss": 3.4992468924056306, + "tokens_seen": 1741357056 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023859578736208626, + "loss": 1.384, + "theoretical_loss": 3.499235702774195, + "tokens_seen": 1741422592 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023858575727181544, + "loss": 1.2192, + "theoretical_loss": 3.499224513681762, + "tokens_seen": 1741488128 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023857572718154462, + "loss": 1.4249, + "theoretical_loss": 3.4992133251282853, + "tokens_seen": 1741553664 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023856569709127383, + "loss": 1.2976, + "theoretical_loss": 3.499202137113719, + "tokens_seen": 1741619200 + }, + { + "epoch": 0.53, + "learning_rate": 0.000238555667001003, + "loss": 1.4056, + "theoretical_loss": 3.499190949638017, + "tokens_seen": 1741684736 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002385456369107322, + "loss": 1.3763, + "theoretical_loss": 3.499179762701133, + "tokens_seen": 1741750272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023853560682046137, + "loss": 1.2919, + "theoretical_loss": 3.4991685763030205, + "tokens_seen": 1741815808 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023852557673019058, + "loss": 1.3127, + "theoretical_loss": 3.4991573904436333, + "tokens_seen": 1741881344 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023851554663991976, + "loss": 1.3898, + "theoretical_loss": 3.499146205122926, + "tokens_seen": 1741946880 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023850551654964894, + "loss": 1.3892, + "theoretical_loss": 3.4991350203408507, + "tokens_seen": 1742012416 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023849548645937812, + "loss": 1.5164, + "theoretical_loss": 3.499123836097363, + "tokens_seen": 1742077952 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023848545636910733, + "loss": 1.3429, + "theoretical_loss": 3.499112652392416, + "tokens_seen": 1742143488 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002384754262788365, + "loss": 1.3642, + "theoretical_loss": 3.499101469225963, + "tokens_seen": 1742209024 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002384653961885657, + "loss": 1.342, + "theoretical_loss": 3.4990902865979585, + "tokens_seen": 1742274560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023845536609829487, + "loss": 1.2665, + "theoretical_loss": 3.499079104508356, + "tokens_seen": 1742340096 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023844533600802406, + "loss": 1.3902, + "theoretical_loss": 3.499067922957109, + "tokens_seen": 1742405632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023843530591775326, + "loss": 1.3026, + "theoretical_loss": 3.4990567419441723, + "tokens_seen": 1742471168 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023842527582748244, + "loss": 1.327, + "theoretical_loss": 3.499045561469499, + "tokens_seen": 1742536704 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023841524573721163, + "loss": 1.346, + "theoretical_loss": 3.4990343815330434, + "tokens_seen": 1742602240 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002384052156469408, + "loss": 1.5267, + "theoretical_loss": 3.4990232021347585, + "tokens_seen": 1742667776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023839518555667002, + "loss": 1.3322, + "theoretical_loss": 3.4990120232745987, + "tokens_seen": 1742733312 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002383851554663992, + "loss": 1.456, + "theoretical_loss": 3.4990008449525183, + "tokens_seen": 1742798848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023837512537612838, + "loss": 1.3783, + "theoretical_loss": 3.4989896671684706, + "tokens_seen": 1742864384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023836509528585756, + "loss": 1.4209, + "theoretical_loss": 3.49897848992241, + "tokens_seen": 1742929920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023835506519558677, + "loss": 1.3259, + "theoretical_loss": 3.4989673132142896, + "tokens_seen": 1742995456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023834503510531598, + "loss": 1.404, + "theoretical_loss": 3.498956137044064, + "tokens_seen": 1743060992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023833500501504516, + "loss": 1.336, + "theoretical_loss": 3.498944961411687, + "tokens_seen": 1743126528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023832497492477434, + "loss": 1.4764, + "theoretical_loss": 3.4989337863171115, + "tokens_seen": 1743192064 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 636817, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3697288036346436, + "objective/train/theoretical_loss": 3.4989226117602925, + "objective/train/tokens_used": 1763717600, + "theoretical_loss": 3.4989226117602925, + "tokens_seen": 1743257600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023831494483450352, + "loss": 1.356, + "theoretical_loss": 3.4989226117602925, + "tokens_seen": 1743257600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023830491474423273, + "loss": 1.3845, + "theoretical_loss": 3.498911437741184, + "tokens_seen": 1743323136 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002382948846539619, + "loss": 1.445, + "theoretical_loss": 3.4989002642597393, + "tokens_seen": 1743388672 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002382848545636911, + "loss": 1.3544, + "theoretical_loss": 3.4988890913159123, + "tokens_seen": 1743454208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023827482447342027, + "loss": 1.3474, + "theoretical_loss": 3.498877918909657, + "tokens_seen": 1743519744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023826479438314945, + "loss": 1.3343, + "theoretical_loss": 3.4988667470409283, + "tokens_seen": 1743585280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023825476429287866, + "loss": 1.3857, + "theoretical_loss": 3.498855575709679, + "tokens_seen": 1743650816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023824473420260784, + "loss": 1.3825, + "theoretical_loss": 3.4988444049158627, + "tokens_seen": 1743716352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023823470411233702, + "loss": 1.3757, + "theoretical_loss": 3.4988332346594344, + "tokens_seen": 1743781888 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002382246740220662, + "loss": 1.2356, + "theoretical_loss": 3.498822064940348, + "tokens_seen": 1743847424 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002382146439317954, + "loss": 1.4842, + "theoretical_loss": 3.498810895758557, + "tokens_seen": 1743912960 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002382046138415246, + "loss": 1.4094, + "theoretical_loss": 3.498799727114015, + "tokens_seen": 1743978496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023819458375125377, + "loss": 1.419, + "theoretical_loss": 3.4987885590066767, + "tokens_seen": 1744044032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023818455366098295, + "loss": 1.3618, + "theoretical_loss": 3.498777391436496, + "tokens_seen": 1744109568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023817452357071214, + "loss": 1.2537, + "theoretical_loss": 3.4987662244034268, + "tokens_seen": 1744175104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023816449348044134, + "loss": 1.2369, + "theoretical_loss": 3.498755057907423, + "tokens_seen": 1744240640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023815446339017052, + "loss": 1.2379, + "theoretical_loss": 3.4987438919484384, + "tokens_seen": 1744306176 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002381444332998997, + "loss": 1.3543, + "theoretical_loss": 3.498732726526427, + "tokens_seen": 1744371712 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002381344032096289, + "loss": 1.3494, + "theoretical_loss": 3.4987215616413434, + "tokens_seen": 1744437248 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002381243731193581, + "loss": 1.3047, + "theoretical_loss": 3.498710397293141, + "tokens_seen": 1744502784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023811434302908728, + "loss": 1.3602, + "theoretical_loss": 3.498699233481774, + "tokens_seen": 1744568320 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023810431293881646, + "loss": 1.3211, + "theoretical_loss": 3.498688070207197, + "tokens_seen": 1744633856 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023809428284854564, + "loss": 1.4012, + "theoretical_loss": 3.4986769074693624, + "tokens_seen": 1744699392 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023808425275827482, + "loss": 1.4427, + "theoretical_loss": 3.498665745268226, + "tokens_seen": 1744764928 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023807422266800403, + "loss": 1.2978, + "theoretical_loss": 3.498654583603741, + "tokens_seen": 1744830464 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002380641925777332, + "loss": 1.3622, + "theoretical_loss": 3.498643422475862, + "tokens_seen": 1744896000 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002380541624874624, + "loss": 1.5354, + "theoretical_loss": 3.4986322618845422, + "tokens_seen": 1744961536 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023804413239719157, + "loss": 1.2773, + "theoretical_loss": 3.4986211018297357, + "tokens_seen": 1745027072 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023803410230692078, + "loss": 1.3312, + "theoretical_loss": 3.498609942311398, + "tokens_seen": 1745092608 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023802407221664996, + "loss": 1.2887, + "theoretical_loss": 3.4985987833294816, + "tokens_seen": 1745158144 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023801404212637914, + "loss": 1.4626, + "theoretical_loss": 3.4985876248839407, + "tokens_seen": 1745223680 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023800401203610832, + "loss": 1.3571, + "theoretical_loss": 3.49857646697473, + "tokens_seen": 1745289216 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023799398194583753, + "loss": 1.3833, + "theoretical_loss": 3.498565309601804, + "tokens_seen": 1745354752 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002379839518555667, + "loss": 1.3862, + "theoretical_loss": 3.498554152765115, + "tokens_seen": 1745420288 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002379739217652959, + "loss": 1.2506, + "theoretical_loss": 3.498542996464619, + "tokens_seen": 1745485824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023796389167502507, + "loss": 1.4597, + "theoretical_loss": 3.4985318407002692, + "tokens_seen": 1745551360 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023795386158475426, + "loss": 1.2989, + "theoretical_loss": 3.4985206854720197, + "tokens_seen": 1745616896 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023794383149448346, + "loss": 1.4246, + "theoretical_loss": 3.498509530779825, + "tokens_seen": 1745682432 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023793380140421265, + "loss": 1.393, + "theoretical_loss": 3.4984983766236386, + "tokens_seen": 1745747968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023792377131394183, + "loss": 1.4045, + "theoretical_loss": 3.498487223003415, + "tokens_seen": 1745813504 + }, + { + "epoch": 0.53, + "learning_rate": 0.000237913741223671, + "loss": 1.3887, + "theoretical_loss": 3.498476069919109, + "tokens_seen": 1745879040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023790371113340022, + "loss": 1.4002, + "theoretical_loss": 3.498464917370673, + "tokens_seen": 1745944576 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002378936810431294, + "loss": 1.3345, + "theoretical_loss": 3.498453765358063, + "tokens_seen": 1746010112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023788365095285858, + "loss": 1.4762, + "theoretical_loss": 3.498442613881232, + "tokens_seen": 1746075648 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023787362086258776, + "loss": 1.3757, + "theoretical_loss": 3.4984314629401347, + "tokens_seen": 1746141184 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023786359077231694, + "loss": 1.4051, + "theoretical_loss": 3.498420312534725, + "tokens_seen": 1746206720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023785356068204615, + "loss": 1.3381, + "theoretical_loss": 3.498409162664957, + "tokens_seen": 1746272256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023784353059177533, + "loss": 1.4399, + "theoretical_loss": 3.4983980133307853, + "tokens_seen": 1746337792 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002378335005015045, + "loss": 1.2877, + "theoretical_loss": 3.498386864532163, + "tokens_seen": 1746403328 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002378234704112337, + "loss": 1.2096, + "theoretical_loss": 3.498375716269046, + "tokens_seen": 1746468864 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 637932, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.407595157623291, + "objective/train/theoretical_loss": 3.498364568541387, + "objective/train/tokens_used": 1766994400, + "theoretical_loss": 3.498364568541387, + "tokens_seen": 1746534400 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002378134403209629, + "loss": 1.3138, + "theoretical_loss": 3.498364568541387, + "tokens_seen": 1746534400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023780341023069208, + "loss": 1.4158, + "theoretical_loss": 3.4983534213491407, + "tokens_seen": 1746599936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023779338014042126, + "loss": 1.3071, + "theoretical_loss": 3.4983422746922614, + "tokens_seen": 1746665472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023778335005015044, + "loss": 1.3895, + "theoretical_loss": 3.4983311285707033, + "tokens_seen": 1746731008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023777331995987962, + "loss": 1.3223, + "theoretical_loss": 3.4983199829844205, + "tokens_seen": 1746796544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023776328986960883, + "loss": 1.3572, + "theoretical_loss": 3.498308837933367, + "tokens_seen": 1746862080 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023775325977933801, + "loss": 1.3756, + "theoretical_loss": 3.4982976934174976, + "tokens_seen": 1746927616 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002377432296890672, + "loss": 1.2771, + "theoretical_loss": 3.4982865494367656, + "tokens_seen": 1746993152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023773319959879638, + "loss": 1.2896, + "theoretical_loss": 3.4982754059911265, + "tokens_seen": 1747058688 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023772316950852558, + "loss": 1.2322, + "theoretical_loss": 3.498264263080533, + "tokens_seen": 1747124224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023771313941825477, + "loss": 1.3115, + "theoretical_loss": 3.498253120704941, + "tokens_seen": 1747189760 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023770310932798395, + "loss": 1.3146, + "theoretical_loss": 3.4982419788643035, + "tokens_seen": 1747255296 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023769307923771313, + "loss": 1.4078, + "theoretical_loss": 3.498230837558576, + "tokens_seen": 1747320832 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002376830491474423, + "loss": 1.3289, + "theoretical_loss": 3.498219696787711, + "tokens_seen": 1747386368 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023767301905717152, + "loss": 1.4827, + "theoretical_loss": 3.498208556551664, + "tokens_seen": 1747451904 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002376629889669007, + "loss": 1.3661, + "theoretical_loss": 3.4981974168503887, + "tokens_seen": 1747517440 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023765295887662988, + "loss": 1.4752, + "theoretical_loss": 3.49818627768384, + "tokens_seen": 1747582976 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023764292878635906, + "loss": 1.2414, + "theoretical_loss": 3.4981751390519715, + "tokens_seen": 1747648512 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023763289869608827, + "loss": 1.3171, + "theoretical_loss": 3.498164000954738, + "tokens_seen": 1747714048 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023762286860581745, + "loss": 1.3889, + "theoretical_loss": 3.498152863392094, + "tokens_seen": 1747779584 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023761283851554663, + "loss": 1.3107, + "theoretical_loss": 3.498141726363993, + "tokens_seen": 1747845120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023760280842527584, + "loss": 1.3559, + "theoretical_loss": 3.4981305898703896, + "tokens_seen": 1747910656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023759277833500502, + "loss": 1.4349, + "theoretical_loss": 3.4981194539112384, + "tokens_seen": 1747976192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023758274824473423, + "loss": 1.2624, + "theoretical_loss": 3.4981083184864934, + "tokens_seen": 1748041728 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002375727181544634, + "loss": 1.4372, + "theoretical_loss": 3.4980971835961094, + "tokens_seen": 1748107264 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002375626880641926, + "loss": 1.3788, + "theoretical_loss": 3.49808604924004, + "tokens_seen": 1748172800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023755265797392177, + "loss": 1.2466, + "theoretical_loss": 3.49807491541824, + "tokens_seen": 1748238336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023754262788365098, + "loss": 1.3507, + "theoretical_loss": 3.4980637821306635, + "tokens_seen": 1748303872 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023753259779338016, + "loss": 1.3069, + "theoretical_loss": 3.4980526493772652, + "tokens_seen": 1748369408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023752256770310934, + "loss": 1.1823, + "theoretical_loss": 3.498041517157999, + "tokens_seen": 1748434944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023751253761283852, + "loss": 1.3228, + "theoretical_loss": 3.49803038547282, + "tokens_seen": 1748500480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023750250752256773, + "loss": 1.2589, + "theoretical_loss": 3.4980192543216813, + "tokens_seen": 1748566016 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002374924774322969, + "loss": 1.3328, + "theoretical_loss": 3.4980081237045386, + "tokens_seen": 1748631552 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002374824473420261, + "loss": 1.3711, + "theoretical_loss": 3.4979969936213458, + "tokens_seen": 1748697088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023747241725175527, + "loss": 1.4328, + "theoretical_loss": 3.497985864072057, + "tokens_seen": 1748762624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023746238716148446, + "loss": 1.326, + "theoretical_loss": 3.4979747350566264, + "tokens_seen": 1748828160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023745235707121366, + "loss": 1.4305, + "theoretical_loss": 3.4979636065750093, + "tokens_seen": 1748893696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023744232698094285, + "loss": 1.4285, + "theoretical_loss": 3.497952478627159, + "tokens_seen": 1748959232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023743229689067203, + "loss": 1.4381, + "theoretical_loss": 3.4979413512130306, + "tokens_seen": 1749024768 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002374222668004012, + "loss": 1.3595, + "theoretical_loss": 3.4979302243325785, + "tokens_seen": 1749090304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023741223671013042, + "loss": 1.3514, + "theoretical_loss": 3.497919097985757, + "tokens_seen": 1749155840 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002374022066198596, + "loss": 1.2824, + "theoretical_loss": 3.49790797217252, + "tokens_seen": 1749221376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023739217652958878, + "loss": 1.3215, + "theoretical_loss": 3.4978968468928224, + "tokens_seen": 1749286912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023738214643931796, + "loss": 1.2903, + "theoretical_loss": 3.497885722146619, + "tokens_seen": 1749352448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023737211634904714, + "loss": 1.3089, + "theoretical_loss": 3.4978745979338637, + "tokens_seen": 1749417984 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023736208625877635, + "loss": 1.3824, + "theoretical_loss": 3.497863474254511, + "tokens_seen": 1749483520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023735205616850553, + "loss": 1.4039, + "theoretical_loss": 3.4978523511085156, + "tokens_seen": 1749549056 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002373420260782347, + "loss": 1.3451, + "theoretical_loss": 3.4978412284958313, + "tokens_seen": 1749614592 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002373319959879639, + "loss": 1.4331, + "theoretical_loss": 3.497830106416414, + "tokens_seen": 1749680128 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002373219658976931, + "loss": 1.2744, + "theoretical_loss": 3.497818984870216, + "tokens_seen": 1749745664 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 639149, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3413631916046143, + "objective/train/theoretical_loss": 3.4978078638571937, + "objective/train/tokens_used": 1770271200, + "theoretical_loss": 3.4978078638571937, + "tokens_seen": 1749811200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023731193580742228, + "loss": 1.3877, + "theoretical_loss": 3.4978078638571937, + "tokens_seen": 1749811200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023730190571715146, + "loss": 1.3242, + "theoretical_loss": 3.4977967433773003, + "tokens_seen": 1749876736 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023729187562688064, + "loss": 1.3827, + "theoretical_loss": 3.497785623430491, + "tokens_seen": 1749942272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023728184553660982, + "loss": 1.436, + "theoretical_loss": 3.4977745040167205, + "tokens_seen": 1750007808 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023727181544633903, + "loss": 1.5076, + "theoretical_loss": 3.4977633851359426, + "tokens_seen": 1750073344 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023726178535606821, + "loss": 1.4245, + "theoretical_loss": 3.497752266788112, + "tokens_seen": 1750138880 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002372517552657974, + "loss": 1.3794, + "theoretical_loss": 3.497741148973183, + "tokens_seen": 1750204416 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023724172517552658, + "loss": 1.45, + "theoretical_loss": 3.497730031691111, + "tokens_seen": 1750269952 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023723169508525578, + "loss": 1.161, + "theoretical_loss": 3.4977189149418493, + "tokens_seen": 1750335488 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023722166499498497, + "loss": 1.2701, + "theoretical_loss": 3.497707798725353, + "tokens_seen": 1750401024 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023721163490471415, + "loss": 1.3309, + "theoretical_loss": 3.497696683041577, + "tokens_seen": 1750466560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023720160481444333, + "loss": 1.2931, + "theoretical_loss": 3.4976855678904752, + "tokens_seen": 1750532096 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002371915747241725, + "loss": 1.384, + "theoretical_loss": 3.497674453272002, + "tokens_seen": 1750597632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023718154463390172, + "loss": 1.361, + "theoretical_loss": 3.497663339186113, + "tokens_seen": 1750663168 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002371715145436309, + "loss": 1.3088, + "theoretical_loss": 3.497652225632762, + "tokens_seen": 1750728704 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023716148445336008, + "loss": 1.4425, + "theoretical_loss": 3.4976411126119036, + "tokens_seen": 1750794240 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023715145436308926, + "loss": 1.2664, + "theoretical_loss": 3.497630000123492, + "tokens_seen": 1750859776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023714142427281847, + "loss": 1.2518, + "theoretical_loss": 3.497618888167482, + "tokens_seen": 1750925312 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023713139418254765, + "loss": 1.3649, + "theoretical_loss": 3.4976077767438287, + "tokens_seen": 1750990848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023712136409227683, + "loss": 1.3006, + "theoretical_loss": 3.497596665852486, + "tokens_seen": 1751056384 + }, + { + "epoch": 0.53, + "learning_rate": 0.000237111334002006, + "loss": 1.384, + "theoretical_loss": 3.497585555493409, + "tokens_seen": 1751121920 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002371013039117352, + "loss": 1.3886, + "theoretical_loss": 3.497574445666552, + "tokens_seen": 1751187456 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002370912738214644, + "loss": 1.3087, + "theoretical_loss": 3.4975633363718694, + "tokens_seen": 1751252992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023708124373119358, + "loss": 1.3752, + "theoretical_loss": 3.497552227609316, + "tokens_seen": 1751318528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023707121364092276, + "loss": 1.4422, + "theoretical_loss": 3.4975411193788464, + "tokens_seen": 1751384064 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023706118355065194, + "loss": 1.3657, + "theoretical_loss": 3.4975300116804156, + "tokens_seen": 1751449600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023705115346038115, + "loss": 1.3451, + "theoretical_loss": 3.497518904513978, + "tokens_seen": 1751515136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023704112337011033, + "loss": 1.3724, + "theoretical_loss": 3.4975077978794875, + "tokens_seen": 1751580672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023703109327983952, + "loss": 1.343, + "theoretical_loss": 3.497496691776899, + "tokens_seen": 1751646208 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002370210631895687, + "loss": 1.4224, + "theoretical_loss": 3.497485586206168, + "tokens_seen": 1751711744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023701103309929788, + "loss": 1.3457, + "theoretical_loss": 3.497474481167248, + "tokens_seen": 1751777280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023700100300902709, + "loss": 1.4887, + "theoretical_loss": 3.4974633766600944, + "tokens_seen": 1751842816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023699097291875627, + "loss": 1.3213, + "theoretical_loss": 3.4974522726846615, + "tokens_seen": 1751908352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023698094282848545, + "loss": 1.3796, + "theoretical_loss": 3.497441169240904, + "tokens_seen": 1751973888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023697091273821463, + "loss": 1.4549, + "theoretical_loss": 3.497430066328777, + "tokens_seen": 1752039424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023696088264794384, + "loss": 1.3981, + "theoretical_loss": 3.4974189639482347, + "tokens_seen": 1752104960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023695085255767302, + "loss": 1.4243, + "theoretical_loss": 3.497407862099232, + "tokens_seen": 1752170496 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002369408224674022, + "loss": 1.4082, + "theoretical_loss": 3.497396760781723, + "tokens_seen": 1752236032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023693079237713138, + "loss": 1.5018, + "theoretical_loss": 3.497385659995663, + "tokens_seen": 1752301568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023692076228686056, + "loss": 1.316, + "theoretical_loss": 3.497374559741006, + "tokens_seen": 1752367104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023691073219658977, + "loss": 1.3521, + "theoretical_loss": 3.4973634600177075, + "tokens_seen": 1752432640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023690070210631895, + "loss": 1.3604, + "theoretical_loss": 3.497352360825722, + "tokens_seen": 1752498176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023689067201604813, + "loss": 1.3469, + "theoretical_loss": 3.497341262165004, + "tokens_seen": 1752563712 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002368806419257773, + "loss": 1.4926, + "theoretical_loss": 3.4973301640355086, + "tokens_seen": 1752629248 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023687061183550652, + "loss": 1.3342, + "theoretical_loss": 3.497319066437189, + "tokens_seen": 1752694784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023686058174523573, + "loss": 1.2787, + "theoretical_loss": 3.497307969370002, + "tokens_seen": 1752760320 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002368505516549649, + "loss": 1.2363, + "theoretical_loss": 3.4972968728339016, + "tokens_seen": 1752825856 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002368405215646941, + "loss": 1.3318, + "theoretical_loss": 3.497285776828842, + "tokens_seen": 1752891392 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002368304914744233, + "loss": 1.3136, + "theoretical_loss": 3.4972746813547784, + "tokens_seen": 1752956928 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023682046138415248, + "loss": 1.3162, + "theoretical_loss": 3.4972635864116652, + "tokens_seen": 1753022464 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 640338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3299673795700073, + "objective/train/theoretical_loss": 3.4972524919994576, + "objective/train/tokens_used": 1773548000, + "theoretical_loss": 3.4972524919994576, + "tokens_seen": 1753088000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023681043129388166, + "loss": 1.3325, + "theoretical_loss": 3.4972524919994576, + "tokens_seen": 1753088000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023680040120361084, + "loss": 1.3329, + "theoretical_loss": 3.49724139811811, + "tokens_seen": 1753153536 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023679037111334002, + "loss": 1.3697, + "theoretical_loss": 3.4972303047675775, + "tokens_seen": 1753219072 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023678034102306923, + "loss": 1.3856, + "theoretical_loss": 3.4972192119478143, + "tokens_seen": 1753284608 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023677031093279841, + "loss": 1.3954, + "theoretical_loss": 3.497208119658775, + "tokens_seen": 1753350144 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002367602808425276, + "loss": 1.3334, + "theoretical_loss": 3.4971970279004156, + "tokens_seen": 1753415680 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023675025075225678, + "loss": 1.4529, + "theoretical_loss": 3.4971859366726905, + "tokens_seen": 1753481216 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023674022066198598, + "loss": 1.3889, + "theoretical_loss": 3.497174845975553, + "tokens_seen": 1753546752 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023673019057171517, + "loss": 1.3537, + "theoretical_loss": 3.49716375580896, + "tokens_seen": 1753612288 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023672016048144435, + "loss": 1.4217, + "theoretical_loss": 3.4971526661728647, + "tokens_seen": 1753677824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023671013039117353, + "loss": 1.3863, + "theoretical_loss": 3.4971415770672225, + "tokens_seen": 1753743360 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002367001003009027, + "loss": 1.4184, + "theoretical_loss": 3.4971304884919885, + "tokens_seen": 1753808896 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023669007021063192, + "loss": 1.3173, + "theoretical_loss": 3.4971194004471173, + "tokens_seen": 1753874432 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002366800401203611, + "loss": 1.3066, + "theoretical_loss": 3.4971083129325633, + "tokens_seen": 1753939968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023667001003009028, + "loss": 1.3538, + "theoretical_loss": 3.4970972259482824, + "tokens_seen": 1754005504 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023665997993981946, + "loss": 1.3866, + "theoretical_loss": 3.497086139494228, + "tokens_seen": 1754071040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023664994984954867, + "loss": 1.4839, + "theoretical_loss": 3.497075053570356, + "tokens_seen": 1754136576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023663991975927785, + "loss": 1.5205, + "theoretical_loss": 3.4970639681766205, + "tokens_seen": 1754202112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023662988966900703, + "loss": 1.3124, + "theoretical_loss": 3.4970528833129766, + "tokens_seen": 1754267648 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002366198595787362, + "loss": 1.4277, + "theoretical_loss": 3.49704179897938, + "tokens_seen": 1754333184 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002366098294884654, + "loss": 1.4298, + "theoretical_loss": 3.497030715175784, + "tokens_seen": 1754398720 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002365997993981946, + "loss": 1.4054, + "theoretical_loss": 3.4970196319021447, + "tokens_seen": 1754464256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023658976930792378, + "loss": 1.3761, + "theoretical_loss": 3.4970085491584166, + "tokens_seen": 1754529792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023657973921765296, + "loss": 1.4713, + "theoretical_loss": 3.496997466944554, + "tokens_seen": 1754595328 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023656970912738214, + "loss": 1.4245, + "theoretical_loss": 3.496986385260513, + "tokens_seen": 1754660864 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023655967903711135, + "loss": 1.4131, + "theoretical_loss": 3.496975304106247, + "tokens_seen": 1754726400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023654964894684053, + "loss": 1.4002, + "theoretical_loss": 3.4969642234817124, + "tokens_seen": 1754791936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023653961885656972, + "loss": 1.3863, + "theoretical_loss": 3.496953143386863, + "tokens_seen": 1754857472 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002365295887662989, + "loss": 1.3339, + "theoretical_loss": 3.4969420638216535, + "tokens_seen": 1754923008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023651955867602808, + "loss": 1.3557, + "theoretical_loss": 3.49693098478604, + "tokens_seen": 1754988544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023650952858575729, + "loss": 1.3795, + "theoretical_loss": 3.496919906279977, + "tokens_seen": 1755054080 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023649949849548647, + "loss": 1.3195, + "theoretical_loss": 3.496908828303419, + "tokens_seen": 1755119616 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023648946840521565, + "loss": 1.4474, + "theoretical_loss": 3.496897750856321, + "tokens_seen": 1755185152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023647943831494483, + "loss": 1.3788, + "theoretical_loss": 3.496886673938638, + "tokens_seen": 1755250688 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023646940822467404, + "loss": 1.3615, + "theoretical_loss": 3.496875597550325, + "tokens_seen": 1755316224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023645937813440322, + "loss": 1.3699, + "theoretical_loss": 3.4968645216913368, + "tokens_seen": 1755381760 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002364493480441324, + "loss": 1.3889, + "theoretical_loss": 3.4968534463616283, + "tokens_seen": 1755447296 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023643931795386158, + "loss": 1.3619, + "theoretical_loss": 3.496842371561155, + "tokens_seen": 1755512832 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023642928786359076, + "loss": 1.32, + "theoretical_loss": 3.496831297289871, + "tokens_seen": 1755578368 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023641925777331997, + "loss": 1.3923, + "theoretical_loss": 3.4968202235477324, + "tokens_seen": 1755643904 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023640922768304915, + "loss": 1.3562, + "theoretical_loss": 3.496809150334693, + "tokens_seen": 1755709440 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023639919759277833, + "loss": 1.3622, + "theoretical_loss": 3.4967980776507086, + "tokens_seen": 1755774976 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002363891675025075, + "loss": 1.2747, + "theoretical_loss": 3.4967870054957335, + "tokens_seen": 1755840512 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023637913741223672, + "loss": 1.3663, + "theoretical_loss": 3.4967759338697233, + "tokens_seen": 1755906048 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002363691073219659, + "loss": 1.3272, + "theoretical_loss": 3.496764862772632, + "tokens_seen": 1755971584 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023635907723169508, + "loss": 1.3925, + "theoretical_loss": 3.496753792204416, + "tokens_seen": 1756037120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023634904714142426, + "loss": 1.4124, + "theoretical_loss": 3.4967427221650293, + "tokens_seen": 1756102656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023633901705115347, + "loss": 1.2724, + "theoretical_loss": 3.4967316526544274, + "tokens_seen": 1756168192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023632898696088265, + "loss": 1.357, + "theoretical_loss": 3.496720583672565, + "tokens_seen": 1756233728 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023631895687061184, + "loss": 1.3965, + "theoretical_loss": 3.496709515219397, + "tokens_seen": 1756299264 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 641471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2963881492614746, + "objective/train/theoretical_loss": 3.4966984472948788, + "objective/train/tokens_used": 1776824800, + "theoretical_loss": 3.4966984472948788, + "tokens_seen": 1756364800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023630892678034102, + "loss": 1.2661, + "theoretical_loss": 3.4966984472948788, + "tokens_seen": 1756364800 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002362988966900702, + "loss": 1.3711, + "theoretical_loss": 3.4966873798989653, + "tokens_seen": 1756430336 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002362888665997994, + "loss": 1.5016, + "theoretical_loss": 3.4966763130316116, + "tokens_seen": 1756495872 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002362788365095286, + "loss": 1.2677, + "theoretical_loss": 3.4966652466927726, + "tokens_seen": 1756561408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023626880641925777, + "loss": 1.395, + "theoretical_loss": 3.496654180882403, + "tokens_seen": 1756626944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023625877632898695, + "loss": 1.3571, + "theoretical_loss": 3.4966431156004587, + "tokens_seen": 1756692480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023624874623871616, + "loss": 1.3004, + "theoretical_loss": 3.496632050846894, + "tokens_seen": 1756758016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023623871614844534, + "loss": 1.3751, + "theoretical_loss": 3.4966209866216644, + "tokens_seen": 1756823552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023622868605817452, + "loss": 1.393, + "theoretical_loss": 3.496609922924725, + "tokens_seen": 1756889088 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002362186559679037, + "loss": 1.3944, + "theoretical_loss": 3.4965988597560305, + "tokens_seen": 1756954624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023620862587763288, + "loss": 1.4881, + "theoretical_loss": 3.496587797115536, + "tokens_seen": 1757020160 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002361985957873621, + "loss": 1.3562, + "theoretical_loss": 3.496576735003197, + "tokens_seen": 1757085696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023618856569709127, + "loss": 1.3089, + "theoretical_loss": 3.496565673418968, + "tokens_seen": 1757151232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023617853560682045, + "loss": 1.3719, + "theoretical_loss": 3.4965546123628046, + "tokens_seen": 1757216768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023616850551654963, + "loss": 1.3913, + "theoretical_loss": 3.4965435518346615, + "tokens_seen": 1757282304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023615847542627884, + "loss": 1.4821, + "theoretical_loss": 3.496532491834494, + "tokens_seen": 1757347840 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023614844533600802, + "loss": 1.3033, + "theoretical_loss": 3.4965214323622575, + "tokens_seen": 1757413376 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002361384152457372, + "loss": 1.4494, + "theoretical_loss": 3.496510373417907, + "tokens_seen": 1757478912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023612838515546639, + "loss": 1.3404, + "theoretical_loss": 3.4964993150013974, + "tokens_seen": 1757544448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023611835506519557, + "loss": 1.4011, + "theoretical_loss": 3.4964882571126834, + "tokens_seen": 1757609984 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002361083249749248, + "loss": 1.376, + "theoretical_loss": 3.496477199751721, + "tokens_seen": 1757675520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023609829488465398, + "loss": 1.4417, + "theoretical_loss": 3.496466142918465, + "tokens_seen": 1757741056 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023608826479438316, + "loss": 1.3671, + "theoretical_loss": 3.4964550866128703, + "tokens_seen": 1757806592 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023607823470411234, + "loss": 1.3779, + "theoretical_loss": 3.4964440308348923, + "tokens_seen": 1757872128 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023606820461384155, + "loss": 1.5541, + "theoretical_loss": 3.4964329755844856, + "tokens_seen": 1757937664 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023605817452357073, + "loss": 1.462, + "theoretical_loss": 3.4964219208616067, + "tokens_seen": 1758003200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023604814443329992, + "loss": 1.4208, + "theoretical_loss": 3.49641086666621, + "tokens_seen": 1758068736 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002360381143430291, + "loss": 1.3604, + "theoretical_loss": 3.4963998129982494, + "tokens_seen": 1758134272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023602808425275828, + "loss": 1.4268, + "theoretical_loss": 3.496388759857682, + "tokens_seen": 1758199808 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023601805416248749, + "loss": 1.4472, + "theoretical_loss": 3.4963777072444624, + "tokens_seen": 1758265344 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023600802407221667, + "loss": 1.3814, + "theoretical_loss": 3.4963666551585453, + "tokens_seen": 1758330880 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023599799398194585, + "loss": 1.3684, + "theoretical_loss": 3.496355603599887, + "tokens_seen": 1758396416 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023598796389167503, + "loss": 1.3534, + "theoretical_loss": 3.496344552568441, + "tokens_seen": 1758461952 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023597793380140424, + "loss": 1.4157, + "theoretical_loss": 3.4963335020641635, + "tokens_seen": 1758527488 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023596790371113342, + "loss": 1.3004, + "theoretical_loss": 3.4963224520870098, + "tokens_seen": 1758593024 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002359578736208626, + "loss": 1.3545, + "theoretical_loss": 3.496311402636935, + "tokens_seen": 1758658560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023594784353059178, + "loss": 1.3025, + "theoretical_loss": 3.4963003537138944, + "tokens_seen": 1758724096 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023593781344032096, + "loss": 1.3917, + "theoretical_loss": 3.4962893053178425, + "tokens_seen": 1758789632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023592778335005017, + "loss": 1.2997, + "theoretical_loss": 3.4962782574487354, + "tokens_seen": 1758855168 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023591775325977935, + "loss": 1.2454, + "theoretical_loss": 3.496267210106528, + "tokens_seen": 1758920704 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023590772316950853, + "loss": 1.4102, + "theoretical_loss": 3.496256163291176, + "tokens_seen": 1758986240 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002358976930792377, + "loss": 1.3914, + "theoretical_loss": 3.496245117002634, + "tokens_seen": 1759051776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023588766298896692, + "loss": 1.3763, + "theoretical_loss": 3.496234071240857, + "tokens_seen": 1759117312 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002358776328986961, + "loss": 1.4644, + "theoretical_loss": 3.4962230260058016, + "tokens_seen": 1759182848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023586760280842528, + "loss": 1.3055, + "theoretical_loss": 3.4962119812974217, + "tokens_seen": 1759248384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023585757271815447, + "loss": 1.4271, + "theoretical_loss": 3.496200937115673, + "tokens_seen": 1759313920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023584754262788367, + "loss": 1.3907, + "theoretical_loss": 3.496189893460511, + "tokens_seen": 1759379456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023583751253761285, + "loss": 1.4902, + "theoretical_loss": 3.4961788503318907, + "tokens_seen": 1759444992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023582748244734204, + "loss": 1.4137, + "theoretical_loss": 3.4961678077297673, + "tokens_seen": 1759510528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023581745235707122, + "loss": 1.4789, + "theoretical_loss": 3.4961567656540966, + "tokens_seen": 1759576064 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 642679, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3076953887939453, + "objective/train/theoretical_loss": 3.4961457241048337, + "objective/train/tokens_used": 1780101600, + "theoretical_loss": 3.4961457241048337, + "tokens_seen": 1759641600 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002358074222668004, + "loss": 1.3797, + "theoretical_loss": 3.4961457241048337, + "tokens_seen": 1759641600 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002357973921765296, + "loss": 1.3443, + "theoretical_loss": 3.4961346830819338, + "tokens_seen": 1759707136 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002357873620862588, + "loss": 1.3021, + "theoretical_loss": 3.496123642585352, + "tokens_seen": 1759772672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023577733199598797, + "loss": 1.3697, + "theoretical_loss": 3.4961126026150433, + "tokens_seen": 1759838208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023576730190571715, + "loss": 1.3889, + "theoretical_loss": 3.4961015631709644, + "tokens_seen": 1759903744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023575727181544636, + "loss": 1.2841, + "theoretical_loss": 3.496090524253069, + "tokens_seen": 1759969280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023574724172517554, + "loss": 1.4297, + "theoretical_loss": 3.496079485861314, + "tokens_seen": 1760034816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023573721163490472, + "loss": 1.4332, + "theoretical_loss": 3.496068447995653, + "tokens_seen": 1760100352 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002357271815446339, + "loss": 1.4392, + "theoretical_loss": 3.496057410656043, + "tokens_seen": 1760165888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023571715145436308, + "loss": 1.4341, + "theoretical_loss": 3.4960463738424385, + "tokens_seen": 1760231424 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002357071213640923, + "loss": 1.3678, + "theoretical_loss": 3.496035337554795, + "tokens_seen": 1760296960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023569709127382147, + "loss": 1.3675, + "theoretical_loss": 3.4960243017930672, + "tokens_seen": 1760362496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023568706118355065, + "loss": 1.3948, + "theoretical_loss": 3.4960132665572115, + "tokens_seen": 1760428032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023567703109327983, + "loss": 1.411, + "theoretical_loss": 3.496002231847182, + "tokens_seen": 1760493568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023566700100300904, + "loss": 1.3589, + "theoretical_loss": 3.495991197662936, + "tokens_seen": 1760559104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023565697091273822, + "loss": 1.3934, + "theoretical_loss": 3.4959801640044272, + "tokens_seen": 1760624640 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002356469408224674, + "loss": 1.2864, + "theoretical_loss": 3.495969130871612, + "tokens_seen": 1760690176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023563691073219659, + "loss": 1.4096, + "theoretical_loss": 3.4959580982644445, + "tokens_seen": 1760755712 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023562688064192577, + "loss": 1.4155, + "theoretical_loss": 3.4959470661828815, + "tokens_seen": 1760821248 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023561685055165497, + "loss": 1.3523, + "theoretical_loss": 3.495936034626878, + "tokens_seen": 1760886784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023560682046138416, + "loss": 1.3742, + "theoretical_loss": 3.4959250035963887, + "tokens_seen": 1760952320 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023559679037111334, + "loss": 1.3943, + "theoretical_loss": 3.49591397309137, + "tokens_seen": 1761017856 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023558676028084252, + "loss": 1.2775, + "theoretical_loss": 3.495902943111777, + "tokens_seen": 1761083392 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023557673019057173, + "loss": 1.2955, + "theoretical_loss": 3.4958919136575646, + "tokens_seen": 1761148928 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002355667001003009, + "loss": 1.3642, + "theoretical_loss": 3.4958808847286886, + "tokens_seen": 1761214464 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002355566700100301, + "loss": 1.3495, + "theoretical_loss": 3.495869856325105, + "tokens_seen": 1761280000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023554663991975927, + "loss": 1.4428, + "theoretical_loss": 3.4958588284467678, + "tokens_seen": 1761345536 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023553660982948845, + "loss": 1.4186, + "theoretical_loss": 3.4958478010936336, + "tokens_seen": 1761411072 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023552657973921766, + "loss": 1.3462, + "theoretical_loss": 3.4958367742656575, + "tokens_seen": 1761476608 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023551654964894684, + "loss": 1.3593, + "theoretical_loss": 3.4958257479627957, + "tokens_seen": 1761542144 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023550651955867602, + "loss": 1.2571, + "theoretical_loss": 3.495814722185002, + "tokens_seen": 1761607680 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002354964894684052, + "loss": 1.4188, + "theoretical_loss": 3.4958036969322337, + "tokens_seen": 1761673216 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002354864593781344, + "loss": 1.3532, + "theoretical_loss": 3.495792672204445, + "tokens_seen": 1761738752 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002354764292878636, + "loss": 1.3651, + "theoretical_loss": 3.495781648001592, + "tokens_seen": 1761804288 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023546639919759277, + "loss": 1.3645, + "theoretical_loss": 3.4957706243236295, + "tokens_seen": 1761869824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023545636910732195, + "loss": 1.451, + "theoretical_loss": 3.495759601170514, + "tokens_seen": 1761935360 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023544633901705113, + "loss": 1.4591, + "theoretical_loss": 3.4957485785422, + "tokens_seen": 1762000896 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023543630892678034, + "loss": 1.3824, + "theoretical_loss": 3.4957375564386437, + "tokens_seen": 1762066432 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023542627883650952, + "loss": 1.3383, + "theoretical_loss": 3.4957265348598, + "tokens_seen": 1762131968 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002354162487462387, + "loss": 1.4192, + "theoretical_loss": 3.495715513805625, + "tokens_seen": 1762197504 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002354062186559679, + "loss": 1.361, + "theoretical_loss": 3.495704493276074, + "tokens_seen": 1762263040 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002353961885656971, + "loss": 1.4441, + "theoretical_loss": 3.4956934732711025, + "tokens_seen": 1762328576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023538615847542628, + "loss": 1.352, + "theoretical_loss": 3.4956824537906654, + "tokens_seen": 1762394112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023537612838515546, + "loss": 1.4087, + "theoretical_loss": 3.4956714348347195, + "tokens_seen": 1762459648 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023536609829488464, + "loss": 1.3536, + "theoretical_loss": 3.4956604164032195, + "tokens_seen": 1762525184 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023535606820461387, + "loss": 1.2498, + "theoretical_loss": 3.4956493984961208, + "tokens_seen": 1762590720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023534603811434305, + "loss": 1.3094, + "theoretical_loss": 3.4956383811133795, + "tokens_seen": 1762656256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023533600802407224, + "loss": 1.4173, + "theoretical_loss": 3.4956273642549514, + "tokens_seen": 1762721792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023532597793380142, + "loss": 1.2948, + "theoretical_loss": 3.495616347920791, + "tokens_seen": 1762787328 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002353159478435306, + "loss": 1.3742, + "theoretical_loss": 3.4956053321108547, + "tokens_seen": 1762852864 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 643830, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.178256869316101, + "objective/train/theoretical_loss": 3.4955943168250974, + "objective/train/tokens_used": 1783378400, + "theoretical_loss": 3.4955943168250974, + "tokens_seen": 1762918400 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002353059177532598, + "loss": 1.2746, + "theoretical_loss": 3.4955943168250974, + "tokens_seen": 1762918400 + }, + { + "epoch": 0.53, + "learning_rate": 0.000235295887662989, + "loss": 1.4718, + "theoretical_loss": 3.4955833020634755, + "tokens_seen": 1762983936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023528585757271817, + "loss": 1.5292, + "theoretical_loss": 3.4955722878259436, + "tokens_seen": 1763049472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023527582748244735, + "loss": 1.2805, + "theoretical_loss": 3.495561274112458, + "tokens_seen": 1763115008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023526579739217656, + "loss": 1.4483, + "theoretical_loss": 3.4955502609229745, + "tokens_seen": 1763180544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023525576730190574, + "loss": 1.306, + "theoretical_loss": 3.4955392482574474, + "tokens_seen": 1763246080 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023524573721163492, + "loss": 1.4092, + "theoretical_loss": 3.4955282361158337, + "tokens_seen": 1763311616 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002352357071213641, + "loss": 1.4358, + "theoretical_loss": 3.495517224498089, + "tokens_seen": 1763377152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023522567703109328, + "loss": 1.3041, + "theoretical_loss": 3.495506213404168, + "tokens_seen": 1763442688 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002352156469408225, + "loss": 1.3888, + "theoretical_loss": 3.495495202834027, + "tokens_seen": 1763508224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023520561685055167, + "loss": 1.3173, + "theoretical_loss": 3.495484192787621, + "tokens_seen": 1763573760 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023519558676028085, + "loss": 1.2391, + "theoretical_loss": 3.495473183264906, + "tokens_seen": 1763639296 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023518555667001003, + "loss": 1.3415, + "theoretical_loss": 3.495462174265838, + "tokens_seen": 1763704832 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023517552657973924, + "loss": 1.3374, + "theoretical_loss": 3.4954511657903717, + "tokens_seen": 1763770368 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023516549648946842, + "loss": 1.3311, + "theoretical_loss": 3.4954401578384635, + "tokens_seen": 1763835904 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002351554663991976, + "loss": 1.4217, + "theoretical_loss": 3.495429150410069, + "tokens_seen": 1763901440 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023514543630892679, + "loss": 1.3318, + "theoretical_loss": 3.4954181435051437, + "tokens_seen": 1763966976 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023513540621865597, + "loss": 1.3518, + "theoretical_loss": 3.495407137123643, + "tokens_seen": 1764032512 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023512537612838517, + "loss": 1.3272, + "theoretical_loss": 3.495396131265523, + "tokens_seen": 1764098048 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023511534603811436, + "loss": 1.4448, + "theoretical_loss": 3.495385125930739, + "tokens_seen": 1764163584 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023510531594784354, + "loss": 1.3739, + "theoretical_loss": 3.495374121119247, + "tokens_seen": 1764229120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023509528585757272, + "loss": 1.3551, + "theoretical_loss": 3.4953631168310024, + "tokens_seen": 1764294656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023508525576730193, + "loss": 1.2715, + "theoretical_loss": 3.4953521130659615, + "tokens_seen": 1764360192 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002350752256770311, + "loss": 1.2839, + "theoretical_loss": 3.4953411098240785, + "tokens_seen": 1764425728 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002350651955867603, + "loss": 1.3028, + "theoretical_loss": 3.4953301071053113, + "tokens_seen": 1764491264 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023505516549648947, + "loss": 1.2819, + "theoretical_loss": 3.4953191049096137, + "tokens_seen": 1764556800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023504513540621865, + "loss": 1.4322, + "theoretical_loss": 3.495308103236942, + "tokens_seen": 1764622336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023503510531594786, + "loss": 1.4468, + "theoretical_loss": 3.4952971020872527, + "tokens_seen": 1764687872 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023502507522567704, + "loss": 1.2846, + "theoretical_loss": 3.4952861014605006, + "tokens_seen": 1764753408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023501504513540622, + "loss": 1.4444, + "theoretical_loss": 3.4952751013566417, + "tokens_seen": 1764818944 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002350050150451354, + "loss": 1.3144, + "theoretical_loss": 3.4952641017756316, + "tokens_seen": 1764884480 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002349949849548646, + "loss": 1.3879, + "theoretical_loss": 3.495253102717426, + "tokens_seen": 1764950016 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002349849548645938, + "loss": 1.3309, + "theoretical_loss": 3.495242104181981, + "tokens_seen": 1765015552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023497492477432297, + "loss": 1.3995, + "theoretical_loss": 3.495231106169252, + "tokens_seen": 1765081088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023496489468405215, + "loss": 1.37, + "theoretical_loss": 3.495220108679195, + "tokens_seen": 1765146624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023495486459378134, + "loss": 1.4453, + "theoretical_loss": 3.4952091117117656, + "tokens_seen": 1765212160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023494483450351054, + "loss": 1.4181, + "theoretical_loss": 3.49519811526692, + "tokens_seen": 1765277696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00023493480441323972, + "loss": 1.4059, + "theoretical_loss": 3.495187119344613, + "tokens_seen": 1765343232 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002349247743229689, + "loss": 1.3776, + "theoretical_loss": 3.495176123944801, + "tokens_seen": 1765408768 + }, + { + "epoch": 0.53, + "learning_rate": 0.0002349147442326981, + "loss": 1.4103, + "theoretical_loss": 3.4951651290674395, + "tokens_seen": 1765474304 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002349047141424273, + "loss": 1.4298, + "theoretical_loss": 3.495154134712485, + "tokens_seen": 1765539840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023489468405215648, + "loss": 1.2867, + "theoretical_loss": 3.4951431408798928, + "tokens_seen": 1765605376 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023488465396188566, + "loss": 1.2642, + "theoretical_loss": 3.4951321475696187, + "tokens_seen": 1765670912 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023487462387161484, + "loss": 1.2746, + "theoretical_loss": 3.495121154781618, + "tokens_seen": 1765736448 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023486459378134402, + "loss": 1.3854, + "theoretical_loss": 3.4951101625158474, + "tokens_seen": 1765801984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023485456369107323, + "loss": 1.3384, + "theoretical_loss": 3.495099170772262, + "tokens_seen": 1765867520 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002348445336008024, + "loss": 1.4243, + "theoretical_loss": 3.4950881795508177, + "tokens_seen": 1765933056 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002348345035105316, + "loss": 1.3233, + "theoretical_loss": 3.4950771888514716, + "tokens_seen": 1765998592 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023482447342026077, + "loss": 1.3974, + "theoretical_loss": 3.495066198674177, + "tokens_seen": 1766064128 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023481444332998998, + "loss": 1.4118, + "theoretical_loss": 3.495055209018892, + "tokens_seen": 1766129664 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 644977, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5729326009750366, + "objective/train/theoretical_loss": 3.495044219885572, + "objective/train/tokens_used": 1786655200, + "theoretical_loss": 3.495044219885572, + "tokens_seen": 1766195200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023480441323971916, + "loss": 1.4431, + "theoretical_loss": 3.495044219885572, + "tokens_seen": 1766195200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023479438314944834, + "loss": 1.2239, + "theoretical_loss": 3.4950332312741716, + "tokens_seen": 1766260736 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023478435305917752, + "loss": 1.4319, + "theoretical_loss": 3.495022243184648, + "tokens_seen": 1766326272 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023477432296890673, + "loss": 1.3463, + "theoretical_loss": 3.495011255616957, + "tokens_seen": 1766391808 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002347642928786359, + "loss": 1.3687, + "theoretical_loss": 3.4950002685710535, + "tokens_seen": 1766457344 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002347542627883651, + "loss": 1.4146, + "theoretical_loss": 3.4949892820468937, + "tokens_seen": 1766522880 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023474423269809427, + "loss": 1.3894, + "theoretical_loss": 3.494978296044434, + "tokens_seen": 1766588416 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023473420260782346, + "loss": 1.4207, + "theoretical_loss": 3.49496731056363, + "tokens_seen": 1766653952 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023472417251755266, + "loss": 1.3681, + "theoretical_loss": 3.494956325604437, + "tokens_seen": 1766719488 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023471414242728184, + "loss": 1.4337, + "theoretical_loss": 3.4949453411668117, + "tokens_seen": 1766785024 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023470411233701103, + "loss": 1.3548, + "theoretical_loss": 3.4949343572507097, + "tokens_seen": 1766850560 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002346940822467402, + "loss": 1.3163, + "theoretical_loss": 3.4949233738560874, + "tokens_seen": 1766916096 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023468405215646942, + "loss": 1.387, + "theoretical_loss": 3.4949123909828996, + "tokens_seen": 1766981632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002346740220661986, + "loss": 1.4914, + "theoretical_loss": 3.4949014086311028, + "tokens_seen": 1767047168 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023466399197592778, + "loss": 1.3455, + "theoretical_loss": 3.4948904268006533, + "tokens_seen": 1767112704 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023465396188565696, + "loss": 1.3928, + "theoretical_loss": 3.494879445491506, + "tokens_seen": 1767178240 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023464393179538614, + "loss": 1.3482, + "theoretical_loss": 3.494868464703618, + "tokens_seen": 1767243776 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023463390170511535, + "loss": 1.3937, + "theoretical_loss": 3.4948574844369444, + "tokens_seen": 1767309312 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023462387161484453, + "loss": 1.3807, + "theoretical_loss": 3.4948465046914414, + "tokens_seen": 1767374848 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002346138415245737, + "loss": 1.4302, + "theoretical_loss": 3.494835525467065, + "tokens_seen": 1767440384 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023460381143430292, + "loss": 1.2699, + "theoretical_loss": 3.494824546763771, + "tokens_seen": 1767505920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023459378134403213, + "loss": 1.4528, + "theoretical_loss": 3.494813568581516, + "tokens_seen": 1767571456 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002345837512537613, + "loss": 1.3178, + "theoretical_loss": 3.494802590920255, + "tokens_seen": 1767636992 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002345737211634905, + "loss": 1.3585, + "theoretical_loss": 3.494791613779944, + "tokens_seen": 1767702528 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023456369107321967, + "loss": 1.3627, + "theoretical_loss": 3.49478063716054, + "tokens_seen": 1767768064 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023455366098294885, + "loss": 1.1763, + "theoretical_loss": 3.494769661061998, + "tokens_seen": 1767833600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023454363089267806, + "loss": 1.3088, + "theoretical_loss": 3.494758685484274, + "tokens_seen": 1767899136 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023453360080240724, + "loss": 1.3414, + "theoretical_loss": 3.4947477104273243, + "tokens_seen": 1767964672 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023452357071213642, + "loss": 1.3301, + "theoretical_loss": 3.494736735891105, + "tokens_seen": 1768030208 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002345135406218656, + "loss": 1.3602, + "theoretical_loss": 3.494725761875572, + "tokens_seen": 1768095744 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002345035105315948, + "loss": 1.3259, + "theoretical_loss": 3.494714788380681, + "tokens_seen": 1768161280 + }, + { + "epoch": 0.54, + "learning_rate": 0.000234493480441324, + "loss": 1.4953, + "theoretical_loss": 3.494703815406388, + "tokens_seen": 1768226816 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023448345035105317, + "loss": 1.4127, + "theoretical_loss": 3.4946928429526496, + "tokens_seen": 1768292352 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023447342026078235, + "loss": 1.2662, + "theoretical_loss": 3.494681871019422, + "tokens_seen": 1768357888 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023446339017051154, + "loss": 1.3822, + "theoretical_loss": 3.4946708996066596, + "tokens_seen": 1768423424 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023445336008024074, + "loss": 1.4261, + "theoretical_loss": 3.49465992871432, + "tokens_seen": 1768488960 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023444332998996992, + "loss": 1.3662, + "theoretical_loss": 3.494648958342359, + "tokens_seen": 1768554496 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002344332998996991, + "loss": 1.2684, + "theoretical_loss": 3.494637988490732, + "tokens_seen": 1768620032 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002344232698094283, + "loss": 1.366, + "theoretical_loss": 3.4946270191593958, + "tokens_seen": 1768685568 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002344132397191575, + "loss": 1.3609, + "theoretical_loss": 3.4946160503483057, + "tokens_seen": 1768751104 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023440320962888668, + "loss": 1.3868, + "theoretical_loss": 3.494605082057418, + "tokens_seen": 1768816640 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023439317953861586, + "loss": 1.3954, + "theoretical_loss": 3.4945941142866888, + "tokens_seen": 1768882176 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023438314944834504, + "loss": 1.3896, + "theoretical_loss": 3.494583147036075, + "tokens_seen": 1768947712 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023437311935807422, + "loss": 1.3557, + "theoretical_loss": 3.494572180305531, + "tokens_seen": 1769013248 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023436308926780343, + "loss": 1.4574, + "theoretical_loss": 3.494561214095014, + "tokens_seen": 1769078784 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002343530591775326, + "loss": 1.3964, + "theoretical_loss": 3.4945502484044804, + "tokens_seen": 1769144320 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002343430290872618, + "loss": 1.2434, + "theoretical_loss": 3.494539283233885, + "tokens_seen": 1769209856 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023433299899699097, + "loss": 1.2955, + "theoretical_loss": 3.4945283185831846, + "tokens_seen": 1769275392 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023432296890672018, + "loss": 1.3029, + "theoretical_loss": 3.4945173544523356, + "tokens_seen": 1769340928 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023431293881644936, + "loss": 1.2531, + "theoretical_loss": 3.494506390841294, + "tokens_seen": 1769406464 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 646001, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2574732303619385, + "objective/train/theoretical_loss": 3.4944954277500155, + "objective/train/tokens_used": 1789932000, + "theoretical_loss": 3.4944954277500155, + "tokens_seen": 1769472000 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023430290872617854, + "loss": 1.2696, + "theoretical_loss": 3.4944954277500155, + "tokens_seen": 1769472000 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023429287863590772, + "loss": 1.3958, + "theoretical_loss": 3.4944844651784566, + "tokens_seen": 1769537536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002342828485456369, + "loss": 1.3448, + "theoretical_loss": 3.4944735031265726, + "tokens_seen": 1769603072 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002342728184553661, + "loss": 1.3122, + "theoretical_loss": 3.494462541594321, + "tokens_seen": 1769668608 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002342627883650953, + "loss": 1.2812, + "theoretical_loss": 3.4944515805816563, + "tokens_seen": 1769734144 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023425275827482447, + "loss": 1.4554, + "theoretical_loss": 3.494440620088536, + "tokens_seen": 1769799680 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023424272818455366, + "loss": 1.349, + "theoretical_loss": 3.494429660114916, + "tokens_seen": 1769865216 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023423269809428286, + "loss": 1.3392, + "theoretical_loss": 3.494418700660752, + "tokens_seen": 1769930752 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023422266800401204, + "loss": 1.3494, + "theoretical_loss": 3.4944077417260004, + "tokens_seen": 1769996288 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023421263791374123, + "loss": 1.4754, + "theoretical_loss": 3.4943967833106173, + "tokens_seen": 1770061824 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002342026078234704, + "loss": 1.2776, + "theoretical_loss": 3.4943858254145583, + "tokens_seen": 1770127360 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023419257773319962, + "loss": 1.4464, + "theoretical_loss": 3.494374868037781, + "tokens_seen": 1770192896 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002341825476429288, + "loss": 1.4079, + "theoretical_loss": 3.49436391118024, + "tokens_seen": 1770258432 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023417251755265798, + "loss": 1.3294, + "theoretical_loss": 3.4943529548418923, + "tokens_seen": 1770323968 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023416248746238716, + "loss": 1.397, + "theoretical_loss": 3.4943419990226943, + "tokens_seen": 1770389504 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023415245737211634, + "loss": 1.2737, + "theoretical_loss": 3.4943310437226014, + "tokens_seen": 1770455040 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023414242728184555, + "loss": 1.4936, + "theoretical_loss": 3.49432008894157, + "tokens_seen": 1770520576 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023413239719157473, + "loss": 1.3331, + "theoretical_loss": 3.4943091346795567, + "tokens_seen": 1770586112 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002341223671013039, + "loss": 1.448, + "theoretical_loss": 3.494298180936518, + "tokens_seen": 1770651648 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002341123370110331, + "loss": 1.3607, + "theoretical_loss": 3.4942872277124093, + "tokens_seen": 1770717184 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002341023069207623, + "loss": 1.2808, + "theoretical_loss": 3.494276275007187, + "tokens_seen": 1770782720 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023409227683049148, + "loss": 1.3898, + "theoretical_loss": 3.494265322820808, + "tokens_seen": 1770848256 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023408224674022066, + "loss": 1.3868, + "theoretical_loss": 3.494254371153227, + "tokens_seen": 1770913792 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023407221664994984, + "loss": 1.3798, + "theoretical_loss": 3.4942434200044015, + "tokens_seen": 1770979328 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023406218655967902, + "loss": 1.3604, + "theoretical_loss": 3.494232469374287, + "tokens_seen": 1771044864 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023405215646940823, + "loss": 1.3081, + "theoretical_loss": 3.494221519262841, + "tokens_seen": 1771110400 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002340421263791374, + "loss": 1.3314, + "theoretical_loss": 3.4942105696700185, + "tokens_seen": 1771175936 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002340320962888666, + "loss": 1.3462, + "theoretical_loss": 3.4941996205957766, + "tokens_seen": 1771241472 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023402206619859578, + "loss": 1.3687, + "theoretical_loss": 3.4941886720400706, + "tokens_seen": 1771307008 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023401203610832498, + "loss": 1.3538, + "theoretical_loss": 3.4941777240028573, + "tokens_seen": 1771372544 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023400200601805416, + "loss": 1.3735, + "theoretical_loss": 3.494166776484093, + "tokens_seen": 1771438080 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023399197592778335, + "loss": 1.3503, + "theoretical_loss": 3.494155829483734, + "tokens_seen": 1771503616 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023398194583751253, + "loss": 1.3232, + "theoretical_loss": 3.494144883001736, + "tokens_seen": 1771569152 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002339719157472417, + "loss": 1.3997, + "theoretical_loss": 3.4941339370380566, + "tokens_seen": 1771634688 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023396188565697092, + "loss": 1.4068, + "theoretical_loss": 3.4941229915926506, + "tokens_seen": 1771700224 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002339518555667001, + "loss": 1.3657, + "theoretical_loss": 3.4941120466654745, + "tokens_seen": 1771765760 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023394182547642928, + "loss": 1.4391, + "theoretical_loss": 3.4941011022564856, + "tokens_seen": 1771831296 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023393179538615846, + "loss": 1.3622, + "theoretical_loss": 3.4940901583656396, + "tokens_seen": 1771896832 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023392176529588767, + "loss": 1.38, + "theoretical_loss": 3.494079214992893, + "tokens_seen": 1771962368 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023391173520561685, + "loss": 1.3925, + "theoretical_loss": 3.4940682721382017, + "tokens_seen": 1772027904 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023390170511534603, + "loss": 1.3655, + "theoretical_loss": 3.4940573298015223, + "tokens_seen": 1772093440 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002338916750250752, + "loss": 1.3335, + "theoretical_loss": 3.494046387982811, + "tokens_seen": 1772158976 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002338816449348044, + "loss": 1.3658, + "theoretical_loss": 3.494035446682024, + "tokens_seen": 1772224512 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002338716148445336, + "loss": 1.3013, + "theoretical_loss": 3.494024505899118, + "tokens_seen": 1772290048 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023386158475426278, + "loss": 1.4237, + "theoretical_loss": 3.4940135656340496, + "tokens_seen": 1772355584 + }, + { + "epoch": 0.54, + "learning_rate": 0.000233851554663992, + "loss": 1.4126, + "theoretical_loss": 3.4940026258867745, + "tokens_seen": 1772421120 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023384152457372117, + "loss": 1.3357, + "theoretical_loss": 3.4939916866572487, + "tokens_seen": 1772486656 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023383149448345038, + "loss": 1.3804, + "theoretical_loss": 3.4939807479454297, + "tokens_seen": 1772552192 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023382146439317956, + "loss": 1.2951, + "theoretical_loss": 3.493969809751273, + "tokens_seen": 1772617728 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023381143430290874, + "loss": 1.3566, + "theoretical_loss": 3.4939588720747357, + "tokens_seen": 1772683264 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 647211, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5260331630706787, + "objective/train/theoretical_loss": 3.493947934915773, + "objective/train/tokens_used": 1793208800, + "theoretical_loss": 3.493947934915773, + "tokens_seen": 1772748800 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023380140421263792, + "loss": 1.4395, + "theoretical_loss": 3.493947934915773, + "tokens_seen": 1772748800 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002337913741223671, + "loss": 1.3493, + "theoretical_loss": 3.493936998274343, + "tokens_seen": 1772814336 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002337813440320963, + "loss": 1.2838, + "theoretical_loss": 3.4939260621504005, + "tokens_seen": 1772879872 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002337713139418255, + "loss": 1.3344, + "theoretical_loss": 3.493915126543902, + "tokens_seen": 1772945408 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023376128385155467, + "loss": 1.4592, + "theoretical_loss": 3.493904191454805, + "tokens_seen": 1773010944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023375125376128386, + "loss": 1.4424, + "theoretical_loss": 3.493893256883065, + "tokens_seen": 1773076480 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023374122367101306, + "loss": 1.2639, + "theoretical_loss": 3.4938823228286386, + "tokens_seen": 1773142016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023373119358074224, + "loss": 1.4488, + "theoretical_loss": 3.493871389291482, + "tokens_seen": 1773207552 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023372116349047143, + "loss": 1.3416, + "theoretical_loss": 3.4938604562715527, + "tokens_seen": 1773273088 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002337111334002006, + "loss": 1.2555, + "theoretical_loss": 3.4938495237688056, + "tokens_seen": 1773338624 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023370110330992982, + "loss": 1.4472, + "theoretical_loss": 3.4938385917831978, + "tokens_seen": 1773404160 + }, + { + "epoch": 0.54, + "learning_rate": 0.000233691073219659, + "loss": 1.4242, + "theoretical_loss": 3.493827660314686, + "tokens_seen": 1773469696 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023368104312938818, + "loss": 1.1693, + "theoretical_loss": 3.4938167293632265, + "tokens_seen": 1773535232 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023367101303911736, + "loss": 1.365, + "theoretical_loss": 3.493805798928775, + "tokens_seen": 1773600768 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023366098294884654, + "loss": 1.3682, + "theoretical_loss": 3.493794869011289, + "tokens_seen": 1773666304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023365095285857575, + "loss": 1.3161, + "theoretical_loss": 3.493783939610725, + "tokens_seen": 1773731840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023364092276830493, + "loss": 1.3441, + "theoretical_loss": 3.493773010727038, + "tokens_seen": 1773797376 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002336308926780341, + "loss": 1.3478, + "theoretical_loss": 3.493762082360186, + "tokens_seen": 1773862912 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002336208625877633, + "loss": 1.3069, + "theoretical_loss": 3.4937511545101243, + "tokens_seen": 1773928448 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002336108324974925, + "loss": 1.4101, + "theoretical_loss": 3.4937402271768105, + "tokens_seen": 1773993984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023360080240722168, + "loss": 1.3412, + "theoretical_loss": 3.493729300360201, + "tokens_seen": 1774059520 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023359077231695086, + "loss": 1.2192, + "theoretical_loss": 3.4937183740602507, + "tokens_seen": 1774125056 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023358074222668004, + "loss": 1.2701, + "theoretical_loss": 3.4937074482769175, + "tokens_seen": 1774190592 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023357071213640922, + "loss": 1.444, + "theoretical_loss": 3.4936965230101578, + "tokens_seen": 1774256128 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023356068204613843, + "loss": 1.3044, + "theoretical_loss": 3.493685598259928, + "tokens_seen": 1774321664 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002335506519558676, + "loss": 1.3731, + "theoretical_loss": 3.4936746740261846, + "tokens_seen": 1774387200 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002335406218655968, + "loss": 1.335, + "theoretical_loss": 3.4936637503088837, + "tokens_seen": 1774452736 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023353059177532598, + "loss": 1.2588, + "theoretical_loss": 3.493652827107982, + "tokens_seen": 1774518272 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023352056168505518, + "loss": 1.2546, + "theoretical_loss": 3.493641904423437, + "tokens_seen": 1774583808 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023351053159478437, + "loss": 1.326, + "theoretical_loss": 3.493630982255203, + "tokens_seen": 1774649344 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023350050150451355, + "loss": 1.3563, + "theoretical_loss": 3.493620060603239, + "tokens_seen": 1774714880 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023349047141424273, + "loss": 1.3518, + "theoretical_loss": 3.4936091394674995, + "tokens_seen": 1774780416 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002334804413239719, + "loss": 1.3914, + "theoretical_loss": 3.4935982188479424, + "tokens_seen": 1774845952 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023347041123370112, + "loss": 1.3479, + "theoretical_loss": 3.493587298744524, + "tokens_seen": 1774911488 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002334603811434303, + "loss": 1.3075, + "theoretical_loss": 3.4935763791572003, + "tokens_seen": 1774977024 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023345035105315948, + "loss": 1.2996, + "theoretical_loss": 3.4935654600859283, + "tokens_seen": 1775042560 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023344032096288866, + "loss": 1.3173, + "theoretical_loss": 3.4935545415306644, + "tokens_seen": 1775108096 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023343029087261787, + "loss": 1.3412, + "theoretical_loss": 3.4935436234913655, + "tokens_seen": 1775173632 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023342026078234705, + "loss": 1.409, + "theoretical_loss": 3.4935327059679873, + "tokens_seen": 1775239168 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023341023069207623, + "loss": 1.3338, + "theoretical_loss": 3.4935217889604875, + "tokens_seen": 1775304704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002334002006018054, + "loss": 1.432, + "theoretical_loss": 3.493510872468822, + "tokens_seen": 1775370240 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002333901705115346, + "loss": 1.4331, + "theoretical_loss": 3.493499956492948, + "tokens_seen": 1775435776 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002333801404212638, + "loss": 1.4399, + "theoretical_loss": 3.493489041032821, + "tokens_seen": 1775501312 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023337011033099298, + "loss": 1.3092, + "theoretical_loss": 3.493478126088398, + "tokens_seen": 1775566848 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023336008024072216, + "loss": 1.2888, + "theoretical_loss": 3.4934672116596364, + "tokens_seen": 1775632384 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023335005015045134, + "loss": 1.2502, + "theoretical_loss": 3.4934562977464925, + "tokens_seen": 1775697920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023334002006018055, + "loss": 1.3455, + "theoretical_loss": 3.493445384348922, + "tokens_seen": 1775763456 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023332998996990973, + "loss": 1.3188, + "theoretical_loss": 3.493434471466882, + "tokens_seen": 1775828992 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023331995987963891, + "loss": 1.3676, + "theoretical_loss": 3.49342355910033, + "tokens_seen": 1775894528 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002333099297893681, + "loss": 1.335, + "theoretical_loss": 3.4934126472492215, + "tokens_seen": 1775960064 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 648417, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3500685691833496, + "objective/train/theoretical_loss": 3.4934017359135137, + "objective/train/tokens_used": 1796485600, + "theoretical_loss": 3.4934017359135137, + "tokens_seen": 1776025600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023329989969909728, + "loss": 1.4166, + "theoretical_loss": 3.4934017359135137, + "tokens_seen": 1776025600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023328986960882649, + "loss": 1.307, + "theoretical_loss": 3.4933908250931633, + "tokens_seen": 1776091136 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023327983951855567, + "loss": 1.2518, + "theoretical_loss": 3.493379914788126, + "tokens_seen": 1776156672 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023326980942828485, + "loss": 1.3298, + "theoretical_loss": 3.49336900499836, + "tokens_seen": 1776222208 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023325977933801403, + "loss": 1.5139, + "theoretical_loss": 3.4933580957238206, + "tokens_seen": 1776287744 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023324974924774324, + "loss": 1.277, + "theoretical_loss": 3.4933471869644652, + "tokens_seen": 1776353280 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023323971915747242, + "loss": 1.2805, + "theoretical_loss": 3.4933362787202507, + "tokens_seen": 1776418816 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002332296890672016, + "loss": 1.4068, + "theoretical_loss": 3.4933253709911325, + "tokens_seen": 1776484352 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023321965897693078, + "loss": 1.3859, + "theoretical_loss": 3.493314463777069, + "tokens_seen": 1776549888 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023320962888665996, + "loss": 1.3642, + "theoretical_loss": 3.493303557078015, + "tokens_seen": 1776615424 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023319959879638917, + "loss": 1.3278, + "theoretical_loss": 3.493292650893929, + "tokens_seen": 1776680960 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023318956870611835, + "loss": 1.4694, + "theoretical_loss": 3.4932817452247664, + "tokens_seen": 1776746496 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023317953861584753, + "loss": 1.3359, + "theoretical_loss": 3.493270840070484, + "tokens_seen": 1776812032 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002331695085255767, + "loss": 1.3805, + "theoretical_loss": 3.4932599354310403, + "tokens_seen": 1776877568 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023315947843530592, + "loss": 1.3315, + "theoretical_loss": 3.493249031306389, + "tokens_seen": 1776943104 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002331494483450351, + "loss": 1.4116, + "theoretical_loss": 3.4932381276964897, + "tokens_seen": 1777008640 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023313941825476428, + "loss": 1.4534, + "theoretical_loss": 3.4932272246012968, + "tokens_seen": 1777074176 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023312938816449346, + "loss": 1.3398, + "theoretical_loss": 3.4932163220207686, + "tokens_seen": 1777139712 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023311935807422267, + "loss": 1.2521, + "theoretical_loss": 3.493205419954861, + "tokens_seen": 1777205248 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023310932798395185, + "loss": 1.2917, + "theoretical_loss": 3.4931945184035307, + "tokens_seen": 1777270784 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023309929789368106, + "loss": 1.2226, + "theoretical_loss": 3.493183617366735, + "tokens_seen": 1777336320 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023308926780341024, + "loss": 1.222, + "theoretical_loss": 3.4931727168444304, + "tokens_seen": 1777401856 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023307923771313942, + "loss": 1.4576, + "theoretical_loss": 3.4931618168365737, + "tokens_seen": 1777467392 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023306920762286863, + "loss": 1.3251, + "theoretical_loss": 3.4931509173431214, + "tokens_seen": 1777532928 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023305917753259781, + "loss": 1.3646, + "theoretical_loss": 3.4931400183640307, + "tokens_seen": 1777598464 + }, + { + "epoch": 0.54, + "learning_rate": 0.000233049147442327, + "loss": 1.3663, + "theoretical_loss": 3.4931291198992582, + "tokens_seen": 1777664000 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023303911735205618, + "loss": 1.2243, + "theoretical_loss": 3.49311822194876, + "tokens_seen": 1777729536 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023302908726178538, + "loss": 1.2535, + "theoretical_loss": 3.493107324512494, + "tokens_seen": 1777795072 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023301905717151457, + "loss": 1.3474, + "theoretical_loss": 3.493096427590416, + "tokens_seen": 1777860608 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023300902708124375, + "loss": 1.3059, + "theoretical_loss": 3.4930855311824835, + "tokens_seen": 1777926144 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023299899699097293, + "loss": 1.2785, + "theoretical_loss": 3.493074635288653, + "tokens_seen": 1777991680 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002329889669007021, + "loss": 1.3981, + "theoretical_loss": 3.493063739908881, + "tokens_seen": 1778057216 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023297893681043132, + "loss": 1.3971, + "theoretical_loss": 3.493052845043125, + "tokens_seen": 1778122752 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002329689067201605, + "loss": 1.3585, + "theoretical_loss": 3.493041950691341, + "tokens_seen": 1778188288 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023295887662988968, + "loss": 1.3505, + "theoretical_loss": 3.4930310568534866, + "tokens_seen": 1778253824 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023294884653961886, + "loss": 1.4231, + "theoretical_loss": 3.4930201635295175, + "tokens_seen": 1778319360 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023293881644934807, + "loss": 1.4375, + "theoretical_loss": 3.493009270719392, + "tokens_seen": 1778384896 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023292878635907725, + "loss": 1.3122, + "theoretical_loss": 3.4929983784230663, + "tokens_seen": 1778450432 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023291875626880643, + "loss": 1.3843, + "theoretical_loss": 3.492987486640496, + "tokens_seen": 1778515968 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002329087261785356, + "loss": 1.3337, + "theoretical_loss": 3.4929765953716405, + "tokens_seen": 1778581504 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002328986960882648, + "loss": 1.4576, + "theoretical_loss": 3.4929657046164544, + "tokens_seen": 1778647040 + }, + { + "epoch": 0.54, + "learning_rate": 0.000232888665997994, + "loss": 1.4074, + "theoretical_loss": 3.4929548143748956, + "tokens_seen": 1778712576 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023287863590772318, + "loss": 1.3407, + "theoretical_loss": 3.4929439246469203, + "tokens_seen": 1778778112 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023286860581745236, + "loss": 1.2924, + "theoretical_loss": 3.492933035432486, + "tokens_seen": 1778843648 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023285857572718154, + "loss": 1.3013, + "theoretical_loss": 3.492922146731549, + "tokens_seen": 1778909184 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023284854563691075, + "loss": 1.3417, + "theoretical_loss": 3.4929112585440674, + "tokens_seen": 1778974720 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023283851554663993, + "loss": 1.3179, + "theoretical_loss": 3.492900370869996, + "tokens_seen": 1779040256 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023282848545636911, + "loss": 1.3864, + "theoretical_loss": 3.492889483709294, + "tokens_seen": 1779105792 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002328184553660983, + "loss": 1.4098, + "theoretical_loss": 3.4928785970619165, + "tokens_seen": 1779171328 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023280842527582748, + "loss": 1.3581, + "theoretical_loss": 3.492867710927821, + "tokens_seen": 1779236864 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 649543, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4306217432022095, + "objective/train/theoretical_loss": 3.4928568253069643, + "objective/train/tokens_used": 1799762400, + "theoretical_loss": 3.4928568253069643, + "tokens_seen": 1779302400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023279839518555669, + "loss": 1.3676, + "theoretical_loss": 3.4928568253069643, + "tokens_seen": 1779302400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023278836509528587, + "loss": 1.3596, + "theoretical_loss": 3.4928459401993033, + "tokens_seen": 1779367936 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023277833500501505, + "loss": 1.3113, + "theoretical_loss": 3.492835055604796, + "tokens_seen": 1779433472 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023276830491474423, + "loss": 1.2706, + "theoretical_loss": 3.492824171523397, + "tokens_seen": 1779499008 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023275827482447344, + "loss": 1.2807, + "theoretical_loss": 3.492813287955065, + "tokens_seen": 1779564544 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023274824473420262, + "loss": 1.3607, + "theoretical_loss": 3.4928024048997566, + "tokens_seen": 1779630080 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002327382146439318, + "loss": 1.4361, + "theoretical_loss": 3.492791522357429, + "tokens_seen": 1779695616 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023272818455366098, + "loss": 1.2776, + "theoretical_loss": 3.4927806403280384, + "tokens_seen": 1779761152 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023271815446339016, + "loss": 1.3516, + "theoretical_loss": 3.4927697588115416, + "tokens_seen": 1779826688 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023270812437311937, + "loss": 1.3649, + "theoretical_loss": 3.492758877807897, + "tokens_seen": 1779892224 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023269809428284855, + "loss": 1.4748, + "theoretical_loss": 3.4927479973170596, + "tokens_seen": 1779957760 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023268806419257773, + "loss": 1.4276, + "theoretical_loss": 3.4927371173389874, + "tokens_seen": 1780023296 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002326780341023069, + "loss": 1.5644, + "theoretical_loss": 3.4927262378736375, + "tokens_seen": 1780088832 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023266800401203612, + "loss": 1.3634, + "theoretical_loss": 3.4927153589209667, + "tokens_seen": 1780154368 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002326579739217653, + "loss": 1.3114, + "theoretical_loss": 3.492704480480932, + "tokens_seen": 1780219904 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023264794383149448, + "loss": 1.3902, + "theoretical_loss": 3.49269360255349, + "tokens_seen": 1780285440 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023263791374122366, + "loss": 1.2148, + "theoretical_loss": 3.492682725138598, + "tokens_seen": 1780350976 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023262788365095287, + "loss": 1.3837, + "theoretical_loss": 3.492671848236213, + "tokens_seen": 1780416512 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023261785356068205, + "loss": 1.295, + "theoretical_loss": 3.4926609718462918, + "tokens_seen": 1780482048 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023260782347041124, + "loss": 1.3034, + "theoretical_loss": 3.4926500959687914, + "tokens_seen": 1780547584 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023259779338014042, + "loss": 1.2905, + "theoretical_loss": 3.492639220603669, + "tokens_seen": 1780613120 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002325877632898696, + "loss": 1.348, + "theoretical_loss": 3.4926283457508815, + "tokens_seen": 1780678656 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002325777331995988, + "loss": 1.348, + "theoretical_loss": 3.4926174714103864, + "tokens_seen": 1780744192 + }, + { + "epoch": 0.54, + "learning_rate": 0.000232567703109328, + "loss": 1.3416, + "theoretical_loss": 3.4926065975821396, + "tokens_seen": 1780809728 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023255767301905717, + "loss": 1.3995, + "theoretical_loss": 3.4925957242660988, + "tokens_seen": 1780875264 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023254764292878635, + "loss": 1.3637, + "theoretical_loss": 3.4925848514622215, + "tokens_seen": 1780940800 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023253761283851556, + "loss": 1.3976, + "theoretical_loss": 3.4925739791704635, + "tokens_seen": 1781006336 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023252758274824474, + "loss": 1.3466, + "theoretical_loss": 3.492563107390783, + "tokens_seen": 1781071872 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023251755265797392, + "loss": 1.2314, + "theoretical_loss": 3.4925522361231365, + "tokens_seen": 1781137408 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002325075225677031, + "loss": 1.4788, + "theoretical_loss": 3.4925413653674813, + "tokens_seen": 1781202944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023249749247743228, + "loss": 1.3514, + "theoretical_loss": 3.4925304951237734, + "tokens_seen": 1781268480 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002324874623871615, + "loss": 1.1714, + "theoretical_loss": 3.4925196253919717, + "tokens_seen": 1781334016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023247743229689067, + "loss": 1.2771, + "theoretical_loss": 3.4925087561720325, + "tokens_seen": 1781399552 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023246740220661985, + "loss": 1.4406, + "theoretical_loss": 3.4924978874639114, + "tokens_seen": 1781465088 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023245737211634903, + "loss": 1.3095, + "theoretical_loss": 3.4924870192675677, + "tokens_seen": 1781530624 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023244734202607824, + "loss": 1.4312, + "theoretical_loss": 3.4924761515829577, + "tokens_seen": 1781596160 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023243731193580742, + "loss": 1.3268, + "theoretical_loss": 3.492465284410038, + "tokens_seen": 1781661696 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002324272818455366, + "loss": 1.421, + "theoretical_loss": 3.4924544177487657, + "tokens_seen": 1781727232 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023241725175526578, + "loss": 1.2599, + "theoretical_loss": 3.4924435515990986, + "tokens_seen": 1781792768 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023240722166499497, + "loss": 1.3606, + "theoretical_loss": 3.492432685960993, + "tokens_seen": 1781858304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023239719157472417, + "loss": 1.3901, + "theoretical_loss": 3.492421820834406, + "tokens_seen": 1781923840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023238716148445336, + "loss": 1.3618, + "theoretical_loss": 3.492410956219296, + "tokens_seen": 1781989376 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023237713139418254, + "loss": 1.4357, + "theoretical_loss": 3.492400092115619, + "tokens_seen": 1782054912 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023236710130391172, + "loss": 1.4113, + "theoretical_loss": 3.492389228523332, + "tokens_seen": 1782120448 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023235707121364095, + "loss": 1.2792, + "theoretical_loss": 3.4923783654423923, + "tokens_seen": 1782185984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023234704112337013, + "loss": 1.2559, + "theoretical_loss": 3.4923675028727574, + "tokens_seen": 1782251520 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023233701103309932, + "loss": 1.2623, + "theoretical_loss": 3.4923566408143842, + "tokens_seen": 1782317056 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002323269809428285, + "loss": 1.3406, + "theoretical_loss": 3.49234577926723, + "tokens_seen": 1782382592 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023231695085255768, + "loss": 1.3739, + "theoretical_loss": 3.4923349182312515, + "tokens_seen": 1782448128 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023230692076228689, + "loss": 1.4162, + "theoretical_loss": 3.4923240577064067, + "tokens_seen": 1782513664 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 650724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5339055061340332, + "objective/train/theoretical_loss": 3.4923131976926514, + "objective/train/tokens_used": 1803039200, + "theoretical_loss": 3.4923131976926514, + "tokens_seen": 1782579200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023229689067201607, + "loss": 1.4408, + "theoretical_loss": 3.4923131976926514, + "tokens_seen": 1782579200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023228686058174525, + "loss": 1.3368, + "theoretical_loss": 3.492302338189944, + "tokens_seen": 1782644736 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023227683049147443, + "loss": 1.3895, + "theoretical_loss": 3.492291479198241, + "tokens_seen": 1782710272 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023226680040120364, + "loss": 1.3658, + "theoretical_loss": 3.4922806207174997, + "tokens_seen": 1782775808 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023225677031093282, + "loss": 1.3623, + "theoretical_loss": 3.4922697627476778, + "tokens_seen": 1782841344 + }, + { + "epoch": 0.54, + "learning_rate": 0.000232246740220662, + "loss": 1.4042, + "theoretical_loss": 3.4922589052887316, + "tokens_seen": 1782906880 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023223671013039118, + "loss": 1.4116, + "theoretical_loss": 3.492248048340619, + "tokens_seen": 1782972416 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023222668004012036, + "loss": 1.4626, + "theoretical_loss": 3.492237191903297, + "tokens_seen": 1783037952 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023221664994984957, + "loss": 1.3696, + "theoretical_loss": 3.4922263359767225, + "tokens_seen": 1783103488 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023220661985957875, + "loss": 1.4294, + "theoretical_loss": 3.492215480560853, + "tokens_seen": 1783169024 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023219658976930793, + "loss": 1.3978, + "theoretical_loss": 3.4922046256556456, + "tokens_seen": 1783234560 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002321865596790371, + "loss": 1.3091, + "theoretical_loss": 3.4921937712610576, + "tokens_seen": 1783300096 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023217652958876632, + "loss": 1.5531, + "theoretical_loss": 3.492182917377046, + "tokens_seen": 1783365632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002321664994984955, + "loss": 1.3689, + "theoretical_loss": 3.492172064003568, + "tokens_seen": 1783431168 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023215646940822468, + "loss": 1.3642, + "theoretical_loss": 3.492161211140581, + "tokens_seen": 1783496704 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023214643931795386, + "loss": 1.3766, + "theoretical_loss": 3.4921503587880425, + "tokens_seen": 1783562240 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023213640922768307, + "loss": 1.4155, + "theoretical_loss": 3.4921395069459096, + "tokens_seen": 1783627776 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023212637913741225, + "loss": 1.3349, + "theoretical_loss": 3.492128655614139, + "tokens_seen": 1783693312 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023211634904714144, + "loss": 1.3407, + "theoretical_loss": 3.492117804792689, + "tokens_seen": 1783758848 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023210631895687062, + "loss": 1.3652, + "theoretical_loss": 3.4921069544815153, + "tokens_seen": 1783824384 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002320962888665998, + "loss": 1.4016, + "theoretical_loss": 3.4920961046805763, + "tokens_seen": 1783889920 + }, + { + "epoch": 0.54, + "learning_rate": 0.000232086258776329, + "loss": 1.324, + "theoretical_loss": 3.4920852553898296, + "tokens_seen": 1783955456 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002320762286860582, + "loss": 1.3618, + "theoretical_loss": 3.4920744066092313, + "tokens_seen": 1784020992 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023206619859578737, + "loss": 1.3917, + "theoretical_loss": 3.4920635583387396, + "tokens_seen": 1784086528 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023205616850551655, + "loss": 1.2735, + "theoretical_loss": 3.492052710578311, + "tokens_seen": 1784152064 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023204613841524576, + "loss": 1.3144, + "theoretical_loss": 3.492041863327904, + "tokens_seen": 1784217600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023203610832497494, + "loss": 1.3329, + "theoretical_loss": 3.4920310165874744, + "tokens_seen": 1784283136 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023202607823470412, + "loss": 1.3947, + "theoretical_loss": 3.49202017035698, + "tokens_seen": 1784348672 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002320160481444333, + "loss": 1.3361, + "theoretical_loss": 3.492009324636379, + "tokens_seen": 1784414208 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023200601805416248, + "loss": 1.4059, + "theoretical_loss": 3.491998479425628, + "tokens_seen": 1784479744 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002319959879638917, + "loss": 1.2462, + "theoretical_loss": 3.491987634724684, + "tokens_seen": 1784545280 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023198595787362087, + "loss": 1.3309, + "theoretical_loss": 3.4919767905335046, + "tokens_seen": 1784610816 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023197592778335005, + "loss": 1.3646, + "theoretical_loss": 3.491965946852047, + "tokens_seen": 1784676352 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023196589769307923, + "loss": 1.3483, + "theoretical_loss": 3.491955103680269, + "tokens_seen": 1784741888 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023195586760280844, + "loss": 1.5202, + "theoretical_loss": 3.4919442610181277, + "tokens_seen": 1784807424 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023194583751253762, + "loss": 1.3554, + "theoretical_loss": 3.49193341886558, + "tokens_seen": 1784872960 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002319358074222668, + "loss": 1.4957, + "theoretical_loss": 3.491922577222584, + "tokens_seen": 1784938496 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023192577733199599, + "loss": 1.3898, + "theoretical_loss": 3.4919117360890963, + "tokens_seen": 1785004032 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023191574724172517, + "loss": 1.4461, + "theoretical_loss": 3.4919008954650748, + "tokens_seen": 1785069568 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023190571715145437, + "loss": 1.4746, + "theoretical_loss": 3.491890055350477, + "tokens_seen": 1785135104 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023189568706118356, + "loss": 1.3401, + "theoretical_loss": 3.491879215745259, + "tokens_seen": 1785200640 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023188565697091274, + "loss": 1.3043, + "theoretical_loss": 3.4918683766493794, + "tokens_seen": 1785266176 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023187562688064192, + "loss": 1.3754, + "theoretical_loss": 3.491857538062795, + "tokens_seen": 1785331712 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023186559679037113, + "loss": 1.3074, + "theoretical_loss": 3.491846699985464, + "tokens_seen": 1785397248 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002318555667001003, + "loss": 1.3742, + "theoretical_loss": 3.491835862417343, + "tokens_seen": 1785462784 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002318455366098295, + "loss": 1.3724, + "theoretical_loss": 3.491825025358389, + "tokens_seen": 1785528320 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023183550651955867, + "loss": 1.2362, + "theoretical_loss": 3.4918141888085605, + "tokens_seen": 1785593856 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023182547642928785, + "loss": 1.3449, + "theoretical_loss": 3.4918033527678145, + "tokens_seen": 1785659392 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023181544633901706, + "loss": 1.3183, + "theoretical_loss": 3.491792517236108, + "tokens_seen": 1785724928 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023180541624874624, + "loss": 1.2765, + "theoretical_loss": 3.4917816822133982, + "tokens_seen": 1785790464 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 652000, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.554517149925232, + "objective/train/theoretical_loss": 3.4917708476996436, + "objective/train/tokens_used": 1806316000, + "theoretical_loss": 3.4917708476996436, + "tokens_seen": 1785856000 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023179538615847542, + "loss": 1.4767, + "theoretical_loss": 3.4917708476996436, + "tokens_seen": 1785856000 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002317853560682046, + "loss": 1.3585, + "theoretical_loss": 3.491760013694801, + "tokens_seen": 1785921536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002317753259779338, + "loss": 1.4397, + "theoretical_loss": 3.491749180198828, + "tokens_seen": 1785987072 + }, + { + "epoch": 0.54, + "learning_rate": 0.000231765295887663, + "loss": 1.3091, + "theoretical_loss": 3.4917383472116814, + "tokens_seen": 1786052608 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023175526579739217, + "loss": 1.3417, + "theoretical_loss": 3.491727514733319, + "tokens_seen": 1786118144 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023174523570712135, + "loss": 1.3933, + "theoretical_loss": 3.4917166827636987, + "tokens_seen": 1786183680 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023173520561685053, + "loss": 1.3027, + "theoretical_loss": 3.491705851302777, + "tokens_seen": 1786249216 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023172517552657974, + "loss": 1.409, + "theoretical_loss": 3.4916950203505124, + "tokens_seen": 1786314752 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023171514543630892, + "loss": 1.2621, + "theoretical_loss": 3.491684189906862, + "tokens_seen": 1786380288 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002317051153460381, + "loss": 1.417, + "theoretical_loss": 3.4916733599717826, + "tokens_seen": 1786445824 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023169508525576729, + "loss": 1.3602, + "theoretical_loss": 3.4916625305452325, + "tokens_seen": 1786511360 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002316850551654965, + "loss": 1.3917, + "theoretical_loss": 3.491651701627169, + "tokens_seen": 1786576896 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023167502507522568, + "loss": 1.2961, + "theoretical_loss": 3.4916408732175492, + "tokens_seen": 1786642432 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023166499498495486, + "loss": 1.2333, + "theoretical_loss": 3.491630045316331, + "tokens_seen": 1786707968 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023165496489468404, + "loss": 1.4877, + "theoretical_loss": 3.4916192179234713, + "tokens_seen": 1786773504 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023164493480441322, + "loss": 1.5061, + "theoretical_loss": 3.491608391038928, + "tokens_seen": 1786839040 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023163490471414243, + "loss": 1.4308, + "theoretical_loss": 3.4915975646626594, + "tokens_seen": 1786904576 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002316248746238716, + "loss": 1.3435, + "theoretical_loss": 3.491586738794621, + "tokens_seen": 1786970112 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002316148445336008, + "loss": 1.3705, + "theoretical_loss": 3.491575913434772, + "tokens_seen": 1787035648 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023160481444333, + "loss": 1.3104, + "theoretical_loss": 3.4915650885830694, + "tokens_seen": 1787101184 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002315947843530592, + "loss": 1.4, + "theoretical_loss": 3.4915542642394706, + "tokens_seen": 1787166720 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002315847542627884, + "loss": 1.3452, + "theoretical_loss": 3.4915434404039334, + "tokens_seen": 1787232256 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023157472417251757, + "loss": 1.3165, + "theoretical_loss": 3.491532617076415, + "tokens_seen": 1787297792 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023156469408224675, + "loss": 1.3837, + "theoretical_loss": 3.491521794256873, + "tokens_seen": 1787363328 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023155466399197596, + "loss": 1.3774, + "theoretical_loss": 3.4915109719452646, + "tokens_seen": 1787428864 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023154463390170514, + "loss": 1.1624, + "theoretical_loss": 3.4915001501415484, + "tokens_seen": 1787494400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023153460381143432, + "loss": 1.3292, + "theoretical_loss": 3.491489328845681, + "tokens_seen": 1787559936 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002315245737211635, + "loss": 1.5193, + "theoretical_loss": 3.4914785080576207, + "tokens_seen": 1787625472 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023151454363089268, + "loss": 1.3836, + "theoretical_loss": 3.491467687777324, + "tokens_seen": 1787691008 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002315045135406219, + "loss": 1.3611, + "theoretical_loss": 3.4914568680047493, + "tokens_seen": 1787756544 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023149448345035107, + "loss": 1.4233, + "theoretical_loss": 3.4914460487398538, + "tokens_seen": 1787822080 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023148445336008025, + "loss": 1.3648, + "theoretical_loss": 3.491435229982595, + "tokens_seen": 1787887616 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023147442326980943, + "loss": 1.2551, + "theoretical_loss": 3.4914244117329307, + "tokens_seen": 1787953152 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023146439317953864, + "loss": 1.3909, + "theoretical_loss": 3.491413593990819, + "tokens_seen": 1788018688 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023145436308926782, + "loss": 1.2743, + "theoretical_loss": 3.4914027767562166, + "tokens_seen": 1788084224 + }, + { + "epoch": 0.54, + "learning_rate": 0.000231444332998997, + "loss": 1.3389, + "theoretical_loss": 3.4913919600290813, + "tokens_seen": 1788149760 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023143430290872619, + "loss": 1.3597, + "theoretical_loss": 3.491381143809371, + "tokens_seen": 1788215296 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023142427281845537, + "loss": 1.3901, + "theoretical_loss": 3.4913703280970427, + "tokens_seen": 1788280832 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023141424272818457, + "loss": 1.3496, + "theoretical_loss": 3.4913595128920547, + "tokens_seen": 1788346368 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023140421263791376, + "loss": 1.2994, + "theoretical_loss": 3.4913486981943644, + "tokens_seen": 1788411904 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023139418254764294, + "loss": 1.3962, + "theoretical_loss": 3.4913378840039293, + "tokens_seen": 1788477440 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023138415245737212, + "loss": 1.3111, + "theoretical_loss": 3.491327070320707, + "tokens_seen": 1788542976 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023137412236710133, + "loss": 1.2821, + "theoretical_loss": 3.491316257144655, + "tokens_seen": 1788608512 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002313640922768305, + "loss": 1.3176, + "theoretical_loss": 3.4913054444757314, + "tokens_seen": 1788674048 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002313540621865597, + "loss": 1.3633, + "theoretical_loss": 3.4912946323138936, + "tokens_seen": 1788739584 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023134403209628887, + "loss": 1.3627, + "theoretical_loss": 3.491283820659099, + "tokens_seen": 1788805120 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023133400200601805, + "loss": 1.3081, + "theoretical_loss": 3.491273009511306, + "tokens_seen": 1788870656 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023132397191574726, + "loss": 1.3412, + "theoretical_loss": 3.491262198870471, + "tokens_seen": 1788936192 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023131394182547644, + "loss": 1.4068, + "theoretical_loss": 3.491251388736553, + "tokens_seen": 1789001728 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023130391173520562, + "loss": 1.3306, + "theoretical_loss": 3.4912405791095082, + "tokens_seen": 1789067264 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 653179, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4443249702453613, + "objective/train/theoretical_loss": 3.491229769989296, + "objective/train/tokens_used": 1809592800, + "theoretical_loss": 3.491229769989296, + "tokens_seen": 1789132800 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002312938816449348, + "loss": 1.328, + "theoretical_loss": 3.491229769989296, + "tokens_seen": 1789132800 + }, + { + "epoch": 0.54, + "learning_rate": 0.000231283851554664, + "loss": 1.3476, + "theoretical_loss": 3.4912189613758726, + "tokens_seen": 1789198336 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002312738214643932, + "loss": 1.2982, + "theoretical_loss": 3.4912081532691963, + "tokens_seen": 1789263872 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023126379137412237, + "loss": 1.3761, + "theoretical_loss": 3.491197345669225, + "tokens_seen": 1789329408 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023125376128385155, + "loss": 1.3519, + "theoretical_loss": 3.4911865385759158, + "tokens_seen": 1789394944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023124373119358073, + "loss": 1.3759, + "theoretical_loss": 3.4911757319892267, + "tokens_seen": 1789460480 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023123370110330994, + "loss": 1.2586, + "theoretical_loss": 3.4911649259091155, + "tokens_seen": 1789526016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023122367101303912, + "loss": 1.4166, + "theoretical_loss": 3.4911541203355396, + "tokens_seen": 1789591552 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002312136409227683, + "loss": 1.3247, + "theoretical_loss": 3.4911433152684572, + "tokens_seen": 1789657088 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002312036108324975, + "loss": 1.4739, + "theoretical_loss": 3.4911325107078257, + "tokens_seen": 1789722624 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002311935807422267, + "loss": 1.3821, + "theoretical_loss": 3.491121706653603, + "tokens_seen": 1789788160 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023118355065195588, + "loss": 1.3748, + "theoretical_loss": 3.4911109031057457, + "tokens_seen": 1789853696 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023117352056168506, + "loss": 1.4808, + "theoretical_loss": 3.4911001000642132, + "tokens_seen": 1789919232 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023116349047141424, + "loss": 1.2958, + "theoretical_loss": 3.491089297528963, + "tokens_seen": 1789984768 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023115346038114342, + "loss": 1.4105, + "theoretical_loss": 3.4910784954999516, + "tokens_seen": 1790050304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023114343029087263, + "loss": 1.2576, + "theoretical_loss": 3.491067693977138, + "tokens_seen": 1790115840 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002311334002006018, + "loss": 1.3796, + "theoretical_loss": 3.4910568929604793, + "tokens_seen": 1790181376 + }, + { + "epoch": 0.54, + "learning_rate": 0.000231123370110331, + "loss": 1.4193, + "theoretical_loss": 3.491046092449933, + "tokens_seen": 1790246912 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023111334002006017, + "loss": 1.2139, + "theoretical_loss": 3.4910352924454577, + "tokens_seen": 1790312448 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023110330992978938, + "loss": 1.2426, + "theoretical_loss": 3.4910244929470107, + "tokens_seen": 1790377984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023109327983951856, + "loss": 1.4359, + "theoretical_loss": 3.4910136939545495, + "tokens_seen": 1790443520 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023108324974924774, + "loss": 1.3101, + "theoretical_loss": 3.4910028954680326, + "tokens_seen": 1790509056 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023107321965897692, + "loss": 1.3724, + "theoretical_loss": 3.490992097487417, + "tokens_seen": 1790574592 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002310631895687061, + "loss": 1.2771, + "theoretical_loss": 3.490981300012661, + "tokens_seen": 1790640128 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002310531594784353, + "loss": 1.1919, + "theoretical_loss": 3.490970503043722, + "tokens_seen": 1790705664 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002310431293881645, + "loss": 1.3525, + "theoretical_loss": 3.4909597065805587, + "tokens_seen": 1790771200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023103309929789367, + "loss": 1.4182, + "theoretical_loss": 3.4909489106231275, + "tokens_seen": 1790836736 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023102306920762286, + "loss": 1.3001, + "theoretical_loss": 3.490938115171387, + "tokens_seen": 1790902272 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023101303911735206, + "loss": 1.3109, + "theoretical_loss": 3.4909273202252953, + "tokens_seen": 1790967808 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023100300902708124, + "loss": 1.4444, + "theoretical_loss": 3.49091652578481, + "tokens_seen": 1791033344 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023099297893681043, + "loss": 1.34, + "theoretical_loss": 3.490905731849888, + "tokens_seen": 1791098880 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002309829488465396, + "loss": 1.3132, + "theoretical_loss": 3.4908949384204884, + "tokens_seen": 1791164416 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023097291875626881, + "loss": 1.2996, + "theoretical_loss": 3.490884145496568, + "tokens_seen": 1791229952 + }, + { + "epoch": 0.54, + "learning_rate": 0.000230962888665998, + "loss": 1.4405, + "theoretical_loss": 3.4908733530780856, + "tokens_seen": 1791295488 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023095285857572718, + "loss": 1.3228, + "theoretical_loss": 3.490862561164999, + "tokens_seen": 1791361024 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023094282848545636, + "loss": 1.4269, + "theoretical_loss": 3.490851769757265, + "tokens_seen": 1791426560 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023093279839518554, + "loss": 1.3183, + "theoretical_loss": 3.490840978854842, + "tokens_seen": 1791492096 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023092276830491475, + "loss": 1.3436, + "theoretical_loss": 3.4908301884576884, + "tokens_seen": 1791557632 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023091273821464393, + "loss": 1.3249, + "theoretical_loss": 3.4908193985657614, + "tokens_seen": 1791623168 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002309027081243731, + "loss": 1.1548, + "theoretical_loss": 3.490808609179019, + "tokens_seen": 1791688704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002308926780341023, + "loss": 1.3581, + "theoretical_loss": 3.490797820297419, + "tokens_seen": 1791754240 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002308826479438315, + "loss": 1.3168, + "theoretical_loss": 3.4907870319209198, + "tokens_seen": 1791819776 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023087261785356068, + "loss": 1.3393, + "theoretical_loss": 3.4907762440494787, + "tokens_seen": 1791885312 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023086258776328986, + "loss": 1.4062, + "theoretical_loss": 3.4907654566830537, + "tokens_seen": 1791950848 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023085255767301907, + "loss": 1.2785, + "theoretical_loss": 3.4907546698216025, + "tokens_seen": 1792016384 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023084252758274825, + "loss": 1.3122, + "theoretical_loss": 3.4907438834650844, + "tokens_seen": 1792081920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023083249749247746, + "loss": 1.3603, + "theoretical_loss": 3.4907330976134547, + "tokens_seen": 1792147456 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023082246740220664, + "loss": 1.2838, + "theoretical_loss": 3.4907223122666737, + "tokens_seen": 1792212992 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023081243731193582, + "loss": 1.3697, + "theoretical_loss": 3.490711527424698, + "tokens_seen": 1792278528 + }, + { + "epoch": 0.54, + "learning_rate": 0.000230802407221665, + "loss": 1.3155, + "theoretical_loss": 3.490700743087486, + "tokens_seen": 1792344064 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 654198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1906723976135254, + "objective/train/theoretical_loss": 3.4906899592549956, + "objective/train/tokens_used": 1812869600, + "theoretical_loss": 3.4906899592549956, + "tokens_seen": 1792409600 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002307923771313942, + "loss": 1.2099, + "theoretical_loss": 3.4906899592549956, + "tokens_seen": 1792409600 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002307823470411234, + "loss": 1.3108, + "theoretical_loss": 3.4906791759271845, + "tokens_seen": 1792475136 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023077231695085257, + "loss": 1.2985, + "theoretical_loss": 3.490668393104011, + "tokens_seen": 1792540672 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023076228686058175, + "loss": 1.2756, + "theoretical_loss": 3.4906576107854326, + "tokens_seen": 1792606208 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023075225677031094, + "loss": 1.4055, + "theoretical_loss": 3.490646828971407, + "tokens_seen": 1792671744 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023074222668004014, + "loss": 1.4079, + "theoretical_loss": 3.4906360476618934, + "tokens_seen": 1792737280 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023073219658976932, + "loss": 1.3001, + "theoretical_loss": 3.4906252668568483, + "tokens_seen": 1792802816 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002307221664994985, + "loss": 1.2506, + "theoretical_loss": 3.4906144865562307, + "tokens_seen": 1792868352 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002307121364092277, + "loss": 1.3926, + "theoretical_loss": 3.490603706759998, + "tokens_seen": 1792933888 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002307021063189569, + "loss": 1.3575, + "theoretical_loss": 3.490592927468108, + "tokens_seen": 1792999424 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023069207622868608, + "loss": 1.3282, + "theoretical_loss": 3.4905821486805197, + "tokens_seen": 1793064960 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023068204613841526, + "loss": 1.2948, + "theoretical_loss": 3.49057137039719, + "tokens_seen": 1793130496 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023067201604814444, + "loss": 1.3049, + "theoretical_loss": 3.490560592618077, + "tokens_seen": 1793196032 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023066198595787362, + "loss": 1.1975, + "theoretical_loss": 3.4905498153431394, + "tokens_seen": 1793261568 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023065195586760283, + "loss": 1.3296, + "theoretical_loss": 3.4905390385723347, + "tokens_seen": 1793327104 + }, + { + "epoch": 0.54, + "learning_rate": 0.000230641925777332, + "loss": 1.3392, + "theoretical_loss": 3.49052826230562, + "tokens_seen": 1793392640 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002306318956870612, + "loss": 1.4216, + "theoretical_loss": 3.490517486542956, + "tokens_seen": 1793458176 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023062186559679037, + "loss": 1.2873, + "theoretical_loss": 3.490506711284297, + "tokens_seen": 1793523712 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023061183550651958, + "loss": 1.1206, + "theoretical_loss": 3.490495936529604, + "tokens_seen": 1793589248 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023060180541624876, + "loss": 1.2356, + "theoretical_loss": 3.490485162278834, + "tokens_seen": 1793654784 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023059177532597794, + "loss": 1.3145, + "theoretical_loss": 3.490474388531945, + "tokens_seen": 1793720320 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023058174523570712, + "loss": 1.3969, + "theoretical_loss": 3.4904636152888946, + "tokens_seen": 1793785856 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002305717151454363, + "loss": 1.246, + "theoretical_loss": 3.4904528425496415, + "tokens_seen": 1793851392 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002305616850551655, + "loss": 1.392, + "theoretical_loss": 3.490442070314143, + "tokens_seen": 1793916928 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002305516549648947, + "loss": 1.4377, + "theoretical_loss": 3.490431298582358, + "tokens_seen": 1793982464 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023054162487462387, + "loss": 1.2995, + "theoretical_loss": 3.490420527354244, + "tokens_seen": 1794048000 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023053159478435306, + "loss": 1.3129, + "theoretical_loss": 3.49040975662976, + "tokens_seen": 1794113536 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023052156469408226, + "loss": 1.2999, + "theoretical_loss": 3.4903989864088625, + "tokens_seen": 1794179072 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023051153460381144, + "loss": 1.3683, + "theoretical_loss": 3.4903882166915103, + "tokens_seen": 1794244608 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023050150451354063, + "loss": 1.4475, + "theoretical_loss": 3.4903774474776617, + "tokens_seen": 1794310144 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002304914744232698, + "loss": 1.3291, + "theoretical_loss": 3.4903666787672747, + "tokens_seen": 1794375680 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023048144433299902, + "loss": 1.2772, + "theoretical_loss": 3.4903559105603073, + "tokens_seen": 1794441216 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002304714142427282, + "loss": 1.3246, + "theoretical_loss": 3.490345142856717, + "tokens_seen": 1794506752 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023046138415245738, + "loss": 1.473, + "theoretical_loss": 3.4903343756564627, + "tokens_seen": 1794572288 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023045135406218656, + "loss": 1.368, + "theoretical_loss": 3.4903236089595024, + "tokens_seen": 1794637824 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023044132397191574, + "loss": 1.2506, + "theoretical_loss": 3.4903128427657935, + "tokens_seen": 1794703360 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023043129388164495, + "loss": 1.3893, + "theoretical_loss": 3.4903020770752953, + "tokens_seen": 1794768896 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023042126379137413, + "loss": 1.3595, + "theoretical_loss": 3.4902913118879644, + "tokens_seen": 1794834432 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002304112337011033, + "loss": 1.2929, + "theoretical_loss": 3.4902805472037604, + "tokens_seen": 1794899968 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002304012036108325, + "loss": 1.3434, + "theoretical_loss": 3.4902697830226406, + "tokens_seen": 1794965504 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002303911735205617, + "loss": 1.3722, + "theoretical_loss": 3.490259019344563, + "tokens_seen": 1795031040 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023038114343029088, + "loss": 1.3461, + "theoretical_loss": 3.490248256169486, + "tokens_seen": 1795096576 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023037111334002006, + "loss": 1.3264, + "theoretical_loss": 3.490237493497368, + "tokens_seen": 1795162112 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023036108324974924, + "loss": 1.3803, + "theoretical_loss": 3.4902267313281667, + "tokens_seen": 1795227648 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023035105315947842, + "loss": 1.419, + "theoretical_loss": 3.4902159696618402, + "tokens_seen": 1795293184 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023034102306920763, + "loss": 1.3819, + "theoretical_loss": 3.490205208498347, + "tokens_seen": 1795358720 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002303309929789368, + "loss": 1.3192, + "theoretical_loss": 3.490194447837645, + "tokens_seen": 1795424256 + }, + { + "epoch": 0.54, + "learning_rate": 0.000230320962888666, + "loss": 1.4146, + "theoretical_loss": 3.490183687679693, + "tokens_seen": 1795489792 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023031093279839518, + "loss": 1.3229, + "theoretical_loss": 3.490172928024448, + "tokens_seen": 1795555328 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023030090270812438, + "loss": 1.4398, + "theoretical_loss": 3.490162168871869, + "tokens_seen": 1795620864 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 655419, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3706367015838623, + "objective/train/theoretical_loss": 3.4901514102219138, + "objective/train/tokens_used": 1816146400, + "theoretical_loss": 3.4901514102219138, + "tokens_seen": 1795686400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023029087261785356, + "loss": 1.3153, + "theoretical_loss": 3.4901514102219138, + "tokens_seen": 1795686400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023028084252758275, + "loss": 1.3277, + "theoretical_loss": 3.490140652074541, + "tokens_seen": 1795751936 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023027081243731193, + "loss": 1.3255, + "theoretical_loss": 3.4901298944297086, + "tokens_seen": 1795817472 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002302607823470411, + "loss": 1.2931, + "theoretical_loss": 3.4901191372873743, + "tokens_seen": 1795883008 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023025075225677032, + "loss": 1.4418, + "theoretical_loss": 3.490108380647497, + "tokens_seen": 1795948544 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002302407221664995, + "loss": 1.4384, + "theoretical_loss": 3.4900976245100344, + "tokens_seen": 1796014080 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023023069207622868, + "loss": 1.329, + "theoretical_loss": 3.4900868688749442, + "tokens_seen": 1796079616 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023022066198595786, + "loss": 1.2998, + "theoretical_loss": 3.4900761137421865, + "tokens_seen": 1796145152 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023021063189568707, + "loss": 1.3843, + "theoretical_loss": 3.4900653591117177, + "tokens_seen": 1796210688 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023020060180541625, + "loss": 1.3966, + "theoretical_loss": 3.490054604983497, + "tokens_seen": 1796276224 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023019057171514543, + "loss": 1.2237, + "theoretical_loss": 3.4900438513574814, + "tokens_seen": 1796341760 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002301805416248746, + "loss": 1.4973, + "theoretical_loss": 3.490033098233631, + "tokens_seen": 1796407296 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002301705115346038, + "loss": 1.4232, + "theoretical_loss": 3.4900223456119024, + "tokens_seen": 1796472832 + }, + { + "epoch": 0.54, + "learning_rate": 0.000230160481444333, + "loss": 1.3469, + "theoretical_loss": 3.4900115934922544, + "tokens_seen": 1796538368 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023015045135406218, + "loss": 1.5425, + "theoretical_loss": 3.4900008418746458, + "tokens_seen": 1796603904 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023014042126379136, + "loss": 1.3434, + "theoretical_loss": 3.4899900907590338, + "tokens_seen": 1796669440 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023013039117352054, + "loss": 1.3153, + "theoretical_loss": 3.4899793401453776, + "tokens_seen": 1796734976 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023012036108324975, + "loss": 1.422, + "theoretical_loss": 3.4899685900336346, + "tokens_seen": 1796800512 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023011033099297893, + "loss": 1.3632, + "theoretical_loss": 3.489957840423764, + "tokens_seen": 1796866048 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023010030090270814, + "loss": 1.248, + "theoretical_loss": 3.489947091315723, + "tokens_seen": 1796931584 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023009027081243732, + "loss": 1.3916, + "theoretical_loss": 3.4899363427094707, + "tokens_seen": 1796997120 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002300802407221665, + "loss": 1.2185, + "theoretical_loss": 3.489925594604965, + "tokens_seen": 1797062656 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002300702106318957, + "loss": 1.4294, + "theoretical_loss": 3.4899148470021646, + "tokens_seen": 1797128192 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002300601805416249, + "loss": 1.2896, + "theoretical_loss": 3.489904099901027, + "tokens_seen": 1797193728 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023005015045135407, + "loss": 1.2824, + "theoretical_loss": 3.4898933533015115, + "tokens_seen": 1797259264 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023004012036108326, + "loss": 1.2801, + "theoretical_loss": 3.489882607203576, + "tokens_seen": 1797324800 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023003009027081246, + "loss": 1.3198, + "theoretical_loss": 3.489871861607178, + "tokens_seen": 1797390336 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023002006018054164, + "loss": 1.2896, + "theoretical_loss": 3.4898611165122766, + "tokens_seen": 1797455872 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023001003009027083, + "loss": 1.4323, + "theoretical_loss": 3.4898503719188305, + "tokens_seen": 1797521408 + }, + { + "epoch": 0.54, + "learning_rate": 0.00023, + "loss": 1.4592, + "theoretical_loss": 3.489839627826797, + "tokens_seen": 1797586944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022998996990972922, + "loss": 1.3458, + "theoretical_loss": 3.489828884236135, + "tokens_seen": 1797652480 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002299799398194584, + "loss": 1.3317, + "theoretical_loss": 3.489818141146803, + "tokens_seen": 1797718016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022996990972918758, + "loss": 1.3536, + "theoretical_loss": 3.489807398558759, + "tokens_seen": 1797783552 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022995987963891676, + "loss": 1.2247, + "theoretical_loss": 3.4897966564719614, + "tokens_seen": 1797849088 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022994984954864594, + "loss": 1.3981, + "theoretical_loss": 3.489785914886369, + "tokens_seen": 1797914624 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022993981945837515, + "loss": 1.3241, + "theoretical_loss": 3.489775173801939, + "tokens_seen": 1797980160 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022992978936810433, + "loss": 1.2799, + "theoretical_loss": 3.489764433218631, + "tokens_seen": 1798045696 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002299197592778335, + "loss": 1.3321, + "theoretical_loss": 3.489753693136402, + "tokens_seen": 1798111232 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002299097291875627, + "loss": 1.2717, + "theoretical_loss": 3.4897429535552122, + "tokens_seen": 1798176768 + }, + { + "epoch": 0.54, + "learning_rate": 0.0002298996990972919, + "loss": 1.3853, + "theoretical_loss": 3.4897322144750182, + "tokens_seen": 1798242304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022988966900702108, + "loss": 1.3496, + "theoretical_loss": 3.4897214758957795, + "tokens_seen": 1798307840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022987963891675026, + "loss": 1.4087, + "theoretical_loss": 3.489710737817454, + "tokens_seen": 1798373376 + }, + { + "epoch": 0.54, + "learning_rate": 0.00022986960882647944, + "loss": 1.3463, + "theoretical_loss": 3.48970000024, + "tokens_seen": 1798438912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022985957873620862, + "loss": 1.4117, + "theoretical_loss": 3.4896892631633762, + "tokens_seen": 1798504448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022984954864593783, + "loss": 1.3406, + "theoretical_loss": 3.489678526587541, + "tokens_seen": 1798569984 + }, + { + "epoch": 0.55, + "learning_rate": 0.000229839518555667, + "loss": 1.2668, + "theoretical_loss": 3.4896677905124527, + "tokens_seen": 1798635520 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002298294884653962, + "loss": 1.4023, + "theoretical_loss": 3.4896570549380694, + "tokens_seen": 1798701056 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022981945837512538, + "loss": 1.3566, + "theoretical_loss": 3.4896463198643497, + "tokens_seen": 1798766592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022980942828485458, + "loss": 1.513, + "theoretical_loss": 3.4896355852912526, + "tokens_seen": 1798832128 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022979939819458376, + "loss": 1.381, + "theoretical_loss": 3.4896248512187356, + "tokens_seen": 1798897664 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 656735, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.337079405784607, + "objective/train/theoretical_loss": 3.4896141176467577, + "objective/train/tokens_used": 1819423200, + "theoretical_loss": 3.4896141176467577, + "tokens_seen": 1798963200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022978936810431295, + "loss": 1.2934, + "theoretical_loss": 3.4896141176467577, + "tokens_seen": 1798963200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022977933801404213, + "loss": 1.2853, + "theoretical_loss": 3.4896033845752763, + "tokens_seen": 1799028736 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002297693079237713, + "loss": 1.2718, + "theoretical_loss": 3.4895926520042515, + "tokens_seen": 1799094272 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022975927783350052, + "loss": 1.3962, + "theoretical_loss": 3.4895819199336406, + "tokens_seen": 1799159808 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002297492477432297, + "loss": 1.3513, + "theoretical_loss": 3.489571188363403, + "tokens_seen": 1799225344 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022973921765295888, + "loss": 1.2373, + "theoretical_loss": 3.4895604572934955, + "tokens_seen": 1799290880 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022972918756268806, + "loss": 1.5135, + "theoretical_loss": 3.489549726723878, + "tokens_seen": 1799356416 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022971915747241727, + "loss": 1.3506, + "theoretical_loss": 3.4895389966545087, + "tokens_seen": 1799421952 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022970912738214645, + "loss": 1.303, + "theoretical_loss": 3.489528267085346, + "tokens_seen": 1799487488 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022969909729187563, + "loss": 1.3863, + "theoretical_loss": 3.4895175380163472, + "tokens_seen": 1799553024 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002296890672016048, + "loss": 1.4626, + "theoretical_loss": 3.4895068094474726, + "tokens_seen": 1799618560 + }, + { + "epoch": 0.55, + "learning_rate": 0.000229679037111334, + "loss": 1.3684, + "theoretical_loss": 3.4894960813786797, + "tokens_seen": 1799684096 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002296690070210632, + "loss": 1.3512, + "theoretical_loss": 3.4894853538099273, + "tokens_seen": 1799749632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022965897693079238, + "loss": 1.3355, + "theoretical_loss": 3.489474626741174, + "tokens_seen": 1799815168 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022964894684052156, + "loss": 1.3411, + "theoretical_loss": 3.4894639001723773, + "tokens_seen": 1799880704 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022963891675025074, + "loss": 1.4835, + "theoretical_loss": 3.4894531741034966, + "tokens_seen": 1799946240 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022962888665997995, + "loss": 1.3271, + "theoretical_loss": 3.489442448534491, + "tokens_seen": 1800011776 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022961885656970913, + "loss": 1.3208, + "theoretical_loss": 3.489431723465317, + "tokens_seen": 1800077312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022960882647943831, + "loss": 1.2931, + "theoretical_loss": 3.489420998895935, + "tokens_seen": 1800142848 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002295987963891675, + "loss": 1.3789, + "theoretical_loss": 3.489410274826303, + "tokens_seen": 1800208384 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022958876629889668, + "loss": 1.3065, + "theoretical_loss": 3.489399551256379, + "tokens_seen": 1800273920 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022957873620862589, + "loss": 1.4849, + "theoretical_loss": 3.4893888281861223, + "tokens_seen": 1800339456 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022956870611835507, + "loss": 1.3748, + "theoretical_loss": 3.489378105615491, + "tokens_seen": 1800404992 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022955867602808425, + "loss": 1.2625, + "theoretical_loss": 3.489367383544443, + "tokens_seen": 1800470528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022954864593781343, + "loss": 1.3322, + "theoretical_loss": 3.489356661972938, + "tokens_seen": 1800536064 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022953861584754264, + "loss": 1.3418, + "theoretical_loss": 3.489345940900934, + "tokens_seen": 1800601600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022952858575727182, + "loss": 1.3606, + "theoretical_loss": 3.48933522032839, + "tokens_seen": 1800667136 + }, + { + "epoch": 0.55, + "learning_rate": 0.000229518555667001, + "loss": 1.4711, + "theoretical_loss": 3.4893245002552638, + "tokens_seen": 1800732672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022950852557673018, + "loss": 1.3784, + "theoretical_loss": 3.489313780681514, + "tokens_seen": 1800798208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022949849548645936, + "loss": 1.2185, + "theoretical_loss": 3.4893030616070995, + "tokens_seen": 1800863744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022948846539618857, + "loss": 1.3224, + "theoretical_loss": 3.4892923430319795, + "tokens_seen": 1800929280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022947843530591775, + "loss": 1.4611, + "theoretical_loss": 3.4892816249561114, + "tokens_seen": 1800994816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022946840521564693, + "loss": 1.1768, + "theoretical_loss": 3.4892709073794546, + "tokens_seen": 1801060352 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002294583751253761, + "loss": 1.3289, + "theoretical_loss": 3.489260190301967, + "tokens_seen": 1801125888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022944834503510532, + "loss": 1.3374, + "theoretical_loss": 3.489249473723608, + "tokens_seen": 1801191424 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002294383149448345, + "loss": 1.3362, + "theoretical_loss": 3.4892387576443356, + "tokens_seen": 1801256960 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022942828485456368, + "loss": 1.5143, + "theoretical_loss": 3.4892280420641084, + "tokens_seen": 1801322496 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022941825476429286, + "loss": 1.3425, + "theoretical_loss": 3.4892173269828852, + "tokens_seen": 1801388032 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022940822467402207, + "loss": 1.2801, + "theoretical_loss": 3.489206612400625, + "tokens_seen": 1801453568 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022939819458375125, + "loss": 1.3414, + "theoretical_loss": 3.489195898317286, + "tokens_seen": 1801519104 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022938816449348043, + "loss": 1.3052, + "theoretical_loss": 3.4891851847328264, + "tokens_seen": 1801584640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022937813440320962, + "loss": 1.1826, + "theoretical_loss": 3.489174471647205, + "tokens_seen": 1801650176 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002293681043129388, + "loss": 1.3588, + "theoretical_loss": 3.4891637590603812, + "tokens_seen": 1801715712 + }, + { + "epoch": 0.55, + "learning_rate": 0.000229358074222668, + "loss": 1.3338, + "theoretical_loss": 3.489153046972313, + "tokens_seen": 1801781248 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002293480441323972, + "loss": 1.3104, + "theoretical_loss": 3.4891423353829594, + "tokens_seen": 1801846784 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002293380140421264, + "loss": 1.2339, + "theoretical_loss": 3.4891316242922787, + "tokens_seen": 1801912320 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022932798395185558, + "loss": 1.3491, + "theoretical_loss": 3.489120913700229, + "tokens_seen": 1801977856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022931795386158478, + "loss": 1.2428, + "theoretical_loss": 3.4891102036067703, + "tokens_seen": 1802043392 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022930792377131397, + "loss": 1.2914, + "theoretical_loss": 3.4890994940118603, + "tokens_seen": 1802108928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022929789368104315, + "loss": 1.3995, + "theoretical_loss": 3.4890887849154577, + "tokens_seen": 1802174464 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 657910, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4427692890167236, + "objective/train/theoretical_loss": 3.4890780763175218, + "objective/train/tokens_used": 1822700000, + "theoretical_loss": 3.4890780763175218, + "tokens_seen": 1802240000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022928786359077233, + "loss": 1.4213, + "theoretical_loss": 3.4890780763175218, + "tokens_seen": 1802240000 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002292778335005015, + "loss": 1.346, + "theoretical_loss": 3.4890673682180107, + "tokens_seen": 1802305536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022926780341023072, + "loss": 1.4074, + "theoretical_loss": 3.489056660616883, + "tokens_seen": 1802371072 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002292577733199599, + "loss": 1.3015, + "theoretical_loss": 3.489045953514098, + "tokens_seen": 1802436608 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022924774322968908, + "loss": 1.4051, + "theoretical_loss": 3.489035246909614, + "tokens_seen": 1802502144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022923771313941826, + "loss": 1.3385, + "theoretical_loss": 3.4890245408033893, + "tokens_seen": 1802567680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022922768304914747, + "loss": 1.3851, + "theoretical_loss": 3.489013835195383, + "tokens_seen": 1802633216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022921765295887665, + "loss": 1.3139, + "theoretical_loss": 3.489003130085554, + "tokens_seen": 1802698752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022920762286860583, + "loss": 1.2512, + "theoretical_loss": 3.488992425473861, + "tokens_seen": 1802764288 + }, + { + "epoch": 0.55, + "learning_rate": 0.000229197592778335, + "loss": 1.4676, + "theoretical_loss": 3.4889817213602625, + "tokens_seen": 1802829824 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002291875626880642, + "loss": 1.4332, + "theoretical_loss": 3.488971017744717, + "tokens_seen": 1802895360 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002291775325977934, + "loss": 1.3318, + "theoretical_loss": 3.4889603146271835, + "tokens_seen": 1802960896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022916750250752258, + "loss": 1.4205, + "theoretical_loss": 3.488949612007621, + "tokens_seen": 1803026432 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022915747241725176, + "loss": 1.3773, + "theoretical_loss": 3.488938909885988, + "tokens_seen": 1803091968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022914744232698094, + "loss": 1.3012, + "theoretical_loss": 3.4889282082622426, + "tokens_seen": 1803157504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022913741223671015, + "loss": 1.3141, + "theoretical_loss": 3.4889175071363443, + "tokens_seen": 1803223040 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022912738214643933, + "loss": 1.3221, + "theoretical_loss": 3.4889068065082522, + "tokens_seen": 1803288576 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022911735205616851, + "loss": 1.2992, + "theoretical_loss": 3.488896106377924, + "tokens_seen": 1803354112 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002291073219658977, + "loss": 1.3144, + "theoretical_loss": 3.488885406745319, + "tokens_seen": 1803419648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022909729187562688, + "loss": 1.4999, + "theoretical_loss": 3.4888747076103956, + "tokens_seen": 1803485184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022908726178535609, + "loss": 1.2643, + "theoretical_loss": 3.4888640089731133, + "tokens_seen": 1803550720 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022907723169508527, + "loss": 1.2023, + "theoretical_loss": 3.4888533108334303, + "tokens_seen": 1803616256 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022906720160481445, + "loss": 1.2387, + "theoretical_loss": 3.4888426131913057, + "tokens_seen": 1803681792 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022905717151454363, + "loss": 1.3379, + "theoretical_loss": 3.4888319160466983, + "tokens_seen": 1803747328 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022904714142427284, + "loss": 1.3432, + "theoretical_loss": 3.488821219399566, + "tokens_seen": 1803812864 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022903711133400202, + "loss": 1.427, + "theoretical_loss": 3.488810523249869, + "tokens_seen": 1803878400 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002290270812437312, + "loss": 1.3568, + "theoretical_loss": 3.488799827597565, + "tokens_seen": 1803943936 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022901705115346038, + "loss": 1.3911, + "theoretical_loss": 3.4887891324426135, + "tokens_seen": 1804009472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022900702106318956, + "loss": 1.3625, + "theoretical_loss": 3.4887784377849727, + "tokens_seen": 1804075008 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022899699097291877, + "loss": 1.4517, + "theoretical_loss": 3.488767743624602, + "tokens_seen": 1804140544 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022898696088264795, + "loss": 1.3318, + "theoretical_loss": 3.4887570499614595, + "tokens_seen": 1804206080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022897693079237713, + "loss": 1.4763, + "theoretical_loss": 3.488746356795505, + "tokens_seen": 1804271616 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002289669007021063, + "loss": 1.2838, + "theoretical_loss": 3.488735664126697, + "tokens_seen": 1804337152 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022895687061183552, + "loss": 1.4162, + "theoretical_loss": 3.4887249719549933, + "tokens_seen": 1804402688 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002289468405215647, + "loss": 1.4417, + "theoretical_loss": 3.488714280280354, + "tokens_seen": 1804468224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022893681043129388, + "loss": 1.3497, + "theoretical_loss": 3.4887035891027374, + "tokens_seen": 1804533760 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022892678034102306, + "loss": 1.3601, + "theoretical_loss": 3.4886928984221024, + "tokens_seen": 1804599296 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022891675025075227, + "loss": 1.348, + "theoretical_loss": 3.4886822082384077, + "tokens_seen": 1804664832 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022890672016048145, + "loss": 1.3833, + "theoretical_loss": 3.4886715185516124, + "tokens_seen": 1804730368 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022889669007021063, + "loss": 1.3443, + "theoretical_loss": 3.4886608293616757, + "tokens_seen": 1804795904 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022888665997993982, + "loss": 1.223, + "theoretical_loss": 3.4886501406685557, + "tokens_seen": 1804861440 + }, + { + "epoch": 0.55, + "learning_rate": 0.000228876629889669, + "loss": 1.2697, + "theoretical_loss": 3.4886394524722117, + "tokens_seen": 1804926976 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002288665997993982, + "loss": 1.348, + "theoretical_loss": 3.488628764772603, + "tokens_seen": 1804992512 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002288565697091274, + "loss": 1.3465, + "theoretical_loss": 3.4886180775696873, + "tokens_seen": 1805058048 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022884653961885657, + "loss": 1.3265, + "theoretical_loss": 3.4886073908634243, + "tokens_seen": 1805123584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022883650952858575, + "loss": 1.4579, + "theoretical_loss": 3.488596704653773, + "tokens_seen": 1805189120 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022882647943831496, + "loss": 1.294, + "theoretical_loss": 3.488586018940692, + "tokens_seen": 1805254656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022881644934804414, + "loss": 1.3482, + "theoretical_loss": 3.4885753337241403, + "tokens_seen": 1805320192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022880641925777332, + "loss": 1.3748, + "theoretical_loss": 3.4885646490040765, + "tokens_seen": 1805385728 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002287963891675025, + "loss": 1.4026, + "theoretical_loss": 3.48855396478046, + "tokens_seen": 1805451264 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 658936, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2090117931365967, + "objective/train/theoretical_loss": 3.4885432810532495, + "objective/train/tokens_used": 1825976800, + "theoretical_loss": 3.4885432810532495, + "tokens_seen": 1805516800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022878635907723168, + "loss": 1.3772, + "theoretical_loss": 3.4885432810532495, + "tokens_seen": 1805516800 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002287763289869609, + "loss": 1.4157, + "theoretical_loss": 3.4885325978224038, + "tokens_seen": 1805582336 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022876629889669007, + "loss": 1.3244, + "theoretical_loss": 3.488521915087882, + "tokens_seen": 1805647872 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022875626880641925, + "loss": 1.4565, + "theoretical_loss": 3.4885112328496426, + "tokens_seen": 1805713408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022874623871614843, + "loss": 1.2732, + "theoretical_loss": 3.4885005511076455, + "tokens_seen": 1805778944 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022873620862587764, + "loss": 1.4421, + "theoretical_loss": 3.4884898698618487, + "tokens_seen": 1805844480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022872617853560682, + "loss": 1.3469, + "theoretical_loss": 3.4884791891122116, + "tokens_seen": 1805910016 + }, + { + "epoch": 0.55, + "learning_rate": 0.000228716148445336, + "loss": 1.3881, + "theoretical_loss": 3.488468508858693, + "tokens_seen": 1805975552 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022870611835506518, + "loss": 1.3272, + "theoretical_loss": 3.488457829101252, + "tokens_seen": 1806041088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022869608826479437, + "loss": 1.3032, + "theoretical_loss": 3.488447149839847, + "tokens_seen": 1806106624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022868605817452357, + "loss": 1.2702, + "theoretical_loss": 3.4884364710744378, + "tokens_seen": 1806172160 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022867602808425276, + "loss": 1.2389, + "theoretical_loss": 3.488425792804983, + "tokens_seen": 1806237696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022866599799398194, + "loss": 1.3086, + "theoretical_loss": 3.488415115031441, + "tokens_seen": 1806303232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022865596790371112, + "loss": 1.3286, + "theoretical_loss": 3.488404437753772, + "tokens_seen": 1806368768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022864593781344033, + "loss": 1.3268, + "theoretical_loss": 3.4883937609719338, + "tokens_seen": 1806434304 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002286359077231695, + "loss": 1.3469, + "theoretical_loss": 3.4883830846858856, + "tokens_seen": 1806499840 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002286258776328987, + "loss": 1.3425, + "theoretical_loss": 3.4883724088955876, + "tokens_seen": 1806565376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022861584754262787, + "loss": 1.2988, + "theoretical_loss": 3.488361733600997, + "tokens_seen": 1806630912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022860581745235705, + "loss": 1.4341, + "theoretical_loss": 3.488351058802074, + "tokens_seen": 1806696448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022859578736208629, + "loss": 1.2999, + "theoretical_loss": 3.4883403844987777, + "tokens_seen": 1806761984 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022858575727181547, + "loss": 1.3483, + "theoretical_loss": 3.4883297106910662, + "tokens_seen": 1806827520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022857572718154465, + "loss": 1.367, + "theoretical_loss": 3.4883190373788997, + "tokens_seen": 1806893056 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022856569709127383, + "loss": 1.4047, + "theoretical_loss": 3.4883083645622355, + "tokens_seen": 1806958592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022855566700100304, + "loss": 1.3262, + "theoretical_loss": 3.4882976922410345, + "tokens_seen": 1807024128 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022854563691073222, + "loss": 1.4319, + "theoretical_loss": 3.4882870204152545, + "tokens_seen": 1807089664 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002285356068204614, + "loss": 1.3784, + "theoretical_loss": 3.488276349084855, + "tokens_seen": 1807155200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022852557673019058, + "loss": 1.3998, + "theoretical_loss": 3.488265678249795, + "tokens_seen": 1807220736 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022851554663991976, + "loss": 1.3733, + "theoretical_loss": 3.4882550079100336, + "tokens_seen": 1807286272 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022850551654964897, + "loss": 1.3106, + "theoretical_loss": 3.4882443380655297, + "tokens_seen": 1807351808 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022849548645937815, + "loss": 1.3774, + "theoretical_loss": 3.4882336687162425, + "tokens_seen": 1807417344 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022848545636910733, + "loss": 1.3329, + "theoretical_loss": 3.4882229998621304, + "tokens_seen": 1807482880 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002284754262788365, + "loss": 1.2942, + "theoretical_loss": 3.4882123315031537, + "tokens_seen": 1807548416 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022846539618856572, + "loss": 1.2556, + "theoretical_loss": 3.4882016636392708, + "tokens_seen": 1807613952 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002284553660982949, + "loss": 1.3589, + "theoretical_loss": 3.4881909962704407, + "tokens_seen": 1807679488 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022844533600802408, + "loss": 1.4849, + "theoretical_loss": 3.4881803293966227, + "tokens_seen": 1807745024 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022843530591775326, + "loss": 1.3756, + "theoretical_loss": 3.488169663017775, + "tokens_seen": 1807810560 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022842527582748247, + "loss": 1.4335, + "theoretical_loss": 3.4881589971338585, + "tokens_seen": 1807876096 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022841524573721165, + "loss": 1.4047, + "theoretical_loss": 3.4881483317448305, + "tokens_seen": 1807941632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022840521564694084, + "loss": 1.3105, + "theoretical_loss": 3.488137666850651, + "tokens_seen": 1808007168 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022839518555667002, + "loss": 1.3503, + "theoretical_loss": 3.488127002451279, + "tokens_seen": 1808072704 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002283851554663992, + "loss": 1.3409, + "theoretical_loss": 3.488116338546674, + "tokens_seen": 1808138240 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002283751253761284, + "loss": 1.315, + "theoretical_loss": 3.488105675136794, + "tokens_seen": 1808203776 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002283650952858576, + "loss": 1.3879, + "theoretical_loss": 3.4880950122215992, + "tokens_seen": 1808269312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022835506519558677, + "loss": 1.2825, + "theoretical_loss": 3.4880843498010483, + "tokens_seen": 1808334848 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022834503510531595, + "loss": 1.3126, + "theoretical_loss": 3.4880736878751, + "tokens_seen": 1808400384 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022833500501504516, + "loss": 1.471, + "theoretical_loss": 3.4880630264437142, + "tokens_seen": 1808465920 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022832497492477434, + "loss": 1.2921, + "theoretical_loss": 3.48805236550685, + "tokens_seen": 1808531456 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022831494483450352, + "loss": 1.4913, + "theoretical_loss": 3.488041705064466, + "tokens_seen": 1808596992 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002283049147442327, + "loss": 1.3027, + "theoretical_loss": 3.4880310451165215, + "tokens_seen": 1808662528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022829488465396188, + "loss": 1.356, + "theoretical_loss": 3.488020385662976, + "tokens_seen": 1808728064 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 660069, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.33481764793396, + "objective/train/theoretical_loss": 3.4880097267037877, + "objective/train/tokens_used": 1829253600, + "theoretical_loss": 3.4880097267037877, + "tokens_seen": 1808793600 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002282848545636911, + "loss": 1.3473, + "theoretical_loss": 3.4880097267037877, + "tokens_seen": 1808793600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022827482447342027, + "loss": 1.3757, + "theoretical_loss": 3.4879990682389175, + "tokens_seen": 1808859136 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022826479438314945, + "loss": 1.3439, + "theoretical_loss": 3.487988410268323, + "tokens_seen": 1808924672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022825476429287863, + "loss": 1.2695, + "theoretical_loss": 3.4879777527919638, + "tokens_seen": 1808990208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022824473420260784, + "loss": 1.3392, + "theoretical_loss": 3.4879670958097995, + "tokens_seen": 1809055744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022823470411233702, + "loss": 1.3345, + "theoretical_loss": 3.4879564393217892, + "tokens_seen": 1809121280 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002282246740220662, + "loss": 1.3816, + "theoretical_loss": 3.4879457833278913, + "tokens_seen": 1809186816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022821464393179538, + "loss": 1.4037, + "theoretical_loss": 3.4879351278280657, + "tokens_seen": 1809252352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022820461384152457, + "loss": 1.346, + "theoretical_loss": 3.4879244728222716, + "tokens_seen": 1809317888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022819458375125377, + "loss": 1.3186, + "theoretical_loss": 3.487913818310468, + "tokens_seen": 1809383424 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022818455366098296, + "loss": 1.2832, + "theoretical_loss": 3.4879031642926144, + "tokens_seen": 1809448960 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022817452357071214, + "loss": 1.2489, + "theoretical_loss": 3.4878925107686696, + "tokens_seen": 1809514496 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022816449348044132, + "loss": 1.3539, + "theoretical_loss": 3.487881857738593, + "tokens_seen": 1809580032 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022815446339017053, + "loss": 1.3319, + "theoretical_loss": 3.4878712052023433, + "tokens_seen": 1809645568 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002281444332998997, + "loss": 1.2835, + "theoretical_loss": 3.487860553159881, + "tokens_seen": 1809711104 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002281344032096289, + "loss": 1.2479, + "theoretical_loss": 3.487849901611164, + "tokens_seen": 1809776640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022812437311935807, + "loss": 1.197, + "theoretical_loss": 3.487839250556153, + "tokens_seen": 1809842176 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022811434302908725, + "loss": 1.418, + "theoretical_loss": 3.4878285999948053, + "tokens_seen": 1809907712 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022810431293881646, + "loss": 1.3822, + "theoretical_loss": 3.4878179499270816, + "tokens_seen": 1809973248 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022809428284854564, + "loss": 1.3507, + "theoretical_loss": 3.487807300352941, + "tokens_seen": 1810038784 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022808425275827482, + "loss": 1.3216, + "theoretical_loss": 3.487796651272342, + "tokens_seen": 1810104320 + }, + { + "epoch": 0.55, + "learning_rate": 0.000228074222668004, + "loss": 1.2948, + "theoretical_loss": 3.4877860026852443, + "tokens_seen": 1810169856 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002280641925777332, + "loss": 1.3067, + "theoretical_loss": 3.4877753545916077, + "tokens_seen": 1810235392 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002280541624874624, + "loss": 1.3919, + "theoretical_loss": 3.487764706991391, + "tokens_seen": 1810300928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022804413239719157, + "loss": 1.3852, + "theoretical_loss": 3.487754059884553, + "tokens_seen": 1810366464 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022803410230692075, + "loss": 1.4456, + "theoretical_loss": 3.487743413271054, + "tokens_seen": 1810432000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022802407221664993, + "loss": 1.329, + "theoretical_loss": 3.487732767150852, + "tokens_seen": 1810497536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022801404212637914, + "loss": 1.3126, + "theoretical_loss": 3.487722121523907, + "tokens_seen": 1810563072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022800401203610832, + "loss": 1.428, + "theoretical_loss": 3.487711476390179, + "tokens_seen": 1810628608 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002279939819458375, + "loss": 1.3656, + "theoretical_loss": 3.4877008317496268, + "tokens_seen": 1810694144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022798395185556669, + "loss": 1.3097, + "theoretical_loss": 3.4876901876022086, + "tokens_seen": 1810759680 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002279739217652959, + "loss": 1.3353, + "theoretical_loss": 3.487679543947885, + "tokens_seen": 1810825216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022796389167502508, + "loss": 1.3124, + "theoretical_loss": 3.4876689007866153, + "tokens_seen": 1810890752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022795386158475426, + "loss": 1.3064, + "theoretical_loss": 3.4876582581183575, + "tokens_seen": 1810956288 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022794383149448344, + "loss": 1.3369, + "theoretical_loss": 3.4876476159430725, + "tokens_seen": 1811021824 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022793380140421262, + "loss": 1.299, + "theoretical_loss": 3.4876369742607194, + "tokens_seen": 1811087360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022792377131394183, + "loss": 1.3135, + "theoretical_loss": 3.4876263330712565, + "tokens_seen": 1811152896 + }, + { + "epoch": 0.55, + "learning_rate": 0.000227913741223671, + "loss": 1.3544, + "theoretical_loss": 3.4876156923746437, + "tokens_seen": 1811218432 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002279037111334002, + "loss": 1.3243, + "theoretical_loss": 3.487605052170841, + "tokens_seen": 1811283968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022789368104312937, + "loss": 1.4147, + "theoretical_loss": 3.4875944124598064, + "tokens_seen": 1811349504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022788365095285858, + "loss": 1.3446, + "theoretical_loss": 3.487583773241501, + "tokens_seen": 1811415040 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022787362086258776, + "loss": 1.3461, + "theoretical_loss": 3.4875731345158822, + "tokens_seen": 1811480576 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022786359077231694, + "loss": 1.324, + "theoretical_loss": 3.4875624962829104, + "tokens_seen": 1811546112 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022785356068204612, + "loss": 1.2648, + "theoretical_loss": 3.4875518585425453, + "tokens_seen": 1811611648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022784353059177536, + "loss": 1.3315, + "theoretical_loss": 3.487541221294746, + "tokens_seen": 1811677184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022783350050150454, + "loss": 1.291, + "theoretical_loss": 3.4875305845394715, + "tokens_seen": 1811742720 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022782347041123372, + "loss": 1.3648, + "theoretical_loss": 3.487519948276681, + "tokens_seen": 1811808256 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002278134403209629, + "loss": 1.2743, + "theoretical_loss": 3.487509312506335, + "tokens_seen": 1811873792 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022780341023069208, + "loss": 1.4024, + "theoretical_loss": 3.4874986772283916, + "tokens_seen": 1811939328 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002277933801404213, + "loss": 1.3296, + "theoretical_loss": 3.487488042442811, + "tokens_seen": 1812004864 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 661289, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.273240566253662, + "objective/train/theoretical_loss": 3.4874774081495525, + "objective/train/tokens_used": 1832530400, + "theoretical_loss": 3.4874774081495525, + "tokens_seen": 1812070400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022778335005015047, + "loss": 1.3175, + "theoretical_loss": 3.4874774081495525, + "tokens_seen": 1812070400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022777331995987965, + "loss": 1.2883, + "theoretical_loss": 3.487466774348575, + "tokens_seen": 1812135936 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022776328986960883, + "loss": 1.3631, + "theoretical_loss": 3.4874561410398393, + "tokens_seen": 1812201472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022775325977933804, + "loss": 1.2508, + "theoretical_loss": 3.4874455082233027, + "tokens_seen": 1812267008 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022774322968906722, + "loss": 1.3159, + "theoretical_loss": 3.487434875898926, + "tokens_seen": 1812332544 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002277331995987964, + "loss": 1.371, + "theoretical_loss": 3.4874242440666685, + "tokens_seen": 1812398080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022772316950852558, + "loss": 1.5025, + "theoretical_loss": 3.487413612726489, + "tokens_seen": 1812463616 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022771313941825477, + "loss": 1.3828, + "theoretical_loss": 3.4874029818783487, + "tokens_seen": 1812529152 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022770310932798397, + "loss": 1.3459, + "theoretical_loss": 3.4873923515222045, + "tokens_seen": 1812594688 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022769307923771316, + "loss": 1.3739, + "theoretical_loss": 3.487381721658018, + "tokens_seen": 1812660224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022768304914744234, + "loss": 1.3288, + "theoretical_loss": 3.487371092285747, + "tokens_seen": 1812725760 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022767301905717152, + "loss": 1.297, + "theoretical_loss": 3.487360463405352, + "tokens_seen": 1812791296 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022766298896690073, + "loss": 1.2989, + "theoretical_loss": 3.4873498350167917, + "tokens_seen": 1812856832 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002276529588766299, + "loss": 1.2597, + "theoretical_loss": 3.4873392071200264, + "tokens_seen": 1812922368 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002276429287863591, + "loss": 1.4339, + "theoretical_loss": 3.4873285797150153, + "tokens_seen": 1812987904 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022763289869608827, + "loss": 1.4139, + "theoretical_loss": 3.4873179528017175, + "tokens_seen": 1813053440 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022762286860581745, + "loss": 1.4523, + "theoretical_loss": 3.487307326380093, + "tokens_seen": 1813118976 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022761283851554666, + "loss": 1.289, + "theoretical_loss": 3.487296700450101, + "tokens_seen": 1813184512 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022760280842527584, + "loss": 1.3046, + "theoretical_loss": 3.487286075011701, + "tokens_seen": 1813250048 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022759277833500502, + "loss": 1.2973, + "theoretical_loss": 3.4872754500648524, + "tokens_seen": 1813315584 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002275827482447342, + "loss": 1.1516, + "theoretical_loss": 3.4872648256095147, + "tokens_seen": 1813381120 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002275727181544634, + "loss": 1.4012, + "theoretical_loss": 3.487254201645648, + "tokens_seen": 1813446656 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002275626880641926, + "loss": 1.4246, + "theoretical_loss": 3.4872435781732105, + "tokens_seen": 1813512192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022755265797392177, + "loss": 1.3295, + "theoretical_loss": 3.487232955192163, + "tokens_seen": 1813577728 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022754262788365095, + "loss": 1.416, + "theoretical_loss": 3.487222332702464, + "tokens_seen": 1813643264 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022753259779338013, + "loss": 1.3479, + "theoretical_loss": 3.487211710704074, + "tokens_seen": 1813708800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022752256770310934, + "loss": 1.3208, + "theoretical_loss": 3.487201089196952, + "tokens_seen": 1813774336 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022751253761283852, + "loss": 1.317, + "theoretical_loss": 3.4871904681810575, + "tokens_seen": 1813839872 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002275025075225677, + "loss": 1.4357, + "theoretical_loss": 3.48717984765635, + "tokens_seen": 1813905408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022749247743229689, + "loss": 1.3918, + "theoretical_loss": 3.4871692276227897, + "tokens_seen": 1813970944 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002274824473420261, + "loss": 1.4123, + "theoretical_loss": 3.487158608080335, + "tokens_seen": 1814036480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022747241725175528, + "loss": 1.4966, + "theoretical_loss": 3.487147989028946, + "tokens_seen": 1814102016 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022746238716148446, + "loss": 1.3248, + "theoretical_loss": 3.4871373704685826, + "tokens_seen": 1814167552 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022745235707121364, + "loss": 1.4268, + "theoretical_loss": 3.487126752399204, + "tokens_seen": 1814233088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022744232698094282, + "loss": 1.3955, + "theoretical_loss": 3.48711613482077, + "tokens_seen": 1814298624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022743229689067203, + "loss": 1.4109, + "theoretical_loss": 3.48710551773324, + "tokens_seen": 1814364160 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002274222668004012, + "loss": 1.2477, + "theoretical_loss": 3.4870949011365733, + "tokens_seen": 1814429696 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002274122367101304, + "loss": 1.3671, + "theoretical_loss": 3.48708428503073, + "tokens_seen": 1814495232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022740220661985957, + "loss": 1.3055, + "theoretical_loss": 3.487073669415669, + "tokens_seen": 1814560768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022739217652958878, + "loss": 1.2629, + "theoretical_loss": 3.487063054291351, + "tokens_seen": 1814626304 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022738214643931796, + "loss": 1.3928, + "theoretical_loss": 3.4870524396577345, + "tokens_seen": 1814691840 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022737211634904714, + "loss": 1.2462, + "theoretical_loss": 3.4870418255147793, + "tokens_seen": 1814757376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022736208625877632, + "loss": 1.278, + "theoretical_loss": 3.487031211862446, + "tokens_seen": 1814822912 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002273520561685055, + "loss": 1.3245, + "theoretical_loss": 3.4870205987006924, + "tokens_seen": 1814888448 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002273420260782347, + "loss": 1.3971, + "theoretical_loss": 3.48700998602948, + "tokens_seen": 1814953984 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002273319959879639, + "loss": 1.3934, + "theoretical_loss": 3.486999373848767, + "tokens_seen": 1815019520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022732196589769307, + "loss": 1.4286, + "theoretical_loss": 3.4869887621585134, + "tokens_seen": 1815085056 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022731193580742225, + "loss": 1.3769, + "theoretical_loss": 3.486978150958679, + "tokens_seen": 1815150592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022730190571715146, + "loss": 1.2716, + "theoretical_loss": 3.486967540249224, + "tokens_seen": 1815216128 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022729187562688064, + "loss": 1.4168, + "theoretical_loss": 3.486956930030107, + "tokens_seen": 1815281664 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 662349, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2711076736450195, + "objective/train/theoretical_loss": 3.4869463203012883, + "objective/train/tokens_used": 1835807200, + "theoretical_loss": 3.4869463203012883, + "tokens_seen": 1815347200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022728184553660983, + "loss": 1.3105, + "theoretical_loss": 3.4869463203012883, + "tokens_seen": 1815347200 + }, + { + "epoch": 0.55, + "learning_rate": 0.000227271815446339, + "loss": 1.4028, + "theoretical_loss": 3.486935711062727, + "tokens_seen": 1815412736 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022726178535606821, + "loss": 1.331, + "theoretical_loss": 3.4869251023143835, + "tokens_seen": 1815478272 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002272517552657974, + "loss": 1.3722, + "theoretical_loss": 3.486914494056217, + "tokens_seen": 1815543808 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022724172517552658, + "loss": 1.5092, + "theoretical_loss": 3.4869038862881867, + "tokens_seen": 1815609344 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022723169508525576, + "loss": 1.3837, + "theoretical_loss": 3.4868932790102534, + "tokens_seen": 1815674880 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022722166499498494, + "loss": 1.3921, + "theoretical_loss": 3.4868826722223756, + "tokens_seen": 1815740416 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022721163490471415, + "loss": 1.2171, + "theoretical_loss": 3.486872065924514, + "tokens_seen": 1815805952 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022720160481444333, + "loss": 1.2644, + "theoretical_loss": 3.486861460116627, + "tokens_seen": 1815871488 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002271915747241725, + "loss": 1.3465, + "theoretical_loss": 3.4868508547986754, + "tokens_seen": 1815937024 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002271815446339017, + "loss": 1.3914, + "theoretical_loss": 3.486840249970619, + "tokens_seen": 1816002560 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002271715145436309, + "loss": 1.3508, + "theoretical_loss": 3.4868296456324166, + "tokens_seen": 1816068096 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022716148445336008, + "loss": 1.4975, + "theoretical_loss": 3.4868190417840284, + "tokens_seen": 1816133632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022715145436308926, + "loss": 1.3717, + "theoretical_loss": 3.486808438425414, + "tokens_seen": 1816199168 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022714142427281844, + "loss": 1.3177, + "theoretical_loss": 3.4867978355565334, + "tokens_seen": 1816264704 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022713139418254762, + "loss": 1.2741, + "theoretical_loss": 3.4867872331773455, + "tokens_seen": 1816330240 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022712136409227683, + "loss": 1.2506, + "theoretical_loss": 3.4867766312878112, + "tokens_seen": 1816395776 + }, + { + "epoch": 0.55, + "learning_rate": 0.000227111334002006, + "loss": 1.3112, + "theoretical_loss": 3.486766029887889, + "tokens_seen": 1816461312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022710130391173522, + "loss": 1.3256, + "theoretical_loss": 3.4867554289775398, + "tokens_seen": 1816526848 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002270912738214644, + "loss": 1.3706, + "theoretical_loss": 3.4867448285567226, + "tokens_seen": 1816592384 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002270812437311936, + "loss": 1.3687, + "theoretical_loss": 3.486734228625397, + "tokens_seen": 1816657920 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002270712136409228, + "loss": 1.3314, + "theoretical_loss": 3.486723629183523, + "tokens_seen": 1816723456 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022706118355065197, + "loss": 1.2839, + "theoretical_loss": 3.4867130302310607, + "tokens_seen": 1816788992 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022705115346038115, + "loss": 1.3905, + "theoretical_loss": 3.486702431767969, + "tokens_seen": 1816854528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022704112337011033, + "loss": 1.3749, + "theoretical_loss": 3.4866918337942088, + "tokens_seen": 1816920064 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022703109327983954, + "loss": 1.3471, + "theoretical_loss": 3.486681236309739, + "tokens_seen": 1816985600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022702106318956872, + "loss": 1.314, + "theoretical_loss": 3.4866706393145197, + "tokens_seen": 1817051136 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002270110330992979, + "loss": 1.3222, + "theoretical_loss": 3.48666004280851, + "tokens_seen": 1817116672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022700100300902709, + "loss": 1.3058, + "theoretical_loss": 3.4866494467916707, + "tokens_seen": 1817182208 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002269909729187563, + "loss": 1.3178, + "theoretical_loss": 3.486638851263961, + "tokens_seen": 1817247744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022698094282848548, + "loss": 1.421, + "theoretical_loss": 3.486628256225341, + "tokens_seen": 1817313280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022697091273821466, + "loss": 1.2737, + "theoretical_loss": 3.48661766167577, + "tokens_seen": 1817378816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022696088264794384, + "loss": 1.3638, + "theoretical_loss": 3.486607067615208, + "tokens_seen": 1817444352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022695085255767302, + "loss": 1.3624, + "theoretical_loss": 3.486596474043615, + "tokens_seen": 1817509888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022694082246740223, + "loss": 1.3141, + "theoretical_loss": 3.4865858809609507, + "tokens_seen": 1817575424 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002269307923771314, + "loss": 1.4284, + "theoretical_loss": 3.486575288367175, + "tokens_seen": 1817640960 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002269207622868606, + "loss": 1.3517, + "theoretical_loss": 3.4865646962622474, + "tokens_seen": 1817706496 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022691073219658977, + "loss": 1.4649, + "theoretical_loss": 3.486554104646128, + "tokens_seen": 1817772032 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022690070210631898, + "loss": 1.3488, + "theoretical_loss": 3.486543513518776, + "tokens_seen": 1817837568 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022689067201604816, + "loss": 1.3686, + "theoretical_loss": 3.4865329228801523, + "tokens_seen": 1817903104 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022688064192577734, + "loss": 1.3209, + "theoretical_loss": 3.486522332730216, + "tokens_seen": 1817968640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022687061183550652, + "loss": 1.2412, + "theoretical_loss": 3.4865117430689274, + "tokens_seen": 1818034176 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002268605817452357, + "loss": 1.386, + "theoretical_loss": 3.4865011538962456, + "tokens_seen": 1818099712 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002268505516549649, + "loss": 1.3941, + "theoretical_loss": 3.486490565212131, + "tokens_seen": 1818165248 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002268405215646941, + "loss": 1.232, + "theoretical_loss": 3.486479977016544, + "tokens_seen": 1818230784 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022683049147442327, + "loss": 1.2914, + "theoretical_loss": 3.486469389309443, + "tokens_seen": 1818296320 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022682046138415245, + "loss": 1.3549, + "theoretical_loss": 3.486458802090789, + "tokens_seen": 1818361856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022681043129388166, + "loss": 1.3979, + "theoretical_loss": 3.4864482153605416, + "tokens_seen": 1818427392 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022680040120361084, + "loss": 1.4329, + "theoretical_loss": 3.4864376291186607, + "tokens_seen": 1818492928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022679037111334003, + "loss": 1.2425, + "theoretical_loss": 3.4864270433651052, + "tokens_seen": 1818558464 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 663450, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3358533382415771, + "objective/train/theoretical_loss": 3.486416458099837, + "objective/train/tokens_used": 1839084000, + "theoretical_loss": 3.486416458099837, + "tokens_seen": 1818624000 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002267803410230692, + "loss": 1.3541, + "theoretical_loss": 3.486416458099837, + "tokens_seen": 1818624000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022677031093279841, + "loss": 1.3841, + "theoretical_loss": 3.486405873322814, + "tokens_seen": 1818689536 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002267602808425276, + "loss": 1.3778, + "theoretical_loss": 3.4863952890339975, + "tokens_seen": 1818755072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022675025075225678, + "loss": 1.3363, + "theoretical_loss": 3.486384705233346, + "tokens_seen": 1818820608 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022674022066198596, + "loss": 1.3202, + "theoretical_loss": 3.4863741219208206, + "tokens_seen": 1818886144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022673019057171514, + "loss": 1.3827, + "theoretical_loss": 3.486363539096381, + "tokens_seen": 1818951680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022672016048144435, + "loss": 1.3647, + "theoretical_loss": 3.486352956759987, + "tokens_seen": 1819017216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022671013039117353, + "loss": 1.2969, + "theoretical_loss": 3.4863423749115983, + "tokens_seen": 1819082752 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002267001003009027, + "loss": 1.2956, + "theoretical_loss": 3.486331793551175, + "tokens_seen": 1819148288 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002266900702106319, + "loss": 1.3666, + "theoretical_loss": 3.4863212126786762, + "tokens_seen": 1819213824 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002266800401203611, + "loss": 1.381, + "theoretical_loss": 3.4863106322940633, + "tokens_seen": 1819279360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022667001003009028, + "loss": 1.4094, + "theoretical_loss": 3.4863000523972953, + "tokens_seen": 1819344896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022665997993981946, + "loss": 1.393, + "theoretical_loss": 3.4862894729883322, + "tokens_seen": 1819410432 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022664994984954864, + "loss": 1.3796, + "theoretical_loss": 3.4862788940671345, + "tokens_seen": 1819475968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022663991975927782, + "loss": 1.3814, + "theoretical_loss": 3.4862683156336614, + "tokens_seen": 1819541504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022662988966900703, + "loss": 1.34, + "theoretical_loss": 3.486257737687873, + "tokens_seen": 1819607040 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002266198595787362, + "loss": 1.3205, + "theoretical_loss": 3.4862471602297296, + "tokens_seen": 1819672576 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002266098294884654, + "loss": 1.3762, + "theoretical_loss": 3.486236583259191, + "tokens_seen": 1819738112 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022659979939819458, + "loss": 1.3053, + "theoretical_loss": 3.4862260067762176, + "tokens_seen": 1819803648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022658976930792378, + "loss": 1.2837, + "theoretical_loss": 3.486215430780769, + "tokens_seen": 1819869184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022657973921765296, + "loss": 1.4751, + "theoretical_loss": 3.486204855272804, + "tokens_seen": 1819934720 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022656970912738215, + "loss": 1.2793, + "theoretical_loss": 3.4861942802522843, + "tokens_seen": 1820000256 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022655967903711133, + "loss": 1.3148, + "theoretical_loss": 3.486183705719169, + "tokens_seen": 1820065792 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002265496489468405, + "loss": 1.443, + "theoretical_loss": 3.486173131673419, + "tokens_seen": 1820131328 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022653961885656972, + "loss": 1.2821, + "theoretical_loss": 3.4861625581149926, + "tokens_seen": 1820196864 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002265295887662989, + "loss": 1.2343, + "theoretical_loss": 3.4861519850438514, + "tokens_seen": 1820262400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022651955867602808, + "loss": 1.4429, + "theoretical_loss": 3.486141412459955, + "tokens_seen": 1820327936 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022650952858575726, + "loss": 1.1897, + "theoretical_loss": 3.486130840363263, + "tokens_seen": 1820393472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022649949849548647, + "loss": 1.3765, + "theoretical_loss": 3.4861202687537354, + "tokens_seen": 1820459008 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022648946840521565, + "loss": 1.3484, + "theoretical_loss": 3.486109697631333, + "tokens_seen": 1820524544 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022647943831494483, + "loss": 1.4332, + "theoretical_loss": 3.486099126996015, + "tokens_seen": 1820590080 + }, + { + "epoch": 0.55, + "learning_rate": 0.000226469408224674, + "loss": 1.5445, + "theoretical_loss": 3.4860885568477413, + "tokens_seen": 1820655616 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002264593781344032, + "loss": 1.3388, + "theoretical_loss": 3.486077987186473, + "tokens_seen": 1820721152 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002264493480441324, + "loss": 1.3426, + "theoretical_loss": 3.486067418012169, + "tokens_seen": 1820786688 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022643931795386158, + "loss": 1.2948, + "theoretical_loss": 3.4860568493247897, + "tokens_seen": 1820852224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022642928786359076, + "loss": 1.3714, + "theoretical_loss": 3.4860462811242954, + "tokens_seen": 1820917760 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022641925777331994, + "loss": 1.3797, + "theoretical_loss": 3.4860357134106463, + "tokens_seen": 1820983296 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022640922768304915, + "loss": 1.3395, + "theoretical_loss": 3.4860251461838017, + "tokens_seen": 1821048832 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022639919759277833, + "loss": 1.4993, + "theoretical_loss": 3.4860145794437223, + "tokens_seen": 1821114368 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022638916750250751, + "loss": 1.3742, + "theoretical_loss": 3.486004013190368, + "tokens_seen": 1821179904 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002263791374122367, + "loss": 1.3851, + "theoretical_loss": 3.4859934474236987, + "tokens_seen": 1821245440 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022636910732196588, + "loss": 1.3613, + "theoretical_loss": 3.4859828821436745, + "tokens_seen": 1821310976 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022635907723169508, + "loss": 1.453, + "theoretical_loss": 3.4859723173502557, + "tokens_seen": 1821376512 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002263490471414243, + "loss": 1.3264, + "theoretical_loss": 3.4859617530434024, + "tokens_seen": 1821442048 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022633901705115347, + "loss": 1.4233, + "theoretical_loss": 3.485951189223074, + "tokens_seen": 1821507584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022632898696088266, + "loss": 1.3585, + "theoretical_loss": 3.4859406258892314, + "tokens_seen": 1821573120 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022631895687061186, + "loss": 1.2319, + "theoretical_loss": 3.4859300630418346, + "tokens_seen": 1821638656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022630892678034104, + "loss": 1.2675, + "theoretical_loss": 3.4859195006808434, + "tokens_seen": 1821704192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022629889669007023, + "loss": 1.3887, + "theoretical_loss": 3.485908938806218, + "tokens_seen": 1821769728 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002262888665997994, + "loss": 1.3901, + "theoretical_loss": 3.485898377417919, + "tokens_seen": 1821835264 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 664649, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2136337757110596, + "objective/train/theoretical_loss": 3.4858878165159055, + "objective/train/tokens_used": 1842360800, + "theoretical_loss": 3.4858878165159055, + "tokens_seen": 1821900800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022627883650952861, + "loss": 1.2902, + "theoretical_loss": 3.4858878165159055, + "tokens_seen": 1821900800 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002262688064192578, + "loss": 1.2673, + "theoretical_loss": 3.485877256100138, + "tokens_seen": 1821966336 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022625877632898698, + "loss": 1.4068, + "theoretical_loss": 3.4858666961705773, + "tokens_seen": 1822031872 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022624874623871616, + "loss": 1.4113, + "theoretical_loss": 3.4858561367271825, + "tokens_seen": 1822097408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022623871614844534, + "loss": 1.3943, + "theoretical_loss": 3.485845577769915, + "tokens_seen": 1822162944 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022622868605817455, + "loss": 1.5087, + "theoretical_loss": 3.4858350192987335, + "tokens_seen": 1822228480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022621865596790373, + "loss": 1.4524, + "theoretical_loss": 3.485824461313599, + "tokens_seen": 1822294016 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002262086258776329, + "loss": 1.3179, + "theoretical_loss": 3.4858139038144715, + "tokens_seen": 1822359552 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002261985957873621, + "loss": 1.3665, + "theoretical_loss": 3.485803346801311, + "tokens_seen": 1822425088 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002261885656970913, + "loss": 1.3733, + "theoretical_loss": 3.485792790274078, + "tokens_seen": 1822490624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022617853560682048, + "loss": 1.3629, + "theoretical_loss": 3.4857822342327323, + "tokens_seen": 1822556160 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022616850551654966, + "loss": 1.2246, + "theoretical_loss": 3.4857716786772346, + "tokens_seen": 1822621696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022615847542627884, + "loss": 1.4666, + "theoretical_loss": 3.4857611236075443, + "tokens_seen": 1822687232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022614844533600802, + "loss": 1.4347, + "theoretical_loss": 3.4857505690236223, + "tokens_seen": 1822752768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022613841524573723, + "loss": 1.3512, + "theoretical_loss": 3.485740014925428, + "tokens_seen": 1822818304 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002261283851554664, + "loss": 1.3285, + "theoretical_loss": 3.485729461312922, + "tokens_seen": 1822883840 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002261183550651956, + "loss": 1.4098, + "theoretical_loss": 3.4857189081860644, + "tokens_seen": 1822949376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022610832497492478, + "loss": 1.3623, + "theoretical_loss": 3.485708355544816, + "tokens_seen": 1823014912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022609829488465398, + "loss": 1.3585, + "theoretical_loss": 3.4856978033891366, + "tokens_seen": 1823080448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022608826479438316, + "loss": 1.2473, + "theoretical_loss": 3.4856872517189856, + "tokens_seen": 1823145984 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022607823470411235, + "loss": 1.2912, + "theoretical_loss": 3.4856767005343245, + "tokens_seen": 1823211520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022606820461384153, + "loss": 1.485, + "theoretical_loss": 3.4856661498351125, + "tokens_seen": 1823277056 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002260581745235707, + "loss": 1.4318, + "theoretical_loss": 3.4856555996213103, + "tokens_seen": 1823342592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022604814443329992, + "loss": 1.4543, + "theoretical_loss": 3.485645049892878, + "tokens_seen": 1823408128 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002260381143430291, + "loss": 1.429, + "theoretical_loss": 3.485634500649777, + "tokens_seen": 1823473664 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022602808425275828, + "loss": 1.3057, + "theoretical_loss": 3.4856239518919647, + "tokens_seen": 1823539200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022601805416248746, + "loss": 1.4478, + "theoretical_loss": 3.4856134036194044, + "tokens_seen": 1823604736 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022600802407221667, + "loss": 1.4164, + "theoretical_loss": 3.485602855832054, + "tokens_seen": 1823670272 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022599799398194585, + "loss": 1.3165, + "theoretical_loss": 3.4855923085298754, + "tokens_seen": 1823735808 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022598796389167503, + "loss": 1.407, + "theoretical_loss": 3.4855817617128277, + "tokens_seen": 1823801344 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002259779338014042, + "loss": 1.3786, + "theoretical_loss": 3.4855712153808724, + "tokens_seen": 1823866880 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002259679037111334, + "loss": 1.3223, + "theoretical_loss": 3.4855606695339683, + "tokens_seen": 1823932416 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002259578736208626, + "loss": 1.3049, + "theoretical_loss": 3.485550124172077, + "tokens_seen": 1823997952 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022594784353059178, + "loss": 1.3864, + "theoretical_loss": 3.4855395792951573, + "tokens_seen": 1824063488 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022593781344032096, + "loss": 1.3937, + "theoretical_loss": 3.4855290349031707, + "tokens_seen": 1824129024 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022592778335005014, + "loss": 1.3694, + "theoretical_loss": 3.4855184909960775, + "tokens_seen": 1824194560 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022591775325977935, + "loss": 1.4267, + "theoretical_loss": 3.4855079475738373, + "tokens_seen": 1824260096 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022590772316950853, + "loss": 1.1623, + "theoretical_loss": 3.4854974046364102, + "tokens_seen": 1824325632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022589769307923771, + "loss": 1.414, + "theoretical_loss": 3.485486862183757, + "tokens_seen": 1824391168 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002258876629889669, + "loss": 1.4866, + "theoretical_loss": 3.4854763202158385, + "tokens_seen": 1824456704 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022587763289869608, + "loss": 1.3631, + "theoretical_loss": 3.485465778732614, + "tokens_seen": 1824522240 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022586760280842528, + "loss": 1.4015, + "theoretical_loss": 3.485455237734045, + "tokens_seen": 1824587776 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022585757271815447, + "loss": 1.2868, + "theoretical_loss": 3.4854446972200903, + "tokens_seen": 1824653312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022584754262788365, + "loss": 1.3514, + "theoretical_loss": 3.485434157190711, + "tokens_seen": 1824718848 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022583751253761283, + "loss": 1.2142, + "theoretical_loss": 3.4854236176458677, + "tokens_seen": 1824784384 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022582748244734204, + "loss": 1.3529, + "theoretical_loss": 3.4854130785855206, + "tokens_seen": 1824849920 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022581745235707122, + "loss": 1.2734, + "theoretical_loss": 3.4854025400096296, + "tokens_seen": 1824915456 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002258074222668004, + "loss": 1.3808, + "theoretical_loss": 3.485392001918155, + "tokens_seen": 1824980992 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022579739217652958, + "loss": 1.3068, + "theoretical_loss": 3.485381464311058, + "tokens_seen": 1825046528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022578736208625876, + "loss": 1.2786, + "theoretical_loss": 3.485370927188298, + "tokens_seen": 1825112064 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 665640, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2016453742980957, + "objective/train/theoretical_loss": 3.4853603905498356, + "objective/train/tokens_used": 1845637600, + "theoretical_loss": 3.4853603905498356, + "tokens_seen": 1825177600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022577733199598797, + "loss": 1.3334, + "theoretical_loss": 3.4853603905498356, + "tokens_seen": 1825177600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022576730190571715, + "loss": 1.3907, + "theoretical_loss": 3.485349854395632, + "tokens_seen": 1825243136 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022575727181544633, + "loss": 1.4238, + "theoretical_loss": 3.4853393187256456, + "tokens_seen": 1825308672 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002257472417251755, + "loss": 1.4544, + "theoretical_loss": 3.4853287835398388, + "tokens_seen": 1825374208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022573721163490472, + "loss": 1.3917, + "theoretical_loss": 3.485318248838171, + "tokens_seen": 1825439744 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002257271815446339, + "loss": 1.23, + "theoretical_loss": 3.485307714620603, + "tokens_seen": 1825505280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022571715145436308, + "loss": 1.3338, + "theoretical_loss": 3.4852971808870947, + "tokens_seen": 1825570816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022570712136409226, + "loss": 1.339, + "theoretical_loss": 3.4852866476376065, + "tokens_seen": 1825636352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022569709127382147, + "loss": 1.3345, + "theoretical_loss": 3.485276114872099, + "tokens_seen": 1825701888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022568706118355065, + "loss": 1.3323, + "theoretical_loss": 3.4852655825905328, + "tokens_seen": 1825767424 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022567703109327983, + "loss": 1.415, + "theoretical_loss": 3.485255050792868, + "tokens_seen": 1825832960 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022566700100300902, + "loss": 1.3589, + "theoretical_loss": 3.485244519479065, + "tokens_seen": 1825898496 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002256569709127382, + "loss": 1.3957, + "theoretical_loss": 3.4852339886490844, + "tokens_seen": 1825964032 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002256469408224674, + "loss": 1.4166, + "theoretical_loss": 3.4852234583028863, + "tokens_seen": 1826029568 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022563691073219659, + "loss": 1.3461, + "theoretical_loss": 3.485212928440431, + "tokens_seen": 1826095104 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022562688064192577, + "loss": 1.3095, + "theoretical_loss": 3.48520239906168, + "tokens_seen": 1826160640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022561685055165495, + "loss": 1.3807, + "theoretical_loss": 3.4851918701665925, + "tokens_seen": 1826226176 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022560682046138416, + "loss": 1.5064, + "theoretical_loss": 3.485181341755129, + "tokens_seen": 1826291712 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022559679037111336, + "loss": 1.3538, + "theoretical_loss": 3.4851708138272506, + "tokens_seen": 1826357248 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022558676028084255, + "loss": 1.3171, + "theoretical_loss": 3.4851602863829174, + "tokens_seen": 1826422784 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022557673019057173, + "loss": 1.3096, + "theoretical_loss": 3.48514975942209, + "tokens_seen": 1826488320 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002255667001003009, + "loss": 1.3129, + "theoretical_loss": 3.4851392329447286, + "tokens_seen": 1826553856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022555667001003012, + "loss": 1.3128, + "theoretical_loss": 3.4851287069507935, + "tokens_seen": 1826619392 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002255466399197593, + "loss": 1.4178, + "theoretical_loss": 3.485118181440246, + "tokens_seen": 1826684928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022553660982948848, + "loss": 1.1476, + "theoretical_loss": 3.4851076564130454, + "tokens_seen": 1826750464 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022552657973921766, + "loss": 1.2837, + "theoretical_loss": 3.485097131869153, + "tokens_seen": 1826816000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022551654964894687, + "loss": 1.356, + "theoretical_loss": 3.4850866078085287, + "tokens_seen": 1826881536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022550651955867605, + "loss": 1.2637, + "theoretical_loss": 3.485076084231134, + "tokens_seen": 1826947072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022549648946840523, + "loss": 1.4394, + "theoretical_loss": 3.485065561136928, + "tokens_seen": 1827012608 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002254864593781344, + "loss": 1.286, + "theoretical_loss": 3.4850550385258723, + "tokens_seen": 1827078144 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002254764292878636, + "loss": 1.2631, + "theoretical_loss": 3.4850445163979265, + "tokens_seen": 1827143680 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002254663991975928, + "loss": 1.3426, + "theoretical_loss": 3.485033994753052, + "tokens_seen": 1827209216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022545636910732198, + "loss": 1.4274, + "theoretical_loss": 3.4850234735912085, + "tokens_seen": 1827274752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022544633901705116, + "loss": 1.3557, + "theoretical_loss": 3.4850129529123564, + "tokens_seen": 1827340288 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022543630892678034, + "loss": 1.204, + "theoretical_loss": 3.4850024327164575, + "tokens_seen": 1827405824 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022542627883650955, + "loss": 1.2906, + "theoretical_loss": 3.484991913003471, + "tokens_seen": 1827471360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022541624874623873, + "loss": 1.4447, + "theoretical_loss": 3.484981393773358, + "tokens_seen": 1827536896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022540621865596791, + "loss": 1.4191, + "theoretical_loss": 3.484970875026079, + "tokens_seen": 1827602432 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002253961885656971, + "loss": 1.4095, + "theoretical_loss": 3.4849603567615937, + "tokens_seen": 1827667968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022538615847542628, + "loss": 1.4485, + "theoretical_loss": 3.4849498389798637, + "tokens_seen": 1827733504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022537612838515548, + "loss": 1.4383, + "theoretical_loss": 3.4849393216808497, + "tokens_seen": 1827799040 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022536609829488467, + "loss": 1.3779, + "theoretical_loss": 3.484928804864511, + "tokens_seen": 1827864576 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022535606820461385, + "loss": 1.3355, + "theoretical_loss": 3.4849182885308094, + "tokens_seen": 1827930112 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022534603811434303, + "loss": 1.3172, + "theoretical_loss": 3.484907772679705, + "tokens_seen": 1827995648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022533600802407224, + "loss": 1.2093, + "theoretical_loss": 3.4848972573111574, + "tokens_seen": 1828061184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022532597793380142, + "loss": 1.4077, + "theoretical_loss": 3.484886742425129, + "tokens_seen": 1828126720 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002253159478435306, + "loss": 1.3217, + "theoretical_loss": 3.4848762280215784, + "tokens_seen": 1828192256 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022530591775325978, + "loss": 1.3082, + "theoretical_loss": 3.4848657141004677, + "tokens_seen": 1828257792 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022529588766298896, + "loss": 1.4559, + "theoretical_loss": 3.484855200661757, + "tokens_seen": 1828323328 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022528585757271817, + "loss": 1.3674, + "theoretical_loss": 3.4848446877054062, + "tokens_seen": 1828388864 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 666768, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.267007827758789, + "objective/train/theoretical_loss": 3.484834175231377, + "objective/train/tokens_used": 1848914400, + "theoretical_loss": 3.484834175231377, + "tokens_seen": 1828454400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022527582748244735, + "loss": 1.3631, + "theoretical_loss": 3.484834175231377, + "tokens_seen": 1828454400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022526579739217653, + "loss": 1.3072, + "theoretical_loss": 3.4848236632396294, + "tokens_seen": 1828519936 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002252557673019057, + "loss": 1.3408, + "theoretical_loss": 3.484813151730124, + "tokens_seen": 1828585472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022524573721163492, + "loss": 1.3987, + "theoretical_loss": 3.484802640702821, + "tokens_seen": 1828651008 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002252357071213641, + "loss": 1.307, + "theoretical_loss": 3.484792130157682, + "tokens_seen": 1828716544 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022522567703109328, + "loss": 1.2477, + "theoretical_loss": 3.484781620094667, + "tokens_seen": 1828782080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022521564694082246, + "loss": 1.3497, + "theoretical_loss": 3.4847711105137367, + "tokens_seen": 1828847616 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022520561685055167, + "loss": 1.2606, + "theoretical_loss": 3.4847606014148518, + "tokens_seen": 1828913152 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022519558676028085, + "loss": 1.4086, + "theoretical_loss": 3.484750092797972, + "tokens_seen": 1828978688 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022518555667001003, + "loss": 1.393, + "theoretical_loss": 3.4847395846630596, + "tokens_seen": 1829044224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022517552657973922, + "loss": 1.38, + "theoretical_loss": 3.4847290770100736, + "tokens_seen": 1829109760 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002251654964894684, + "loss": 1.2023, + "theoretical_loss": 3.484718569838976, + "tokens_seen": 1829175296 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002251554663991976, + "loss": 1.3473, + "theoretical_loss": 3.484708063149726, + "tokens_seen": 1829240832 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022514543630892679, + "loss": 1.2822, + "theoretical_loss": 3.484697556942286, + "tokens_seen": 1829306368 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022513540621865597, + "loss": 1.2812, + "theoretical_loss": 3.484687051216615, + "tokens_seen": 1829371904 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022512537612838515, + "loss": 1.3951, + "theoretical_loss": 3.4846765459726745, + "tokens_seen": 1829437440 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022511534603811436, + "loss": 1.3302, + "theoretical_loss": 3.484666041210425, + "tokens_seen": 1829502976 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022510531594784354, + "loss": 1.4462, + "theoretical_loss": 3.484655536929827, + "tokens_seen": 1829568512 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022509528585757272, + "loss": 1.3801, + "theoretical_loss": 3.4846450331308416, + "tokens_seen": 1829634048 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002250852557673019, + "loss": 1.3042, + "theoretical_loss": 3.484634529813429, + "tokens_seen": 1829699584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022507522567703108, + "loss": 1.3129, + "theoretical_loss": 3.48462402697755, + "tokens_seen": 1829765120 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002250651955867603, + "loss": 1.2751, + "theoretical_loss": 3.4846135246231658, + "tokens_seen": 1829830656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022505516549648947, + "loss": 1.5265, + "theoretical_loss": 3.4846030227502363, + "tokens_seen": 1829896192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022504513540621865, + "loss": 1.3598, + "theoretical_loss": 3.484592521358722, + "tokens_seen": 1829961728 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022503510531594783, + "loss": 1.2796, + "theoretical_loss": 3.4845820204485847, + "tokens_seen": 1830027264 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022502507522567704, + "loss": 1.3273, + "theoretical_loss": 3.4845715200197844, + "tokens_seen": 1830092800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022501504513540622, + "loss": 1.2895, + "theoretical_loss": 3.484561020072282, + "tokens_seen": 1830158336 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002250050150451354, + "loss": 1.3167, + "theoretical_loss": 3.484550520606038, + "tokens_seen": 1830223872 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022499498495486458, + "loss": 1.3918, + "theoretical_loss": 3.484540021621013, + "tokens_seen": 1830289408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022498495486459377, + "loss": 1.4174, + "theoretical_loss": 3.484529523117168, + "tokens_seen": 1830354944 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022497492477432297, + "loss": 1.3625, + "theoretical_loss": 3.4845190250944635, + "tokens_seen": 1830420480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022496489468405215, + "loss": 1.3774, + "theoretical_loss": 3.484508527552861, + "tokens_seen": 1830486016 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022495486459378134, + "loss": 1.2886, + "theoretical_loss": 3.48449803049232, + "tokens_seen": 1830551552 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022494483450351052, + "loss": 1.392, + "theoretical_loss": 3.4844875339128016, + "tokens_seen": 1830617088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022493480441323973, + "loss": 1.2176, + "theoretical_loss": 3.4844770378142673, + "tokens_seen": 1830682624 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002249247743229689, + "loss": 1.4114, + "theoretical_loss": 3.484466542196677, + "tokens_seen": 1830748160 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002249147442326981, + "loss": 1.2989, + "theoretical_loss": 3.484456047059992, + "tokens_seen": 1830813696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022490471414242727, + "loss": 1.3451, + "theoretical_loss": 3.484445552404173, + "tokens_seen": 1830879232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022489468405215645, + "loss": 1.2818, + "theoretical_loss": 3.48443505822918, + "tokens_seen": 1830944768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022488465396188566, + "loss": 1.5012, + "theoretical_loss": 3.484424564534975, + "tokens_seen": 1831010304 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022487462387161484, + "loss": 1.3237, + "theoretical_loss": 3.4844140713215177, + "tokens_seen": 1831075840 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022486459378134402, + "loss": 1.4341, + "theoretical_loss": 3.4844035785887693, + "tokens_seen": 1831141376 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002248545636910732, + "loss": 1.3003, + "theoretical_loss": 3.4843930863366905, + "tokens_seen": 1831206912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022484453360080244, + "loss": 1.297, + "theoretical_loss": 3.4843825945652425, + "tokens_seen": 1831272448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022483450351053162, + "loss": 1.2665, + "theoretical_loss": 3.484372103274385, + "tokens_seen": 1831337984 + }, + { + "epoch": 0.55, + "learning_rate": 0.0002248244734202608, + "loss": 1.348, + "theoretical_loss": 3.4843616124640797, + "tokens_seen": 1831403520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00022481444332998998, + "loss": 1.3076, + "theoretical_loss": 3.4843511221342873, + "tokens_seen": 1831469056 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022480441323971916, + "loss": 1.3342, + "theoretical_loss": 3.484340632284969, + "tokens_seen": 1831534592 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022479438314944837, + "loss": 1.2366, + "theoretical_loss": 3.4843301429160842, + "tokens_seen": 1831600128 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022478435305917755, + "loss": 1.3549, + "theoretical_loss": 3.484319654027595, + "tokens_seen": 1831665664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 668070, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.296405553817749, + "objective/train/theoretical_loss": 3.484309165619462, + "objective/train/tokens_used": 1852191200, + "theoretical_loss": 3.484309165619462, + "tokens_seen": 1831731200 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022477432296890673, + "loss": 1.3419, + "theoretical_loss": 3.484309165619462, + "tokens_seen": 1831731200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002247642928786359, + "loss": 1.2762, + "theoretical_loss": 3.484298677691645, + "tokens_seen": 1831796736 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022475426278836512, + "loss": 1.3665, + "theoretical_loss": 3.484288190244106, + "tokens_seen": 1831862272 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002247442326980943, + "loss": 1.3065, + "theoretical_loss": 3.484277703276806, + "tokens_seen": 1831927808 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022473420260782348, + "loss": 1.4421, + "theoretical_loss": 3.484267216789705, + "tokens_seen": 1831993344 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022472417251755266, + "loss": 1.3761, + "theoretical_loss": 3.4842567307827643, + "tokens_seen": 1832058880 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022471414242728187, + "loss": 1.2778, + "theoretical_loss": 3.484246245255944, + "tokens_seen": 1832124416 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022470411233701105, + "loss": 1.3234, + "theoretical_loss": 3.484235760209206, + "tokens_seen": 1832189952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022469408224674023, + "loss": 1.3827, + "theoretical_loss": 3.4842252756425105, + "tokens_seen": 1832255488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022468405215646942, + "loss": 1.333, + "theoretical_loss": 3.4842147915558184, + "tokens_seen": 1832321024 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002246740220661986, + "loss": 1.3726, + "theoretical_loss": 3.4842043079490908, + "tokens_seen": 1832386560 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002246639919759278, + "loss": 1.3906, + "theoretical_loss": 3.4841938248222886, + "tokens_seen": 1832452096 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022465396188565699, + "loss": 1.3778, + "theoretical_loss": 3.4841833421753723, + "tokens_seen": 1832517632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022464393179538617, + "loss": 1.5006, + "theoretical_loss": 3.484172860008303, + "tokens_seen": 1832583168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022463390170511535, + "loss": 1.3625, + "theoretical_loss": 3.4841623783210416, + "tokens_seen": 1832648704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022462387161484456, + "loss": 1.3509, + "theoretical_loss": 3.4841518971135486, + "tokens_seen": 1832714240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022461384152457374, + "loss": 1.2736, + "theoretical_loss": 3.4841414163857856, + "tokens_seen": 1832779776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022460381143430292, + "loss": 1.3462, + "theoretical_loss": 3.484130936137713, + "tokens_seen": 1832845312 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002245937813440321, + "loss": 1.2497, + "theoretical_loss": 3.4841204563692916, + "tokens_seen": 1832910848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022458375125376128, + "loss": 1.2907, + "theoretical_loss": 3.484109977080483, + "tokens_seen": 1832976384 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002245737211634905, + "loss": 1.256, + "theoretical_loss": 3.484099498271247, + "tokens_seen": 1833041920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022456369107321967, + "loss": 1.4239, + "theoretical_loss": 3.4840890199415453, + "tokens_seen": 1833107456 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022455366098294885, + "loss": 1.3882, + "theoretical_loss": 3.4840785420913387, + "tokens_seen": 1833172992 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022454363089267803, + "loss": 1.3638, + "theoretical_loss": 3.484068064720588, + "tokens_seen": 1833238528 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022453360080240724, + "loss": 1.2849, + "theoretical_loss": 3.4840575878292546, + "tokens_seen": 1833304064 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022452357071213642, + "loss": 1.4262, + "theoretical_loss": 3.484047111417298, + "tokens_seen": 1833369600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002245135406218656, + "loss": 1.3995, + "theoretical_loss": 3.484036635484681, + "tokens_seen": 1833435136 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022450351053159478, + "loss": 1.387, + "theoretical_loss": 3.4840261600313633, + "tokens_seen": 1833500672 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022449348044132397, + "loss": 1.3732, + "theoretical_loss": 3.4840156850573063, + "tokens_seen": 1833566208 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022448345035105317, + "loss": 1.4111, + "theoretical_loss": 3.4840052105624704, + "tokens_seen": 1833631744 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022447342026078236, + "loss": 1.3939, + "theoretical_loss": 3.4839947365468174, + "tokens_seen": 1833697280 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022446339017051154, + "loss": 1.3663, + "theoretical_loss": 3.483984263010308, + "tokens_seen": 1833762816 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022445336008024072, + "loss": 1.1629, + "theoretical_loss": 3.483973789952903, + "tokens_seen": 1833828352 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022444332998996993, + "loss": 1.4239, + "theoretical_loss": 3.4839633173745628, + "tokens_seen": 1833893888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002244332998996991, + "loss": 1.3572, + "theoretical_loss": 3.4839528452752493, + "tokens_seen": 1833959424 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002244232698094283, + "loss": 1.3753, + "theoretical_loss": 3.483942373654923, + "tokens_seen": 1834024960 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022441323971915747, + "loss": 1.3265, + "theoretical_loss": 3.483931902513545, + "tokens_seen": 1834090496 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022440320962888665, + "loss": 1.3068, + "theoretical_loss": 3.4839214318510763, + "tokens_seen": 1834156032 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022439317953861586, + "loss": 1.3008, + "theoretical_loss": 3.483910961667478, + "tokens_seen": 1834221568 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022438314944834504, + "loss": 1.2529, + "theoretical_loss": 3.483900491962711, + "tokens_seen": 1834287104 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022437311935807422, + "loss": 1.4188, + "theoretical_loss": 3.4838900227367358, + "tokens_seen": 1834352640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002243630892678034, + "loss": 1.4088, + "theoretical_loss": 3.4838795539895138, + "tokens_seen": 1834418176 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002243530591775326, + "loss": 1.366, + "theoretical_loss": 3.483869085721006, + "tokens_seen": 1834483712 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002243430290872618, + "loss": 1.3147, + "theoretical_loss": 3.483858617931174, + "tokens_seen": 1834549248 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022433299899699097, + "loss": 1.2441, + "theoretical_loss": 3.483848150619978, + "tokens_seen": 1834614784 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022432296890672015, + "loss": 1.3059, + "theoretical_loss": 3.483837683787379, + "tokens_seen": 1834680320 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022431293881644933, + "loss": 1.4459, + "theoretical_loss": 3.4838272174333387, + "tokens_seen": 1834745856 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022430290872617854, + "loss": 1.3745, + "theoretical_loss": 3.4838167515578173, + "tokens_seen": 1834811392 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022429287863590772, + "loss": 1.3703, + "theoretical_loss": 3.4838062861607764, + "tokens_seen": 1834876928 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002242828485456369, + "loss": 1.3656, + "theoretical_loss": 3.483795821242177, + "tokens_seen": 1834942464 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 669245, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3102073669433594, + "objective/train/theoretical_loss": 3.48378535680198, + "objective/train/tokens_used": 1855468000, + "theoretical_loss": 3.48378535680198, + "tokens_seen": 1835008000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022427281845536609, + "loss": 1.2737, + "theoretical_loss": 3.48378535680198, + "tokens_seen": 1835008000 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002242627883650953, + "loss": 1.3259, + "theoretical_loss": 3.4837748928401466, + "tokens_seen": 1835073536 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022425275827482448, + "loss": 1.3209, + "theoretical_loss": 3.4837644293566377, + "tokens_seen": 1835139072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022424272818455366, + "loss": 1.4039, + "theoretical_loss": 3.483753966351414, + "tokens_seen": 1835204608 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022423269809428284, + "loss": 1.3581, + "theoretical_loss": 3.4837435038244378, + "tokens_seen": 1835270144 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022422266800401202, + "loss": 1.3708, + "theoretical_loss": 3.4837330417756682, + "tokens_seen": 1835335680 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022421263791374123, + "loss": 1.2819, + "theoretical_loss": 3.483722580205068, + "tokens_seen": 1835401216 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002242026078234704, + "loss": 1.3947, + "theoretical_loss": 3.4837121191125977, + "tokens_seen": 1835466752 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002241925777331996, + "loss": 1.2355, + "theoretical_loss": 3.483701658498218, + "tokens_seen": 1835532288 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022418254764292877, + "loss": 1.3776, + "theoretical_loss": 3.483691198361891, + "tokens_seen": 1835597824 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022417251755265798, + "loss": 1.4429, + "theoretical_loss": 3.483680738703576, + "tokens_seen": 1835663360 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022416248746238716, + "loss": 1.3355, + "theoretical_loss": 3.4836702795232357, + "tokens_seen": 1835728896 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022415245737211634, + "loss": 1.2908, + "theoretical_loss": 3.4836598208208307, + "tokens_seen": 1835794432 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022414242728184552, + "loss": 1.4092, + "theoretical_loss": 3.483649362596322, + "tokens_seen": 1835859968 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002241323971915747, + "loss": 1.3652, + "theoretical_loss": 3.4836389048496708, + "tokens_seen": 1835925504 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002241223671013039, + "loss": 1.3869, + "theoretical_loss": 3.4836284475808386, + "tokens_seen": 1835991040 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002241123370110331, + "loss": 1.5008, + "theoretical_loss": 3.4836179907897855, + "tokens_seen": 1836056576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022410230692076227, + "loss": 1.3625, + "theoretical_loss": 3.4836075344764734, + "tokens_seen": 1836122112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022409227683049148, + "loss": 1.3487, + "theoretical_loss": 3.483597078640863, + "tokens_seen": 1836187648 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002240822467402207, + "loss": 1.3354, + "theoretical_loss": 3.4835866232829167, + "tokens_seen": 1836253184 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022407221664994987, + "loss": 1.3186, + "theoretical_loss": 3.4835761684025934, + "tokens_seen": 1836318720 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022406218655967905, + "loss": 1.4416, + "theoretical_loss": 3.483565713999856, + "tokens_seen": 1836384256 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022405215646940823, + "loss": 1.3981, + "theoretical_loss": 3.483555260074665, + "tokens_seen": 1836449792 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022404212637913744, + "loss": 1.3206, + "theoretical_loss": 3.4835448066269814, + "tokens_seen": 1836515328 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022403209628886662, + "loss": 1.3239, + "theoretical_loss": 3.483534353656767, + "tokens_seen": 1836580864 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002240220661985958, + "loss": 1.2647, + "theoretical_loss": 3.4835239011639825, + "tokens_seen": 1836646400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022401203610832498, + "loss": 1.1931, + "theoretical_loss": 3.483513449148589, + "tokens_seen": 1836711936 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022400200601805417, + "loss": 1.3941, + "theoretical_loss": 3.4835029976105476, + "tokens_seen": 1836777472 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022399197592778337, + "loss": 1.2703, + "theoretical_loss": 3.4834925465498197, + "tokens_seen": 1836843008 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022398194583751256, + "loss": 1.2359, + "theoretical_loss": 3.483482095966366, + "tokens_seen": 1836908544 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022397191574724174, + "loss": 1.389, + "theoretical_loss": 3.4834716458601482, + "tokens_seen": 1836974080 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022396188565697092, + "loss": 1.2461, + "theoretical_loss": 3.483461196231128, + "tokens_seen": 1837039616 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022395185556670013, + "loss": 1.2964, + "theoretical_loss": 3.4834507470792655, + "tokens_seen": 1837105152 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002239418254764293, + "loss": 1.2996, + "theoretical_loss": 3.4834402984045223, + "tokens_seen": 1837170688 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002239317953861585, + "loss": 1.3273, + "theoretical_loss": 3.4834298502068597, + "tokens_seen": 1837236224 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022392176529588767, + "loss": 1.3053, + "theoretical_loss": 3.4834194024862386, + "tokens_seen": 1837301760 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022391173520561685, + "loss": 1.3681, + "theoretical_loss": 3.483408955242621, + "tokens_seen": 1837367296 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022390170511534606, + "loss": 1.3486, + "theoretical_loss": 3.483398508475967, + "tokens_seen": 1837432832 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022389167502507524, + "loss": 1.3753, + "theoretical_loss": 3.4833880621862385, + "tokens_seen": 1837498368 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022388164493480442, + "loss": 1.4463, + "theoretical_loss": 3.4833776163733967, + "tokens_seen": 1837563904 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002238716148445336, + "loss": 1.3084, + "theoretical_loss": 3.4833671710374023, + "tokens_seen": 1837629440 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002238615847542628, + "loss": 1.2707, + "theoretical_loss": 3.4833567261782177, + "tokens_seen": 1837694976 + }, + { + "epoch": 0.56, + "learning_rate": 0.000223851554663992, + "loss": 1.291, + "theoretical_loss": 3.4833462817958027, + "tokens_seen": 1837760512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022384152457372117, + "loss": 1.3735, + "theoretical_loss": 3.4833358378901194, + "tokens_seen": 1837826048 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022383149448345035, + "loss": 1.2916, + "theoretical_loss": 3.4833253944611284, + "tokens_seen": 1837891584 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022382146439317953, + "loss": 1.2217, + "theoretical_loss": 3.483314951508792, + "tokens_seen": 1837957120 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022381143430290874, + "loss": 1.341, + "theoretical_loss": 3.4833045090330703, + "tokens_seen": 1838022656 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022380140421263792, + "loss": 1.4026, + "theoretical_loss": 3.483294067033926, + "tokens_seen": 1838088192 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002237913741223671, + "loss": 1.3581, + "theoretical_loss": 3.4832836255113184, + "tokens_seen": 1838153728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022378134403209629, + "loss": 1.3186, + "theoretical_loss": 3.48327318446521, + "tokens_seen": 1838219264 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 670495, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3910541534423828, + "objective/train/theoretical_loss": 3.483262743895562, + "objective/train/tokens_used": 1858744800, + "theoretical_loss": 3.483262743895562, + "tokens_seen": 1838284800 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002237713139418255, + "loss": 1.4146, + "theoretical_loss": 3.483262743895562, + "tokens_seen": 1838284800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022376128385155468, + "loss": 1.3065, + "theoretical_loss": 3.4832523038023355, + "tokens_seen": 1838350336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022375125376128386, + "loss": 1.3704, + "theoretical_loss": 3.483241864185492, + "tokens_seen": 1838415872 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022374122367101304, + "loss": 1.4217, + "theoretical_loss": 3.483231425044993, + "tokens_seen": 1838481408 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022373119358074222, + "loss": 1.3113, + "theoretical_loss": 3.4832209863807986, + "tokens_seen": 1838546944 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022372116349047143, + "loss": 1.3501, + "theoretical_loss": 3.4832105481928712, + "tokens_seen": 1838612480 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002237111334002006, + "loss": 1.4772, + "theoretical_loss": 3.483200110481172, + "tokens_seen": 1838678016 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002237011033099298, + "loss": 1.3894, + "theoretical_loss": 3.4831896732456613, + "tokens_seen": 1838743552 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022369107321965897, + "loss": 1.4191, + "theoretical_loss": 3.483179236486302, + "tokens_seen": 1838809088 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022368104312938818, + "loss": 1.4609, + "theoretical_loss": 3.483168800203054, + "tokens_seen": 1838874624 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022367101303911736, + "loss": 1.3748, + "theoretical_loss": 3.4831583643958797, + "tokens_seen": 1838940160 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022366098294884654, + "loss": 1.358, + "theoretical_loss": 3.4831479290647396, + "tokens_seen": 1839005696 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022365095285857572, + "loss": 1.2637, + "theoretical_loss": 3.483137494209595, + "tokens_seen": 1839071232 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002236409227683049, + "loss": 1.3677, + "theoretical_loss": 3.4831270598304087, + "tokens_seen": 1839136768 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002236308926780341, + "loss": 1.3463, + "theoretical_loss": 3.48311662592714, + "tokens_seen": 1839202304 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002236208625877633, + "loss": 1.2594, + "theoretical_loss": 3.4831061924997515, + "tokens_seen": 1839267840 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022361083249749247, + "loss": 1.3659, + "theoretical_loss": 3.4830957595482035, + "tokens_seen": 1839333376 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022360080240722165, + "loss": 1.3414, + "theoretical_loss": 3.483085327072459, + "tokens_seen": 1839398912 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022359077231695086, + "loss": 1.3478, + "theoretical_loss": 3.4830748950724777, + "tokens_seen": 1839464448 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022358074222668004, + "loss": 1.2457, + "theoretical_loss": 3.4830644635482217, + "tokens_seen": 1839529984 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022357071213640923, + "loss": 1.3227, + "theoretical_loss": 3.483054032499653, + "tokens_seen": 1839595520 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002235606820461384, + "loss": 1.3611, + "theoretical_loss": 3.483043601926731, + "tokens_seen": 1839661056 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022355065195586761, + "loss": 1.3596, + "theoretical_loss": 3.483033171829419, + "tokens_seen": 1839726592 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002235406218655968, + "loss": 1.3805, + "theoretical_loss": 3.4830227422076776, + "tokens_seen": 1839792128 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022353059177532598, + "loss": 1.3898, + "theoretical_loss": 3.4830123130614687, + "tokens_seen": 1839857664 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022352056168505516, + "loss": 1.3814, + "theoretical_loss": 3.4830018843907524, + "tokens_seen": 1839923200 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022351053159478434, + "loss": 1.3476, + "theoretical_loss": 3.4829914561954913, + "tokens_seen": 1839988736 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022350050150451355, + "loss": 1.3797, + "theoretical_loss": 3.4829810284756464, + "tokens_seen": 1840054272 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022349047141424273, + "loss": 1.4762, + "theoretical_loss": 3.4829706012311794, + "tokens_seen": 1840119808 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002234804413239719, + "loss": 1.3684, + "theoretical_loss": 3.4829601744620513, + "tokens_seen": 1840185344 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002234704112337011, + "loss": 1.2806, + "theoretical_loss": 3.4829497481682234, + "tokens_seen": 1840250880 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002234603811434303, + "loss": 1.3105, + "theoretical_loss": 3.482939322349657, + "tokens_seen": 1840316416 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022345035105315948, + "loss": 1.3386, + "theoretical_loss": 3.482928897006315, + "tokens_seen": 1840381952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022344032096288866, + "loss": 1.3245, + "theoretical_loss": 3.4829184721381568, + "tokens_seen": 1840447488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022343029087261784, + "loss": 1.2783, + "theoretical_loss": 3.482908047745145, + "tokens_seen": 1840513024 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022342026078234702, + "loss": 1.4295, + "theoretical_loss": 3.48289762382724, + "tokens_seen": 1840578560 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022341023069207623, + "loss": 1.3223, + "theoretical_loss": 3.4828872003844045, + "tokens_seen": 1840644096 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002234002006018054, + "loss": 1.339, + "theoretical_loss": 3.482876777416599, + "tokens_seen": 1840709632 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002233901705115346, + "loss": 1.4042, + "theoretical_loss": 3.482866354923786, + "tokens_seen": 1840775168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022338014042126377, + "loss": 1.2891, + "theoretical_loss": 3.4828559329059257, + "tokens_seen": 1840840704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022337011033099298, + "loss": 1.2508, + "theoretical_loss": 3.48284551136298, + "tokens_seen": 1840906240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022336008024072216, + "loss": 1.4436, + "theoretical_loss": 3.482835090294911, + "tokens_seen": 1840971776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022335005015045135, + "loss": 1.4002, + "theoretical_loss": 3.482824669701679, + "tokens_seen": 1841037312 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022334002006018055, + "loss": 1.4683, + "theoretical_loss": 3.4828142495832464, + "tokens_seen": 1841102848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022332998996990973, + "loss": 1.4074, + "theoretical_loss": 3.4828038299395745, + "tokens_seen": 1841168384 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022331995987963894, + "loss": 1.2817, + "theoretical_loss": 3.482793410770624, + "tokens_seen": 1841233920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022330992978936812, + "loss": 1.2408, + "theoretical_loss": 3.4827829920763573, + "tokens_seen": 1841299456 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002232998996990973, + "loss": 1.3291, + "theoretical_loss": 3.482772573856736, + "tokens_seen": 1841364992 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022328986960882649, + "loss": 1.4092, + "theoretical_loss": 3.4827621561117206, + "tokens_seen": 1841430528 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002232798395185557, + "loss": 1.3338, + "theoretical_loss": 3.4827517388412734, + "tokens_seen": 1841496064 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 671743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5256602764129639, + "objective/train/theoretical_loss": 3.4827413220453556, + "objective/train/tokens_used": 1862021600, + "theoretical_loss": 3.4827413220453556, + "tokens_seen": 1841561600 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022326980942828488, + "loss": 1.3457, + "theoretical_loss": 3.4827413220453556, + "tokens_seen": 1841561600 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022325977933801406, + "loss": 1.297, + "theoretical_loss": 3.482730905723929, + "tokens_seen": 1841627136 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022324974924774324, + "loss": 1.2606, + "theoretical_loss": 3.482720489876954, + "tokens_seen": 1841692672 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022323971915747242, + "loss": 1.2845, + "theoretical_loss": 3.4827100745043937, + "tokens_seen": 1841758208 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022322968906720163, + "loss": 1.2994, + "theoretical_loss": 3.4826996596062085, + "tokens_seen": 1841823744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002232196589769308, + "loss": 1.2661, + "theoretical_loss": 3.4826892451823603, + "tokens_seen": 1841889280 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022320962888666, + "loss": 1.3723, + "theoretical_loss": 3.4826788312328105, + "tokens_seen": 1841954816 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022319959879638917, + "loss": 1.4277, + "theoretical_loss": 3.4826684177575213, + "tokens_seen": 1842020352 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022318956870611838, + "loss": 1.3529, + "theoretical_loss": 3.482658004756453, + "tokens_seen": 1842085888 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022317953861584756, + "loss": 1.3385, + "theoretical_loss": 3.4826475922295677, + "tokens_seen": 1842151424 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022316950852557674, + "loss": 1.4126, + "theoretical_loss": 3.4826371801768277, + "tokens_seen": 1842216960 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022315947843530592, + "loss": 1.3311, + "theoretical_loss": 3.4826267685981933, + "tokens_seen": 1842282496 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022315947843530592, + "loss": 1.301, + "theoretical_loss": 3.482616357493627, + "tokens_seen": 1842348032 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002231494483450351, + "loss": 1.375, + "theoretical_loss": 3.4826059468630897, + "tokens_seen": 1842413568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002231394182547643, + "loss": 1.2813, + "theoretical_loss": 3.482595536706543, + "tokens_seen": 1842479104 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002231293881644935, + "loss": 1.4713, + "theoretical_loss": 3.482585127023949, + "tokens_seen": 1842544640 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022311935807422267, + "loss": 1.3354, + "theoretical_loss": 3.482574717815269, + "tokens_seen": 1842610176 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022310932798395185, + "loss": 1.5063, + "theoretical_loss": 3.4825643090804643, + "tokens_seen": 1842675712 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022309929789368106, + "loss": 1.3448, + "theoretical_loss": 3.4825539008194966, + "tokens_seen": 1842741248 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022308926780341024, + "loss": 1.2926, + "theoretical_loss": 3.4825434930323276, + "tokens_seen": 1842806784 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022307923771313943, + "loss": 1.3595, + "theoretical_loss": 3.482533085718919, + "tokens_seen": 1842872320 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002230692076228686, + "loss": 1.3615, + "theoretical_loss": 3.4825226788792323, + "tokens_seen": 1842937856 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022305917753259781, + "loss": 1.3114, + "theoretical_loss": 3.4825122725132287, + "tokens_seen": 1843003392 + }, + { + "epoch": 0.56, + "learning_rate": 0.000223049147442327, + "loss": 1.3551, + "theoretical_loss": 3.4825018666208702, + "tokens_seen": 1843068928 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022303911735205618, + "loss": 1.3816, + "theoretical_loss": 3.4824914612021187, + "tokens_seen": 1843134464 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022302908726178536, + "loss": 1.3005, + "theoretical_loss": 3.482481056256935, + "tokens_seen": 1843200000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022301905717151454, + "loss": 1.3484, + "theoretical_loss": 3.4824706517852815, + "tokens_seen": 1843265536 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022300902708124375, + "loss": 1.3052, + "theoretical_loss": 3.4824602477871194, + "tokens_seen": 1843331072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022299899699097293, + "loss": 1.3444, + "theoretical_loss": 3.48244984426241, + "tokens_seen": 1843396608 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002229889669007021, + "loss": 1.3838, + "theoretical_loss": 3.482439441211116, + "tokens_seen": 1843462144 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002229789368104313, + "loss": 1.2002, + "theoretical_loss": 3.4824290386331977, + "tokens_seen": 1843527680 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002229689067201605, + "loss": 1.2151, + "theoretical_loss": 3.4824186365286174, + "tokens_seen": 1843593216 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022295887662988968, + "loss": 1.3624, + "theoretical_loss": 3.4824082348973366, + "tokens_seen": 1843658752 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022294884653961886, + "loss": 1.3221, + "theoretical_loss": 3.4823978337393173, + "tokens_seen": 1843724288 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022293881644934804, + "loss": 1.2915, + "theoretical_loss": 3.4823874330545213, + "tokens_seen": 1843789824 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022292878635907722, + "loss": 1.3224, + "theoretical_loss": 3.4823770328429093, + "tokens_seen": 1843855360 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022291875626880643, + "loss": 1.302, + "theoretical_loss": 3.4823666331044434, + "tokens_seen": 1843920896 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002229087261785356, + "loss": 1.3484, + "theoretical_loss": 3.4823562338390857, + "tokens_seen": 1843986432 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002228986960882648, + "loss": 1.2601, + "theoretical_loss": 3.4823458350467975, + "tokens_seen": 1844051968 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022288866599799397, + "loss": 1.305, + "theoretical_loss": 3.4823354367275403, + "tokens_seen": 1844117504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022287863590772318, + "loss": 1.3279, + "theoretical_loss": 3.482325038881276, + "tokens_seen": 1844183040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022286860581745236, + "loss": 1.3427, + "theoretical_loss": 3.482314641507966, + "tokens_seen": 1844248576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022285857572718155, + "loss": 1.3679, + "theoretical_loss": 3.4823042446075725, + "tokens_seen": 1844314112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022284854563691073, + "loss": 1.2127, + "theoretical_loss": 3.4822938481800567, + "tokens_seen": 1844379648 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002228385155466399, + "loss": 1.3256, + "theoretical_loss": 3.4822834522253805, + "tokens_seen": 1844445184 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022282848545636912, + "loss": 1.1632, + "theoretical_loss": 3.482273056743506, + "tokens_seen": 1844510720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002228184553660983, + "loss": 1.2365, + "theoretical_loss": 3.482262661734394, + "tokens_seen": 1844576256 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022280842527582748, + "loss": 1.525, + "theoretical_loss": 3.4822522671980067, + "tokens_seen": 1844641792 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022279839518555666, + "loss": 1.352, + "theoretical_loss": 3.4822418731343063, + "tokens_seen": 1844707328 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022278836509528587, + "loss": 1.3146, + "theoretical_loss": 3.4822314795432536, + "tokens_seen": 1844772864 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 672951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3730549812316895, + "objective/train/theoretical_loss": 3.4822210864248113, + "objective/train/tokens_used": 1865298400, + "theoretical_loss": 3.4822210864248113, + "tokens_seen": 1844838400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022277833500501505, + "loss": 1.4368, + "theoretical_loss": 3.4822210864248113, + "tokens_seen": 1844838400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022276830491474423, + "loss": 1.3763, + "theoretical_loss": 3.4822106937789394, + "tokens_seen": 1844903936 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002227582748244734, + "loss": 1.3941, + "theoretical_loss": 3.4822003016056016, + "tokens_seen": 1844969472 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002227482447342026, + "loss": 1.3728, + "theoretical_loss": 3.4821899099047586, + "tokens_seen": 1845035008 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002227382146439318, + "loss": 1.2455, + "theoretical_loss": 3.4821795186763724, + "tokens_seen": 1845100544 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022272818455366098, + "loss": 1.1866, + "theoretical_loss": 3.482169127920405, + "tokens_seen": 1845166080 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022271815446339016, + "loss": 1.336, + "theoretical_loss": 3.4821587376368175, + "tokens_seen": 1845231616 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022270812437311934, + "loss": 1.3616, + "theoretical_loss": 3.482148347825572, + "tokens_seen": 1845297152 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022269809428284855, + "loss": 1.2346, + "theoretical_loss": 3.4821379584866303, + "tokens_seen": 1845362688 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022268806419257773, + "loss": 1.3622, + "theoretical_loss": 3.482127569619954, + "tokens_seen": 1845428224 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022267803410230691, + "loss": 1.2269, + "theoretical_loss": 3.482117181225505, + "tokens_seen": 1845493760 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002226680040120361, + "loss": 1.3656, + "theoretical_loss": 3.482106793303245, + "tokens_seen": 1845559296 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022265797392176528, + "loss": 1.4138, + "theoretical_loss": 3.4820964058531354, + "tokens_seen": 1845624832 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022264794383149448, + "loss": 1.4205, + "theoretical_loss": 3.482086018875139, + "tokens_seen": 1845690368 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022263791374122367, + "loss": 1.395, + "theoretical_loss": 3.4820756323692166, + "tokens_seen": 1845755904 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022262788365095285, + "loss": 1.3979, + "theoretical_loss": 3.4820652463353303, + "tokens_seen": 1845821440 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022261785356068203, + "loss": 1.3754, + "theoretical_loss": 3.4820548607734425, + "tokens_seen": 1845886976 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022260782347041124, + "loss": 1.2478, + "theoretical_loss": 3.4820444756835136, + "tokens_seen": 1845952512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022259779338014042, + "loss": 1.453, + "theoretical_loss": 3.4820340910655068, + "tokens_seen": 1846018048 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022258776328986963, + "loss": 1.2586, + "theoretical_loss": 3.482023706919383, + "tokens_seen": 1846083584 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002225777331995988, + "loss": 1.375, + "theoretical_loss": 3.482013323245104, + "tokens_seen": 1846149120 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022256770310932801, + "loss": 1.3887, + "theoretical_loss": 3.4820029400426327, + "tokens_seen": 1846214656 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002225576730190572, + "loss": 1.3022, + "theoretical_loss": 3.481992557311929, + "tokens_seen": 1846280192 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022254764292878638, + "loss": 1.2425, + "theoretical_loss": 3.481982175052957, + "tokens_seen": 1846345728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022253761283851556, + "loss": 1.3397, + "theoretical_loss": 3.481971793265677, + "tokens_seen": 1846411264 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022252758274824474, + "loss": 1.4936, + "theoretical_loss": 3.481961411950051, + "tokens_seen": 1846476800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022251755265797395, + "loss": 1.3189, + "theoretical_loss": 3.4819510311060418, + "tokens_seen": 1846542336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022250752256770313, + "loss": 1.2985, + "theoretical_loss": 3.4819406507336095, + "tokens_seen": 1846607872 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002224974924774323, + "loss": 1.297, + "theoretical_loss": 3.4819302708327173, + "tokens_seen": 1846673408 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002224874623871615, + "loss": 1.3413, + "theoretical_loss": 3.4819198914033267, + "tokens_seen": 1846738944 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002224774322968907, + "loss": 1.3116, + "theoretical_loss": 3.4819095124454, + "tokens_seen": 1846804480 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022246740220661988, + "loss": 1.4217, + "theoretical_loss": 3.4818991339588976, + "tokens_seen": 1846870016 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022245737211634906, + "loss": 1.3211, + "theoretical_loss": 3.4818887559437828, + "tokens_seen": 1846935552 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022244734202607824, + "loss": 1.2982, + "theoretical_loss": 3.481878378400017, + "tokens_seen": 1847001088 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022243731193580742, + "loss": 1.3322, + "theoretical_loss": 3.4818680013275625, + "tokens_seen": 1847066624 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022242728184553663, + "loss": 1.367, + "theoretical_loss": 3.4818576247263806, + "tokens_seen": 1847132160 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002224172517552658, + "loss": 1.3845, + "theoretical_loss": 3.481847248596433, + "tokens_seen": 1847197696 + }, + { + "epoch": 0.56, + "learning_rate": 0.000222407221664995, + "loss": 1.2745, + "theoretical_loss": 3.4818368729376816, + "tokens_seen": 1847263232 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022239719157472418, + "loss": 1.3652, + "theoretical_loss": 3.4818264977500895, + "tokens_seen": 1847328768 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022238716148445338, + "loss": 1.2583, + "theoretical_loss": 3.4818161230336173, + "tokens_seen": 1847394304 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022237713139418256, + "loss": 1.2624, + "theoretical_loss": 3.481805748788227, + "tokens_seen": 1847459840 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022236710130391175, + "loss": 1.3603, + "theoretical_loss": 3.4817953750138813, + "tokens_seen": 1847525376 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022235707121364093, + "loss": 1.3528, + "theoretical_loss": 3.481785001710541, + "tokens_seen": 1847590912 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002223470411233701, + "loss": 1.2316, + "theoretical_loss": 3.481774628878169, + "tokens_seen": 1847656448 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022233701103309932, + "loss": 1.3251, + "theoretical_loss": 3.481764256516726, + "tokens_seen": 1847721984 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002223269809428285, + "loss": 1.3311, + "theoretical_loss": 3.4817538846261757, + "tokens_seen": 1847787520 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022231695085255768, + "loss": 1.2159, + "theoretical_loss": 3.481743513206479, + "tokens_seen": 1847853056 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022230692076228686, + "loss": 1.4419, + "theoretical_loss": 3.4817331422575974, + "tokens_seen": 1847918592 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022229689067201607, + "loss": 1.4017, + "theoretical_loss": 3.4817227717794936, + "tokens_seen": 1847984128 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022228686058174525, + "loss": 1.3606, + "theoretical_loss": 3.4817124017721293, + "tokens_seen": 1848049664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 673981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3688905239105225, + "objective/train/theoretical_loss": 3.481702032235466, + "objective/train/tokens_used": 1868575200, + "theoretical_loss": 3.481702032235466, + "tokens_seen": 1848115200 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022227683049147443, + "loss": 1.3722, + "theoretical_loss": 3.481702032235466, + "tokens_seen": 1848115200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002222668004012036, + "loss": 1.3577, + "theoretical_loss": 3.4816916631694665, + "tokens_seen": 1848180736 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002222567703109328, + "loss": 1.4362, + "theoretical_loss": 3.481681294574092, + "tokens_seen": 1848246272 + }, + { + "epoch": 0.56, + "learning_rate": 0.000222246740220662, + "loss": 1.3527, + "theoretical_loss": 3.4816709264493046, + "tokens_seen": 1848311808 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022223671013039118, + "loss": 1.4322, + "theoretical_loss": 3.481660558795067, + "tokens_seen": 1848377344 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022222668004012036, + "loss": 1.5237, + "theoretical_loss": 3.4816501916113394, + "tokens_seen": 1848442880 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022221664994984954, + "loss": 1.4134, + "theoretical_loss": 3.4816398248980858, + "tokens_seen": 1848508416 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022220661985957875, + "loss": 1.3024, + "theoretical_loss": 3.481629458655267, + "tokens_seen": 1848573952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022219658976930793, + "loss": 1.3058, + "theoretical_loss": 3.4816190928828457, + "tokens_seen": 1848639488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022218655967903711, + "loss": 1.4269, + "theoretical_loss": 3.4816087275807828, + "tokens_seen": 1848705024 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002221765295887663, + "loss": 1.3743, + "theoretical_loss": 3.481598362749041, + "tokens_seen": 1848770560 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022216649949849548, + "loss": 1.3961, + "theoretical_loss": 3.4815879983875826, + "tokens_seen": 1848836096 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022215646940822468, + "loss": 1.2825, + "theoretical_loss": 3.481577634496369, + "tokens_seen": 1848901632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022214643931795387, + "loss": 1.4809, + "theoretical_loss": 3.4815672710753622, + "tokens_seen": 1848967168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022213640922768305, + "loss": 1.399, + "theoretical_loss": 3.4815569081245243, + "tokens_seen": 1849032704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022212637913741223, + "loss": 1.4454, + "theoretical_loss": 3.481546545643818, + "tokens_seen": 1849098240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022211634904714144, + "loss": 1.3799, + "theoretical_loss": 3.4815361836332044, + "tokens_seen": 1849163776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022210631895687062, + "loss": 1.4416, + "theoretical_loss": 3.481525822092646, + "tokens_seen": 1849229312 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002220962888665998, + "loss": 1.3977, + "theoretical_loss": 3.481515461022104, + "tokens_seen": 1849294848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022208625877632898, + "loss": 1.3089, + "theoretical_loss": 3.4815051004215416, + "tokens_seen": 1849360384 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022207622868605816, + "loss": 1.3675, + "theoretical_loss": 3.4814947402909207, + "tokens_seen": 1849425920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022206619859578737, + "loss": 1.3462, + "theoretical_loss": 3.481484380630202, + "tokens_seen": 1849491456 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022205616850551655, + "loss": 1.4544, + "theoretical_loss": 3.4814740214393494, + "tokens_seen": 1849556992 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022204613841524573, + "loss": 1.3641, + "theoretical_loss": 3.4814636627183235, + "tokens_seen": 1849622528 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002220361083249749, + "loss": 1.4117, + "theoretical_loss": 3.4814533044670872, + "tokens_seen": 1849688064 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022202607823470412, + "loss": 1.2912, + "theoretical_loss": 3.481442946685602, + "tokens_seen": 1849753600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002220160481444333, + "loss": 1.4019, + "theoretical_loss": 3.4814325893738296, + "tokens_seen": 1849819136 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022200601805416248, + "loss": 1.3212, + "theoretical_loss": 3.481422232531733, + "tokens_seen": 1849884672 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022199598796389166, + "loss": 1.3798, + "theoretical_loss": 3.4814118761592745, + "tokens_seen": 1849950208 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022198595787362084, + "loss": 1.4432, + "theoretical_loss": 3.481401520256415, + "tokens_seen": 1850015744 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022197592778335005, + "loss": 1.41, + "theoretical_loss": 3.4813911648231173, + "tokens_seen": 1850081280 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022196589769307923, + "loss": 1.2012, + "theoretical_loss": 3.481380809859343, + "tokens_seen": 1850146816 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022195586760280842, + "loss": 1.3986, + "theoretical_loss": 3.4813704553650546, + "tokens_seen": 1850212352 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002219458375125376, + "loss": 1.2474, + "theoretical_loss": 3.481360101340214, + "tokens_seen": 1850277888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002219358074222668, + "loss": 1.377, + "theoretical_loss": 3.4813497477847837, + "tokens_seen": 1850343424 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022192577733199599, + "loss": 1.3247, + "theoretical_loss": 3.4813393946987254, + "tokens_seen": 1850408960 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022191574724172517, + "loss": 1.3273, + "theoretical_loss": 3.481329042082001, + "tokens_seen": 1850474496 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022190571715145435, + "loss": 1.3736, + "theoretical_loss": 3.481318689934573, + "tokens_seen": 1850540032 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022189568706118356, + "loss": 1.3267, + "theoretical_loss": 3.481308338256403, + "tokens_seen": 1850605568 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022188565697091274, + "loss": 1.4422, + "theoretical_loss": 3.481297987047454, + "tokens_seen": 1850671104 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022187562688064192, + "loss": 1.4593, + "theoretical_loss": 3.4812876363076874, + "tokens_seen": 1850736640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002218655967903711, + "loss": 1.335, + "theoretical_loss": 3.4812772860370655, + "tokens_seen": 1850802176 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022185556670010028, + "loss": 1.4451, + "theoretical_loss": 3.4812669362355506, + "tokens_seen": 1850867712 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022184553660982952, + "loss": 1.4264, + "theoretical_loss": 3.4812565869031045, + "tokens_seen": 1850933248 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002218355065195587, + "loss": 1.426, + "theoretical_loss": 3.48124623803969, + "tokens_seen": 1850998784 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022182547642928788, + "loss": 1.2945, + "theoretical_loss": 3.481235889645268, + "tokens_seen": 1851064320 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022181544633901706, + "loss": 1.405, + "theoretical_loss": 3.4812255417198017, + "tokens_seen": 1851129856 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022180541624874627, + "loss": 1.3175, + "theoretical_loss": 3.4812151942632523, + "tokens_seen": 1851195392 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022179538615847545, + "loss": 1.285, + "theoretical_loss": 3.4812048472755834, + "tokens_seen": 1851260928 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022178535606820463, + "loss": 1.238, + "theoretical_loss": 3.4811945007567564, + "tokens_seen": 1851326464 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 675199, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4188995361328125, + "objective/train/theoretical_loss": 3.4811841547067326, + "objective/train/tokens_used": 1871852000, + "theoretical_loss": 3.4811841547067326, + "tokens_seen": 1851392000 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002217753259779338, + "loss": 1.3093, + "theoretical_loss": 3.4811841547067326, + "tokens_seen": 1851392000 + }, + { + "epoch": 0.56, + "learning_rate": 0.000221765295887663, + "loss": 1.4321, + "theoretical_loss": 3.4811738091254756, + "tokens_seen": 1851457536 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002217552657973922, + "loss": 1.3088, + "theoretical_loss": 3.4811634640129467, + "tokens_seen": 1851523072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022174523570712138, + "loss": 1.4089, + "theoretical_loss": 3.4811531193691083, + "tokens_seen": 1851588608 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022173520561685056, + "loss": 1.3841, + "theoretical_loss": 3.4811427751939226, + "tokens_seen": 1851654144 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022172517552657974, + "loss": 1.4077, + "theoretical_loss": 3.481132431487352, + "tokens_seen": 1851719680 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022171514543630895, + "loss": 1.2288, + "theoretical_loss": 3.481122088249358, + "tokens_seen": 1851785216 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022170511534603813, + "loss": 1.2653, + "theoretical_loss": 3.481111745479904, + "tokens_seen": 1851850752 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022169508525576731, + "loss": 1.3308, + "theoretical_loss": 3.4811014031789504, + "tokens_seen": 1851916288 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002216850551654965, + "loss": 1.32, + "theoretical_loss": 3.4810910613464614, + "tokens_seen": 1851981824 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022167502507522568, + "loss": 1.3951, + "theoretical_loss": 3.4810807199823977, + "tokens_seen": 1852047360 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022166499498495488, + "loss": 1.3146, + "theoretical_loss": 3.481070379086722, + "tokens_seen": 1852112896 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022165496489468407, + "loss": 1.4111, + "theoretical_loss": 3.481060038659397, + "tokens_seen": 1852178432 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022164493480441325, + "loss": 1.4257, + "theoretical_loss": 3.4810496987003843, + "tokens_seen": 1852243968 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022163490471414243, + "loss": 1.1897, + "theoretical_loss": 3.481039359209646, + "tokens_seen": 1852309504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022162487462387164, + "loss": 1.3702, + "theoretical_loss": 3.4810290201871448, + "tokens_seen": 1852375040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022161484453360082, + "loss": 1.3472, + "theoretical_loss": 3.481018681632843, + "tokens_seen": 1852440576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022160481444333, + "loss": 1.3815, + "theoretical_loss": 3.4810083435467023, + "tokens_seen": 1852506112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022159478435305918, + "loss": 1.3307, + "theoretical_loss": 3.480998005928685, + "tokens_seen": 1852571648 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022158475426278836, + "loss": 1.3186, + "theoretical_loss": 3.480987668778754, + "tokens_seen": 1852637184 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022157472417251757, + "loss": 1.2663, + "theoretical_loss": 3.480977332096871, + "tokens_seen": 1852702720 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022156469408224675, + "loss": 1.4057, + "theoretical_loss": 3.4809669958829983, + "tokens_seen": 1852768256 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022155466399197593, + "loss": 1.3716, + "theoretical_loss": 3.4809566601370983, + "tokens_seen": 1852833792 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002215446339017051, + "loss": 1.3496, + "theoretical_loss": 3.4809463248591332, + "tokens_seen": 1852899328 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022153460381143432, + "loss": 1.3285, + "theoretical_loss": 3.4809359900490655, + "tokens_seen": 1852964864 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002215245737211635, + "loss": 1.4245, + "theoretical_loss": 3.480925655706857, + "tokens_seen": 1853030400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022151454363089268, + "loss": 1.3054, + "theoretical_loss": 3.48091532183247, + "tokens_seen": 1853095936 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022150451354062186, + "loss": 1.2441, + "theoretical_loss": 3.480904988425867, + "tokens_seen": 1853161472 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022149448345035105, + "loss": 1.3083, + "theoretical_loss": 3.4808946554870106, + "tokens_seen": 1853227008 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022148445336008025, + "loss": 1.2639, + "theoretical_loss": 3.4808843230158626, + "tokens_seen": 1853292544 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022147442326980943, + "loss": 1.3611, + "theoretical_loss": 3.480873991012385, + "tokens_seen": 1853358080 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022146439317953862, + "loss": 1.2681, + "theoretical_loss": 3.480863659476541, + "tokens_seen": 1853423616 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002214543630892678, + "loss": 1.4899, + "theoretical_loss": 3.4808533284082928, + "tokens_seen": 1853489152 + }, + { + "epoch": 0.56, + "learning_rate": 0.000221444332998997, + "loss": 1.3705, + "theoretical_loss": 3.480842997807602, + "tokens_seen": 1853554688 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022143430290872619, + "loss": 1.3297, + "theoretical_loss": 3.480832667674431, + "tokens_seen": 1853620224 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022142427281845537, + "loss": 1.3477, + "theoretical_loss": 3.4808223380087426, + "tokens_seen": 1853685760 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022141424272818455, + "loss": 1.2849, + "theoretical_loss": 3.480812008810499, + "tokens_seen": 1853751296 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022140421263791376, + "loss": 1.3639, + "theoretical_loss": 3.4808016800796624, + "tokens_seen": 1853816832 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022139418254764294, + "loss": 1.2417, + "theoretical_loss": 3.480791351816195, + "tokens_seen": 1853882368 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022138415245737212, + "loss": 1.3454, + "theoretical_loss": 3.480781024020059, + "tokens_seen": 1853947904 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002213741223671013, + "loss": 1.3166, + "theoretical_loss": 3.4807706966912173, + "tokens_seen": 1854013440 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022136409227683048, + "loss": 1.4, + "theoretical_loss": 3.480760369829632, + "tokens_seen": 1854078976 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002213540621865597, + "loss": 1.3968, + "theoretical_loss": 3.4807500434352647, + "tokens_seen": 1854144512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022134403209628887, + "loss": 1.3191, + "theoretical_loss": 3.4807397175080794, + "tokens_seen": 1854210048 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022133400200601805, + "loss": 1.3886, + "theoretical_loss": 3.480729392048037, + "tokens_seen": 1854275584 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022132397191574723, + "loss": 1.3296, + "theoretical_loss": 3.4807190670551007, + "tokens_seen": 1854341120 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022131394182547644, + "loss": 1.4024, + "theoretical_loss": 3.480708742529232, + "tokens_seen": 1854406656 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022130391173520562, + "loss": 1.4278, + "theoretical_loss": 3.480698418470394, + "tokens_seen": 1854472192 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002212938816449348, + "loss": 1.3492, + "theoretical_loss": 3.480688094878549, + "tokens_seen": 1854537728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022128385155466398, + "loss": 1.3778, + "theoretical_loss": 3.480677771753659, + "tokens_seen": 1854603264 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 676483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4642289876937866, + "objective/train/theoretical_loss": 3.4806674490956864, + "objective/train/tokens_used": 1875128800, + "theoretical_loss": 3.4806674490956864, + "tokens_seen": 1854668800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022127382146439317, + "loss": 1.5075, + "theoretical_loss": 3.4806674490956864, + "tokens_seen": 1854668800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022126379137412237, + "loss": 1.2927, + "theoretical_loss": 3.4806571269045943, + "tokens_seen": 1854734336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022125376128385155, + "loss": 1.4053, + "theoretical_loss": 3.4806468051803443, + "tokens_seen": 1854799872 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022124373119358074, + "loss": 1.3649, + "theoretical_loss": 3.480636483922899, + "tokens_seen": 1854865408 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022123370110330992, + "loss": 1.3494, + "theoretical_loss": 3.480626163132221, + "tokens_seen": 1854930944 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022122367101303913, + "loss": 1.3422, + "theoretical_loss": 3.4806158428082723, + "tokens_seen": 1854996480 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002212136409227683, + "loss": 1.3232, + "theoretical_loss": 3.480605522951016, + "tokens_seen": 1855062016 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002212036108324975, + "loss": 1.4269, + "theoretical_loss": 3.480595203560414, + "tokens_seen": 1855127552 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022119358074222667, + "loss": 1.3477, + "theoretical_loss": 3.4805848846364285, + "tokens_seen": 1855193088 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022118355065195585, + "loss": 1.2927, + "theoretical_loss": 3.480574566179022, + "tokens_seen": 1855258624 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022117352056168506, + "loss": 1.3076, + "theoretical_loss": 3.4805642481881574, + "tokens_seen": 1855324160 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022116349047141424, + "loss": 1.279, + "theoretical_loss": 3.480553930663797, + "tokens_seen": 1855389696 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022115346038114342, + "loss": 1.3816, + "theoretical_loss": 3.480543613605903, + "tokens_seen": 1855455232 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002211434302908726, + "loss": 1.3909, + "theoretical_loss": 3.4805332970144383, + "tokens_seen": 1855520768 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002211334002006018, + "loss": 1.2702, + "theoretical_loss": 3.480522980889364, + "tokens_seen": 1855586304 + }, + { + "epoch": 0.56, + "learning_rate": 0.000221123370110331, + "loss": 1.303, + "theoretical_loss": 3.4805126652306444, + "tokens_seen": 1855651840 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022111334002006017, + "loss": 1.286, + "theoretical_loss": 3.4805023500382406, + "tokens_seen": 1855717376 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022110330992978935, + "loss": 1.3008, + "theoretical_loss": 3.480492035312116, + "tokens_seen": 1855782912 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022109327983951856, + "loss": 1.3275, + "theoretical_loss": 3.480481721052232, + "tokens_seen": 1855848448 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022108324974924777, + "loss": 1.4283, + "theoretical_loss": 3.480471407258552, + "tokens_seen": 1855913984 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022107321965897695, + "loss": 1.3669, + "theoretical_loss": 3.480461093931037, + "tokens_seen": 1855979520 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022106318956870613, + "loss": 1.4723, + "theoretical_loss": 3.480450781069652, + "tokens_seen": 1856045056 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002210531594784353, + "loss": 1.3443, + "theoretical_loss": 3.4804404686743577, + "tokens_seen": 1856110592 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022104312938816452, + "loss": 1.2934, + "theoretical_loss": 3.4804301567451166, + "tokens_seen": 1856176128 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002210330992978937, + "loss": 1.2758, + "theoretical_loss": 3.4804198452818915, + "tokens_seen": 1856241664 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022102306920762288, + "loss": 1.2379, + "theoretical_loss": 3.480409534284645, + "tokens_seen": 1856307200 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022101303911735206, + "loss": 1.3296, + "theoretical_loss": 3.4803992237533397, + "tokens_seen": 1856372736 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022100300902708125, + "loss": 1.2802, + "theoretical_loss": 3.4803889136879373, + "tokens_seen": 1856438272 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022099297893681045, + "loss": 1.2974, + "theoretical_loss": 3.4803786040884015, + "tokens_seen": 1856503808 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022098294884653963, + "loss": 1.383, + "theoretical_loss": 3.480368294954694, + "tokens_seen": 1856569344 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022097291875626882, + "loss": 1.3776, + "theoretical_loss": 3.4803579862867773, + "tokens_seen": 1856634880 + }, + { + "epoch": 0.56, + "learning_rate": 0.000220962888665998, + "loss": 1.3644, + "theoretical_loss": 3.4803476780846143, + "tokens_seen": 1856700416 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002209528585757272, + "loss": 1.4472, + "theoretical_loss": 3.4803373703481673, + "tokens_seen": 1856765952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022094282848545639, + "loss": 1.3058, + "theoretical_loss": 3.4803270630773984, + "tokens_seen": 1856831488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022093279839518557, + "loss": 1.2731, + "theoretical_loss": 3.480316756272271, + "tokens_seen": 1856897024 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022092276830491475, + "loss": 1.3709, + "theoretical_loss": 3.4803064499327476, + "tokens_seen": 1856962560 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022091273821464396, + "loss": 1.346, + "theoretical_loss": 3.48029614405879, + "tokens_seen": 1857028096 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022090270812437314, + "loss": 1.4103, + "theoretical_loss": 3.480285838650361, + "tokens_seen": 1857093632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022089267803410232, + "loss": 1.3801, + "theoretical_loss": 3.480275533707423, + "tokens_seen": 1857159168 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002208826479438315, + "loss": 1.2121, + "theoretical_loss": 3.480265229229939, + "tokens_seen": 1857224704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022087261785356068, + "loss": 1.3502, + "theoretical_loss": 3.4802549252178716, + "tokens_seen": 1857290240 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002208625877632899, + "loss": 1.4158, + "theoretical_loss": 3.4802446216711824, + "tokens_seen": 1857355776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022085255767301907, + "loss": 1.3485, + "theoretical_loss": 3.4802343185898352, + "tokens_seen": 1857421312 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022084252758274825, + "loss": 1.3491, + "theoretical_loss": 3.480224015973792, + "tokens_seen": 1857486848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022083249749247743, + "loss": 1.3991, + "theoretical_loss": 3.480213713823015, + "tokens_seen": 1857552384 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022082246740220664, + "loss": 1.3623, + "theoretical_loss": 3.4802034121374676, + "tokens_seen": 1857617920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022081243731193582, + "loss": 1.2529, + "theoretical_loss": 3.480193110917112, + "tokens_seen": 1857683456 + }, + { + "epoch": 0.56, + "learning_rate": 0.000220802407221665, + "loss": 1.3887, + "theoretical_loss": 3.48018281016191, + "tokens_seen": 1857748992 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022079237713139418, + "loss": 1.3827, + "theoretical_loss": 3.4801725098718257, + "tokens_seen": 1857814528 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022078234704112337, + "loss": 1.2454, + "theoretical_loss": 3.480162210046821, + "tokens_seen": 1857880064 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 677697, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2380220890045166, + "objective/train/theoretical_loss": 3.4801519106868577, + "objective/train/tokens_used": 1878405600, + "theoretical_loss": 3.4801519106868577, + "tokens_seen": 1857945600 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022077231695085257, + "loss": 1.3122, + "theoretical_loss": 3.4801519106868577, + "tokens_seen": 1857945600 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022076228686058175, + "loss": 1.2784, + "theoretical_loss": 3.480141611791899, + "tokens_seen": 1858011136 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022075225677031094, + "loss": 1.3203, + "theoretical_loss": 3.480131313361908, + "tokens_seen": 1858076672 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022074222668004012, + "loss": 1.3498, + "theoretical_loss": 3.480121015396847, + "tokens_seen": 1858142208 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022073219658976933, + "loss": 1.3586, + "theoretical_loss": 3.480110717896679, + "tokens_seen": 1858207744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002207221664994985, + "loss": 1.3746, + "theoretical_loss": 3.480100420861365, + "tokens_seen": 1858273280 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002207121364092277, + "loss": 1.3446, + "theoretical_loss": 3.4800901242908693, + "tokens_seen": 1858338816 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022070210631895687, + "loss": 1.3009, + "theoretical_loss": 3.480079828185154, + "tokens_seen": 1858404352 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022069207622868605, + "loss": 1.3045, + "theoretical_loss": 3.480069532544182, + "tokens_seen": 1858469888 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022068204613841526, + "loss": 1.2516, + "theoretical_loss": 3.4800592373679153, + "tokens_seen": 1858535424 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022067201604814444, + "loss": 1.2878, + "theoretical_loss": 3.4800489426563166, + "tokens_seen": 1858600960 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022066198595787362, + "loss": 1.4221, + "theoretical_loss": 3.4800386484093497, + "tokens_seen": 1858666496 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002206519558676028, + "loss": 1.3836, + "theoretical_loss": 3.4800283546269757, + "tokens_seen": 1858732032 + }, + { + "epoch": 0.56, + "learning_rate": 0.000220641925777332, + "loss": 1.3627, + "theoretical_loss": 3.480018061309158, + "tokens_seen": 1858797568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002206318956870612, + "loss": 1.3806, + "theoretical_loss": 3.4800077684558595, + "tokens_seen": 1858863104 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022062186559679037, + "loss": 1.2513, + "theoretical_loss": 3.4799974760670427, + "tokens_seen": 1858928640 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022061183550651955, + "loss": 1.2726, + "theoretical_loss": 3.4799871841426695, + "tokens_seen": 1858994176 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022060180541624873, + "loss": 1.2705, + "theoretical_loss": 3.4799768926827035, + "tokens_seen": 1859059712 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022059177532597794, + "loss": 1.2513, + "theoretical_loss": 3.4799666016871074, + "tokens_seen": 1859125248 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022058174523570712, + "loss": 1.3278, + "theoretical_loss": 3.479956311155843, + "tokens_seen": 1859190784 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002205717151454363, + "loss": 1.3339, + "theoretical_loss": 3.4799460210888737, + "tokens_seen": 1859256320 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022056168505516549, + "loss": 1.2978, + "theoretical_loss": 3.4799357314861625, + "tokens_seen": 1859321856 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002205516549648947, + "loss": 1.3184, + "theoretical_loss": 3.479925442347671, + "tokens_seen": 1859387392 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022054162487462387, + "loss": 1.317, + "theoretical_loss": 3.479915153673363, + "tokens_seen": 1859452928 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022053159478435306, + "loss": 1.2364, + "theoretical_loss": 3.4799048654632, + "tokens_seen": 1859518464 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022052156469408224, + "loss": 1.3608, + "theoretical_loss": 3.479894577717146, + "tokens_seen": 1859584000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022051153460381142, + "loss": 1.2756, + "theoretical_loss": 3.479884290435163, + "tokens_seen": 1859649536 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022050150451354063, + "loss": 1.3242, + "theoretical_loss": 3.4798740036172138, + "tokens_seen": 1859715072 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002204914744232698, + "loss": 1.4631, + "theoretical_loss": 3.479863717263261, + "tokens_seen": 1859780608 + }, + { + "epoch": 0.56, + "learning_rate": 0.000220481444332999, + "loss": 1.259, + "theoretical_loss": 3.4798534313732676, + "tokens_seen": 1859846144 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022047141424272817, + "loss": 1.461, + "theoretical_loss": 3.4798431459471963, + "tokens_seen": 1859911680 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022046138415245738, + "loss": 1.2943, + "theoretical_loss": 3.4798328609850095, + "tokens_seen": 1859977216 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022045135406218656, + "loss": 1.3124, + "theoretical_loss": 3.4798225764866704, + "tokens_seen": 1860042752 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022044132397191574, + "loss": 1.2283, + "theoretical_loss": 3.4798122924521415, + "tokens_seen": 1860108288 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022043129388164492, + "loss": 1.332, + "theoretical_loss": 3.4798020088813852, + "tokens_seen": 1860173824 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002204212637913741, + "loss": 1.37, + "theoretical_loss": 3.4797917257743647, + "tokens_seen": 1860239360 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002204112337011033, + "loss": 1.37, + "theoretical_loss": 3.479781443131043, + "tokens_seen": 1860304896 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002204012036108325, + "loss": 1.3203, + "theoretical_loss": 3.479771160951382, + "tokens_seen": 1860370432 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022039117352056167, + "loss": 1.4164, + "theoretical_loss": 3.479760879235345, + "tokens_seen": 1860435968 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022038114343029085, + "loss": 1.3054, + "theoretical_loss": 3.4797505979828953, + "tokens_seen": 1860501504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022037111334002006, + "loss": 1.4755, + "theoretical_loss": 3.4797403171939942, + "tokens_seen": 1860567040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022036108324974924, + "loss": 1.3489, + "theoretical_loss": 3.479730036868606, + "tokens_seen": 1860632576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022035105315947842, + "loss": 1.3032, + "theoretical_loss": 3.479719757006693, + "tokens_seen": 1860698112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022034102306920763, + "loss": 1.329, + "theoretical_loss": 3.4797094776082176, + "tokens_seen": 1860763648 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022033099297893684, + "loss": 1.3335, + "theoretical_loss": 3.4796991986731425, + "tokens_seen": 1860829184 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022032096288866602, + "loss": 1.3043, + "theoretical_loss": 3.479688920201431, + "tokens_seen": 1860894720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002203109327983952, + "loss": 1.2992, + "theoretical_loss": 3.4796786421930457, + "tokens_seen": 1860960256 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022030090270812438, + "loss": 1.4145, + "theoretical_loss": 3.4796683646479494, + "tokens_seen": 1861025792 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022029087261785357, + "loss": 1.4332, + "theoretical_loss": 3.479658087566105, + "tokens_seen": 1861091328 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022028084252758277, + "loss": 1.3187, + "theoretical_loss": 3.4796478109474753, + "tokens_seen": 1861156864 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 678968, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3703534603118896, + "objective/train/theoretical_loss": 3.4796375347920225, + "objective/train/tokens_used": 1881682400, + "theoretical_loss": 3.4796375347920225, + "tokens_seen": 1861222400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022027081243731195, + "loss": 1.3287, + "theoretical_loss": 3.4796375347920225, + "tokens_seen": 1861222400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022026078234704114, + "loss": 1.3734, + "theoretical_loss": 3.4796272590997104, + "tokens_seen": 1861287936 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022025075225677032, + "loss": 1.2615, + "theoretical_loss": 3.4796169838705016, + "tokens_seen": 1861353472 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022024072216649953, + "loss": 1.3587, + "theoretical_loss": 3.479606709104358, + "tokens_seen": 1861419008 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002202306920762287, + "loss": 1.2962, + "theoretical_loss": 3.4795964348012434, + "tokens_seen": 1861484544 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002202206619859579, + "loss": 1.3196, + "theoretical_loss": 3.4795861609611203, + "tokens_seen": 1861550080 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022021063189568707, + "loss": 1.3574, + "theoretical_loss": 3.4795758875839518, + "tokens_seen": 1861615616 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022020060180541625, + "loss": 1.3534, + "theoretical_loss": 3.4795656146697, + "tokens_seen": 1861681152 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022019057171514546, + "loss": 1.37, + "theoretical_loss": 3.4795553422183287, + "tokens_seen": 1861746688 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022018054162487464, + "loss": 1.4341, + "theoretical_loss": 3.4795450702298005, + "tokens_seen": 1861812224 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022017051153460382, + "loss": 1.4614, + "theoretical_loss": 3.4795347987040777, + "tokens_seen": 1861877760 + }, + { + "epoch": 0.56, + "learning_rate": 0.000220160481444333, + "loss": 1.3343, + "theoretical_loss": 3.4795245276411233, + "tokens_seen": 1861943296 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002201504513540622, + "loss": 1.3883, + "theoretical_loss": 3.4795142570409006, + "tokens_seen": 1862008832 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002201404212637914, + "loss": 1.2752, + "theoretical_loss": 3.4795039869033726, + "tokens_seen": 1862074368 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022013039117352057, + "loss": 1.2052, + "theoretical_loss": 3.4794937172285016, + "tokens_seen": 1862139904 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022012036108324975, + "loss": 1.3326, + "theoretical_loss": 3.479483448016251, + "tokens_seen": 1862205440 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022011033099297893, + "loss": 1.3586, + "theoretical_loss": 3.479473179266583, + "tokens_seen": 1862270976 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022010030090270814, + "loss": 1.3722, + "theoretical_loss": 3.4794629109794606, + "tokens_seen": 1862336512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022009027081243732, + "loss": 1.3401, + "theoretical_loss": 3.479452643154848, + "tokens_seen": 1862402048 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002200802407221665, + "loss": 1.3738, + "theoretical_loss": 3.479442375792706, + "tokens_seen": 1862467584 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022007021063189569, + "loss": 1.3759, + "theoretical_loss": 3.4794321088929987, + "tokens_seen": 1862533120 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002200601805416249, + "loss": 1.3012, + "theoretical_loss": 3.4794218424556895, + "tokens_seen": 1862598656 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022005015045135408, + "loss": 1.3, + "theoretical_loss": 3.4794115764807403, + "tokens_seen": 1862664192 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022004012036108326, + "loss": 1.3533, + "theoretical_loss": 3.479401310968114, + "tokens_seen": 1862729728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022003009027081244, + "loss": 1.3827, + "theoretical_loss": 3.479391045917774, + "tokens_seen": 1862795264 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022002006018054162, + "loss": 1.3739, + "theoretical_loss": 3.4793807813296835, + "tokens_seen": 1862860800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022001003009027083, + "loss": 1.2972, + "theoretical_loss": 3.479370517203805, + "tokens_seen": 1862926336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00022, + "loss": 1.3562, + "theoretical_loss": 3.4793602535401016, + "tokens_seen": 1862991872 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002199899699097292, + "loss": 1.3428, + "theoretical_loss": 3.4793499903385356, + "tokens_seen": 1863057408 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021997993981945837, + "loss": 1.3691, + "theoretical_loss": 3.4793397275990703, + "tokens_seen": 1863122944 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021996990972918758, + "loss": 1.3711, + "theoretical_loss": 3.479329465321669, + "tokens_seen": 1863188480 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021995987963891676, + "loss": 1.3594, + "theoretical_loss": 3.4793192035062948, + "tokens_seen": 1863254016 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021994984954864594, + "loss": 1.4716, + "theoretical_loss": 3.47930894215291, + "tokens_seen": 1863319552 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021993981945837512, + "loss": 1.4347, + "theoretical_loss": 3.4792986812614775, + "tokens_seen": 1863385088 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002199297893681043, + "loss": 1.2209, + "theoretical_loss": 3.479288420831961, + "tokens_seen": 1863450624 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002199197592778335, + "loss": 1.3806, + "theoretical_loss": 3.479278160864323, + "tokens_seen": 1863516160 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002199097291875627, + "loss": 1.2838, + "theoretical_loss": 3.4792679013585266, + "tokens_seen": 1863581696 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021989969909729187, + "loss": 1.4213, + "theoretical_loss": 3.4792576423145345, + "tokens_seen": 1863647232 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021988966900702105, + "loss": 1.3477, + "theoretical_loss": 3.4792473837323095, + "tokens_seen": 1863712768 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021987963891675026, + "loss": 1.3686, + "theoretical_loss": 3.4792371256118155, + "tokens_seen": 1863778304 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021986960882647944, + "loss": 1.2666, + "theoretical_loss": 3.479226867953014, + "tokens_seen": 1863843840 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021985957873620862, + "loss": 1.3675, + "theoretical_loss": 3.47921661075587, + "tokens_seen": 1863909376 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002198495486459378, + "loss": 1.4005, + "theoretical_loss": 3.4792063540203446, + "tokens_seen": 1863974912 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021983951855566701, + "loss": 1.3756, + "theoretical_loss": 3.479196097746402, + "tokens_seen": 1864040448 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002198294884653962, + "loss": 1.423, + "theoretical_loss": 3.4791858419340045, + "tokens_seen": 1864105984 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021981945837512538, + "loss": 1.3283, + "theoretical_loss": 3.479175586583116, + "tokens_seen": 1864171520 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021980942828485456, + "loss": 1.3846, + "theoretical_loss": 3.4791653316936983, + "tokens_seen": 1864237056 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021979939819458374, + "loss": 1.3299, + "theoretical_loss": 3.479155077265715, + "tokens_seen": 1864302592 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021978936810431295, + "loss": 1.3919, + "theoretical_loss": 3.479144823299129, + "tokens_seen": 1864368128 + }, + { + "epoch": 0.56, + "learning_rate": 0.00021977933801404213, + "loss": 1.2116, + "theoretical_loss": 3.479134569793904, + "tokens_seen": 1864433664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 680280, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1499179601669312, + "objective/train/theoretical_loss": 3.479124316750002, + "objective/train/tokens_used": 1884959200, + "theoretical_loss": 3.479124316750002, + "tokens_seen": 1864499200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0002197693079237713, + "loss": 1.2381, + "theoretical_loss": 3.479124316750002, + "tokens_seen": 1864499200 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002197592778335005, + "loss": 1.3605, + "theoretical_loss": 3.479114064167387, + "tokens_seen": 1864564736 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002197492477432297, + "loss": 1.435, + "theoretical_loss": 3.479103812046021, + "tokens_seen": 1864630272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021973921765295888, + "loss": 1.2274, + "theoretical_loss": 3.4790935603858677, + "tokens_seen": 1864695808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021972918756268806, + "loss": 1.2931, + "theoretical_loss": 3.4790833091868905, + "tokens_seen": 1864761344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021971915747241724, + "loss": 1.2833, + "theoretical_loss": 3.4790730584490515, + "tokens_seen": 1864826880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021970912738214642, + "loss": 1.2898, + "theoretical_loss": 3.479062808172314, + "tokens_seen": 1864892416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021969909729187563, + "loss": 1.3553, + "theoretical_loss": 3.479052558356642, + "tokens_seen": 1864957952 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002196890672016048, + "loss": 1.3729, + "theoretical_loss": 3.479042309001997, + "tokens_seen": 1865023488 + }, + { + "epoch": 0.57, + "learning_rate": 0.000219679037111334, + "loss": 1.416, + "theoretical_loss": 3.4790320601083433, + "tokens_seen": 1865089024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021966900702106317, + "loss": 1.2798, + "theoretical_loss": 3.4790218116756435, + "tokens_seen": 1865154560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021965897693079238, + "loss": 1.3326, + "theoretical_loss": 3.4790115637038608, + "tokens_seen": 1865220096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021964894684052156, + "loss": 1.3187, + "theoretical_loss": 3.4790013161929583, + "tokens_seen": 1865285632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021963891675025075, + "loss": 1.3698, + "theoretical_loss": 3.478991069142899, + "tokens_seen": 1865351168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021962888665997993, + "loss": 1.3847, + "theoretical_loss": 3.4789808225536456, + "tokens_seen": 1865416704 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002196188565697091, + "loss": 1.4077, + "theoretical_loss": 3.478970576425162, + "tokens_seen": 1865482240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021960882647943832, + "loss": 1.3834, + "theoretical_loss": 3.4789603307574106, + "tokens_seen": 1865547776 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002195987963891675, + "loss": 1.3694, + "theoretical_loss": 3.478950085550355, + "tokens_seen": 1865613312 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002195887662988967, + "loss": 1.3442, + "theoretical_loss": 3.478939840803958, + "tokens_seen": 1865678848 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021957873620862589, + "loss": 1.3704, + "theoretical_loss": 3.478929596518183, + "tokens_seen": 1865744384 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002195687061183551, + "loss": 1.2855, + "theoretical_loss": 3.4789193526929925, + "tokens_seen": 1865809920 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021955867602808428, + "loss": 1.3744, + "theoretical_loss": 3.4789091093283497, + "tokens_seen": 1865875456 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021954864593781346, + "loss": 1.3123, + "theoretical_loss": 3.4788988664242186, + "tokens_seen": 1865940992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021953861584754264, + "loss": 1.3136, + "theoretical_loss": 3.478888623980562, + "tokens_seen": 1866006528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021952858575727182, + "loss": 1.4508, + "theoretical_loss": 3.478878381997342, + "tokens_seen": 1866072064 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021951855566700103, + "loss": 1.3966, + "theoretical_loss": 3.478868140474523, + "tokens_seen": 1866137600 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002195085255767302, + "loss": 1.3776, + "theoretical_loss": 3.478857899412067, + "tokens_seen": 1866203136 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002194984954864594, + "loss": 1.3535, + "theoretical_loss": 3.478847658809938, + "tokens_seen": 1866268672 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021948846539618857, + "loss": 1.3862, + "theoretical_loss": 3.4788374186680997, + "tokens_seen": 1866334208 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021947843530591778, + "loss": 1.4099, + "theoretical_loss": 3.4788271789865135, + "tokens_seen": 1866399744 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021946840521564696, + "loss": 1.3235, + "theoretical_loss": 3.478816939765144, + "tokens_seen": 1866465280 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021945837512537614, + "loss": 1.3198, + "theoretical_loss": 3.478806701003954, + "tokens_seen": 1866530816 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021944834503510532, + "loss": 1.2541, + "theoretical_loss": 3.478796462702906, + "tokens_seen": 1866596352 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002194383149448345, + "loss": 1.3344, + "theoretical_loss": 3.4787862248619645, + "tokens_seen": 1866661888 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002194282848545637, + "loss": 1.3231, + "theoretical_loss": 3.4787759874810913, + "tokens_seen": 1866727424 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002194182547642929, + "loss": 1.4478, + "theoretical_loss": 3.4787657505602505, + "tokens_seen": 1866792960 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021940822467402207, + "loss": 1.2701, + "theoretical_loss": 3.478755514099405, + "tokens_seen": 1866858496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021939819458375125, + "loss": 1.2732, + "theoretical_loss": 3.4787452780985175, + "tokens_seen": 1866924032 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021938816449348046, + "loss": 1.4047, + "theoretical_loss": 3.4787350425575516, + "tokens_seen": 1866989568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021937813440320964, + "loss": 1.2935, + "theoretical_loss": 3.478724807476471, + "tokens_seen": 1867055104 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021936810431293882, + "loss": 1.3853, + "theoretical_loss": 3.478714572855238, + "tokens_seen": 1867120640 + }, + { + "epoch": 0.57, + "learning_rate": 0.000219358074222668, + "loss": 1.2997, + "theoretical_loss": 3.4787043386938166, + "tokens_seen": 1867186176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021934804413239721, + "loss": 1.2994, + "theoretical_loss": 3.4786941049921687, + "tokens_seen": 1867251712 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002193380140421264, + "loss": 1.488, + "theoretical_loss": 3.4786838717502593, + "tokens_seen": 1867317248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021932798395185558, + "loss": 1.3323, + "theoretical_loss": 3.4786736389680506, + "tokens_seen": 1867382784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021931795386158476, + "loss": 1.3552, + "theoretical_loss": 3.4786634066455058, + "tokens_seen": 1867448320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021930792377131394, + "loss": 1.3187, + "theoretical_loss": 3.478653174782588, + "tokens_seen": 1867513856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021929789368104315, + "loss": 1.3264, + "theoretical_loss": 3.4786429433792607, + "tokens_seen": 1867579392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021928786359077233, + "loss": 1.351, + "theoretical_loss": 3.4786327124354877, + "tokens_seen": 1867644928 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002192778335005015, + "loss": 1.2978, + "theoretical_loss": 3.4786224819512306, + "tokens_seen": 1867710464 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 681429, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2940229177474976, + "objective/train/theoretical_loss": 3.478612251926455, + "objective/train/tokens_used": 1888236000, + "theoretical_loss": 3.478612251926455, + "tokens_seen": 1867776000 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002192678034102307, + "loss": 1.2987, + "theoretical_loss": 3.478612251926455, + "tokens_seen": 1867776000 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002192577733199599, + "loss": 1.2994, + "theoretical_loss": 3.478602022361122, + "tokens_seen": 1867841536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021924774322968908, + "loss": 1.4053, + "theoretical_loss": 3.4785917932551955, + "tokens_seen": 1867907072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021923771313941826, + "loss": 1.3254, + "theoretical_loss": 3.478581564608639, + "tokens_seen": 1867972608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021922768304914744, + "loss": 1.368, + "theoretical_loss": 3.4785713364214166, + "tokens_seen": 1868038144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021921765295887662, + "loss": 1.3517, + "theoretical_loss": 3.47856110869349, + "tokens_seen": 1868103680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021920762286860583, + "loss": 1.3093, + "theoretical_loss": 3.4785508814248227, + "tokens_seen": 1868169216 + }, + { + "epoch": 0.57, + "learning_rate": 0.000219197592778335, + "loss": 1.288, + "theoretical_loss": 3.478540654615379, + "tokens_seen": 1868234752 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002191875626880642, + "loss": 1.2979, + "theoretical_loss": 3.4785304282651213, + "tokens_seen": 1868300288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021917753259779337, + "loss": 1.2709, + "theoretical_loss": 3.478520202374013, + "tokens_seen": 1868365824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021916750250752258, + "loss": 1.3699, + "theoretical_loss": 3.478509976942018, + "tokens_seen": 1868431360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021915747241725176, + "loss": 1.2719, + "theoretical_loss": 3.478499751969098, + "tokens_seen": 1868496896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021914744232698095, + "loss": 1.2878, + "theoretical_loss": 3.4784895274552183, + "tokens_seen": 1868562432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021913741223671013, + "loss": 1.3502, + "theoretical_loss": 3.478479303400341, + "tokens_seen": 1868627968 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002191273821464393, + "loss": 1.5032, + "theoretical_loss": 3.4784690798044298, + "tokens_seen": 1868693504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021911735205616852, + "loss": 1.3216, + "theoretical_loss": 3.478458856667448, + "tokens_seen": 1868759040 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002191073219658977, + "loss": 1.2867, + "theoretical_loss": 3.478448633989358, + "tokens_seen": 1868824576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021909729187562688, + "loss": 1.4663, + "theoretical_loss": 3.4784384117701244, + "tokens_seen": 1868890112 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021908726178535606, + "loss": 1.4304, + "theoretical_loss": 3.4784281900097103, + "tokens_seen": 1868955648 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021907723169508527, + "loss": 1.276, + "theoretical_loss": 3.4784179687080785, + "tokens_seen": 1869021184 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021906720160481445, + "loss": 1.3625, + "theoretical_loss": 3.4784077478651922, + "tokens_seen": 1869086720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021905717151454363, + "loss": 1.3489, + "theoretical_loss": 3.4783975274810155, + "tokens_seen": 1869152256 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002190471414242728, + "loss": 1.443, + "theoretical_loss": 3.478387307555511, + "tokens_seen": 1869217792 + }, + { + "epoch": 0.57, + "learning_rate": 0.000219037111334002, + "loss": 1.3281, + "theoretical_loss": 3.478377088088642, + "tokens_seen": 1869283328 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002190270812437312, + "loss": 1.3528, + "theoretical_loss": 3.4783668690803733, + "tokens_seen": 1869348864 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021901705115346038, + "loss": 1.3323, + "theoretical_loss": 3.478356650530666, + "tokens_seen": 1869414400 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021900702106318956, + "loss": 1.4092, + "theoretical_loss": 3.4783464324394853, + "tokens_seen": 1869479936 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021899699097291874, + "loss": 1.3679, + "theoretical_loss": 3.4783362148067933, + "tokens_seen": 1869545472 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021898696088264795, + "loss": 1.3732, + "theoretical_loss": 3.4783259976325542, + "tokens_seen": 1869611008 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021897693079237713, + "loss": 1.3932, + "theoretical_loss": 3.4783157809167307, + "tokens_seen": 1869676544 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021896690070210631, + "loss": 1.4148, + "theoretical_loss": 3.4783055646592866, + "tokens_seen": 1869742080 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002189568706118355, + "loss": 1.4153, + "theoretical_loss": 3.4782953488601853, + "tokens_seen": 1869807616 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021894684052156468, + "loss": 1.3922, + "theoretical_loss": 3.47828513351939, + "tokens_seen": 1869873152 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021893681043129388, + "loss": 1.3179, + "theoretical_loss": 3.478274918636864, + "tokens_seen": 1869938688 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021892678034102307, + "loss": 1.4712, + "theoretical_loss": 3.478264704212571, + "tokens_seen": 1870004224 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021891675025075225, + "loss": 1.323, + "theoretical_loss": 3.4782544902464743, + "tokens_seen": 1870069760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021890672016048143, + "loss": 1.4002, + "theoretical_loss": 3.4782442767385366, + "tokens_seen": 1870135296 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021889669007021064, + "loss": 1.369, + "theoretical_loss": 3.4782340636887223, + "tokens_seen": 1870200832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021888665997993982, + "loss": 1.3511, + "theoretical_loss": 3.4782238510969945, + "tokens_seen": 1870266368 + }, + { + "epoch": 0.57, + "learning_rate": 0.000218876629889669, + "loss": 1.3061, + "theoretical_loss": 3.4782136389633163, + "tokens_seen": 1870331904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021886659979939818, + "loss": 1.3788, + "theoretical_loss": 3.478203427287651, + "tokens_seen": 1870397440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021885656970912736, + "loss": 1.4103, + "theoretical_loss": 3.4781932160699625, + "tokens_seen": 1870462976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021884653961885657, + "loss": 1.3213, + "theoretical_loss": 3.478183005310214, + "tokens_seen": 1870528512 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021883650952858578, + "loss": 1.2384, + "theoretical_loss": 3.478172795008369, + "tokens_seen": 1870594048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021882647943831496, + "loss": 1.3392, + "theoretical_loss": 3.478162585164391, + "tokens_seen": 1870659584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021881644934804414, + "loss": 1.2498, + "theoretical_loss": 3.478152375778243, + "tokens_seen": 1870725120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021880641925777335, + "loss": 1.3088, + "theoretical_loss": 3.4781421668498886, + "tokens_seen": 1870790656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021879638916750253, + "loss": 1.4187, + "theoretical_loss": 3.4781319583792913, + "tokens_seen": 1870856192 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002187863590772317, + "loss": 1.2542, + "theoretical_loss": 3.478121750366415, + "tokens_seen": 1870921728 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002187763289869609, + "loss": 1.4952, + "theoretical_loss": 3.478111542811222, + "tokens_seen": 1870987264 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 682627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3522326946258545, + "objective/train/theoretical_loss": 3.4781013357136774, + "objective/train/tokens_used": 1891512800, + "theoretical_loss": 3.4781013357136774, + "tokens_seen": 1871052800 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002187662988966901, + "loss": 1.3135, + "theoretical_loss": 3.4781013357136774, + "tokens_seen": 1871052800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021875626880641928, + "loss": 1.3686, + "theoretical_loss": 3.4780911290737433, + "tokens_seen": 1871118336 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021874623871614846, + "loss": 1.4847, + "theoretical_loss": 3.478080922891383, + "tokens_seen": 1871183872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021873620862587764, + "loss": 1.3601, + "theoretical_loss": 3.4780707171665615, + "tokens_seen": 1871249408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021872617853560682, + "loss": 1.407, + "theoretical_loss": 3.478060511899241, + "tokens_seen": 1871314944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021871614844533603, + "loss": 1.4497, + "theoretical_loss": 3.478050307089385, + "tokens_seen": 1871380480 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002187061183550652, + "loss": 1.3688, + "theoretical_loss": 3.4780401027369576, + "tokens_seen": 1871446016 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002186960882647944, + "loss": 1.3147, + "theoretical_loss": 3.478029898841922, + "tokens_seen": 1871511552 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021868605817452357, + "loss": 1.3156, + "theoretical_loss": 3.4780196954042415, + "tokens_seen": 1871577088 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021867602808425278, + "loss": 1.2626, + "theoretical_loss": 3.4780094924238796, + "tokens_seen": 1871642624 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021866599799398196, + "loss": 1.366, + "theoretical_loss": 3.4779992899008, + "tokens_seen": 1871708160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021865596790371115, + "loss": 1.3635, + "theoretical_loss": 3.477989087834966, + "tokens_seen": 1871773696 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021864593781344033, + "loss": 1.4025, + "theoretical_loss": 3.477978886226341, + "tokens_seen": 1871839232 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002186359077231695, + "loss": 1.4598, + "theoretical_loss": 3.477968685074889, + "tokens_seen": 1871904768 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021862587763289872, + "loss": 1.3926, + "theoretical_loss": 3.477958484380573, + "tokens_seen": 1871970304 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002186158475426279, + "loss": 1.325, + "theoretical_loss": 3.477948284143357, + "tokens_seen": 1872035840 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021860581745235708, + "loss": 1.4552, + "theoretical_loss": 3.477938084363204, + "tokens_seen": 1872101376 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021859578736208626, + "loss": 1.3236, + "theoretical_loss": 3.477927885040078, + "tokens_seen": 1872166912 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021858575727181547, + "loss": 1.4127, + "theoretical_loss": 3.477917686173942, + "tokens_seen": 1872232448 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021857572718154465, + "loss": 1.3869, + "theoretical_loss": 3.4779074877647593, + "tokens_seen": 1872297984 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021856569709127383, + "loss": 1.4264, + "theoretical_loss": 3.4778972898124945, + "tokens_seen": 1872363520 + }, + { + "epoch": 0.57, + "learning_rate": 0.000218555667001003, + "loss": 1.3605, + "theoretical_loss": 3.477887092317111, + "tokens_seen": 1872429056 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002185456369107322, + "loss": 1.2963, + "theoretical_loss": 3.4778768952785715, + "tokens_seen": 1872494592 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002185356068204614, + "loss": 1.3529, + "theoretical_loss": 3.4778666986968396, + "tokens_seen": 1872560128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021852557673019058, + "loss": 1.335, + "theoretical_loss": 3.4778565025718797, + "tokens_seen": 1872625664 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021851554663991976, + "loss": 1.3799, + "theoretical_loss": 3.477846306903655, + "tokens_seen": 1872691200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021850551654964894, + "loss": 1.3028, + "theoretical_loss": 3.4778361116921284, + "tokens_seen": 1872756736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021849548645937815, + "loss": 1.3277, + "theoretical_loss": 3.477825916937264, + "tokens_seen": 1872822272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021848545636910733, + "loss": 1.4263, + "theoretical_loss": 3.4778157226390256, + "tokens_seen": 1872887808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021847542627883651, + "loss": 1.2701, + "theoretical_loss": 3.4778055287973766, + "tokens_seen": 1872953344 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002184653961885657, + "loss": 1.3839, + "theoretical_loss": 3.4777953354122806, + "tokens_seen": 1873018880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021845536609829488, + "loss": 1.3869, + "theoretical_loss": 3.4777851424837007, + "tokens_seen": 1873084416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021844533600802408, + "loss": 1.3191, + "theoretical_loss": 3.477774950011601, + "tokens_seen": 1873149952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021843530591775327, + "loss": 1.4495, + "theoretical_loss": 3.4777647579959448, + "tokens_seen": 1873215488 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021842527582748245, + "loss": 1.3794, + "theoretical_loss": 3.477754566436696, + "tokens_seen": 1873281024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021841524573721163, + "loss": 1.3673, + "theoretical_loss": 3.4777443753338178, + "tokens_seen": 1873346560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021840521564694084, + "loss": 1.2798, + "theoretical_loss": 3.4777341846872742, + "tokens_seen": 1873412096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021839518555667002, + "loss": 1.327, + "theoretical_loss": 3.477723994497029, + "tokens_seen": 1873477632 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002183851554663992, + "loss": 1.3833, + "theoretical_loss": 3.477713804763045, + "tokens_seen": 1873543168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021837512537612838, + "loss": 1.4489, + "theoretical_loss": 3.4777036154852863, + "tokens_seen": 1873608704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021836509528585756, + "loss": 1.2948, + "theoretical_loss": 3.4776934266637167, + "tokens_seen": 1873674240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021835506519558677, + "loss": 1.3736, + "theoretical_loss": 3.477683238298299, + "tokens_seen": 1873739776 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021834503510531595, + "loss": 1.3095, + "theoretical_loss": 3.477673050388998, + "tokens_seen": 1873805312 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021833500501504513, + "loss": 1.4577, + "theoretical_loss": 3.4776628629357766, + "tokens_seen": 1873870848 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002183249749247743, + "loss": 1.2863, + "theoretical_loss": 3.4776526759385984, + "tokens_seen": 1873936384 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021831494483450352, + "loss": 1.357, + "theoretical_loss": 3.477642489397428, + "tokens_seen": 1874001920 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002183049147442327, + "loss": 1.2575, + "theoretical_loss": 3.477632303312227, + "tokens_seen": 1874067456 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021829488465396188, + "loss": 1.3775, + "theoretical_loss": 3.4776221176829614, + "tokens_seen": 1874132992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021828485456369106, + "loss": 1.33, + "theoretical_loss": 3.477611932509593, + "tokens_seen": 1874198528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021827482447342024, + "loss": 1.4469, + "theoretical_loss": 3.4776017477920864, + "tokens_seen": 1874264064 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 683764, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.466771125793457, + "objective/train/theoretical_loss": 3.477591563530405, + "objective/train/tokens_used": 1894789600, + "theoretical_loss": 3.477591563530405, + "tokens_seen": 1874329600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021826479438314945, + "loss": 1.4282, + "theoretical_loss": 3.477591563530405, + "tokens_seen": 1874329600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021825476429287863, + "loss": 1.3137, + "theoretical_loss": 3.477581379724513, + "tokens_seen": 1874395136 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021824473420260782, + "loss": 1.4763, + "theoretical_loss": 3.477571196374373, + "tokens_seen": 1874460672 + }, + { + "epoch": 0.57, + "learning_rate": 0.000218234704112337, + "loss": 1.4586, + "theoretical_loss": 3.4775610134799497, + "tokens_seen": 1874526208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002182246740220662, + "loss": 1.4288, + "theoretical_loss": 3.477550831041206, + "tokens_seen": 1874591744 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021821464393179539, + "loss": 1.3168, + "theoretical_loss": 3.4775406490581062, + "tokens_seen": 1874657280 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021820461384152457, + "loss": 1.4671, + "theoretical_loss": 3.4775304675306136, + "tokens_seen": 1874722816 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021819458375125375, + "loss": 1.4223, + "theoretical_loss": 3.4775202864586916, + "tokens_seen": 1874788352 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021818455366098296, + "loss": 1.3333, + "theoretical_loss": 3.477510105842305, + "tokens_seen": 1874853888 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021817452357071214, + "loss": 1.325, + "theoretical_loss": 3.4774999256814163, + "tokens_seen": 1874919424 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021816449348044132, + "loss": 1.2963, + "theoretical_loss": 3.4774897459759897, + "tokens_seen": 1874984960 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002181544633901705, + "loss": 1.3859, + "theoretical_loss": 3.477479566725989, + "tokens_seen": 1875050496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021814443329989968, + "loss": 1.4016, + "theoretical_loss": 3.477469387931378, + "tokens_seen": 1875116032 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002181344032096289, + "loss": 1.2773, + "theoretical_loss": 3.4774592095921197, + "tokens_seen": 1875181568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021812437311935807, + "loss": 1.4067, + "theoretical_loss": 3.4774490317081783, + "tokens_seen": 1875247104 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021811434302908725, + "loss": 1.2456, + "theoretical_loss": 3.477438854279518, + "tokens_seen": 1875312640 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021810431293881643, + "loss": 1.3059, + "theoretical_loss": 3.4774286773061016, + "tokens_seen": 1875378176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021809428284854564, + "loss": 1.4548, + "theoretical_loss": 3.477418500787894, + "tokens_seen": 1875443712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021808425275827485, + "loss": 1.3115, + "theoretical_loss": 3.4774083247248573, + "tokens_seen": 1875509248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021807422266800403, + "loss": 1.2882, + "theoretical_loss": 3.477398149116957, + "tokens_seen": 1875574784 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002180641925777332, + "loss": 1.3927, + "theoretical_loss": 3.4773879739641553, + "tokens_seen": 1875640320 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002180541624874624, + "loss": 1.4057, + "theoretical_loss": 3.477377799266417, + "tokens_seen": 1875705856 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002180441323971916, + "loss": 1.3811, + "theoretical_loss": 3.4773676250237058, + "tokens_seen": 1875771392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021803410230692078, + "loss": 1.3321, + "theoretical_loss": 3.477357451235984, + "tokens_seen": 1875836928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021802407221664996, + "loss": 1.368, + "theoretical_loss": 3.4773472779032177, + "tokens_seen": 1875902464 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021801404212637914, + "loss": 1.5218, + "theoretical_loss": 3.477337105025369, + "tokens_seen": 1875968000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021800401203610835, + "loss": 1.4261, + "theoretical_loss": 3.477326932602402, + "tokens_seen": 1876033536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021799398194583753, + "loss": 1.304, + "theoretical_loss": 3.477316760634281, + "tokens_seen": 1876099072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021798395185556671, + "loss": 1.4102, + "theoretical_loss": 3.4773065891209693, + "tokens_seen": 1876164608 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002179739217652959, + "loss": 1.2765, + "theoretical_loss": 3.477296418062431, + "tokens_seen": 1876230144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021796389167502508, + "loss": 1.3735, + "theoretical_loss": 3.477286247458629, + "tokens_seen": 1876295680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021795386158475428, + "loss": 1.3496, + "theoretical_loss": 3.477276077309528, + "tokens_seen": 1876361216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021794383149448347, + "loss": 1.3592, + "theoretical_loss": 3.4772659076150916, + "tokens_seen": 1876426752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021793380140421265, + "loss": 1.329, + "theoretical_loss": 3.477255738375283, + "tokens_seen": 1876492288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021792377131394183, + "loss": 1.3163, + "theoretical_loss": 3.4772455695900675, + "tokens_seen": 1876557824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021791374122367104, + "loss": 1.3425, + "theoretical_loss": 3.4772354012594073, + "tokens_seen": 1876623360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021790371113340022, + "loss": 1.3269, + "theoretical_loss": 3.4772252333832667, + "tokens_seen": 1876688896 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002178936810431294, + "loss": 1.3618, + "theoretical_loss": 3.4772150659616097, + "tokens_seen": 1876754432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021788365095285858, + "loss": 1.3372, + "theoretical_loss": 3.4772048989944, + "tokens_seen": 1876819968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021787362086258776, + "loss": 1.419, + "theoretical_loss": 3.4771947324816024, + "tokens_seen": 1876885504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021786359077231697, + "loss": 1.3182, + "theoretical_loss": 3.477184566423179, + "tokens_seen": 1876951040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021785356068204615, + "loss": 1.2929, + "theoretical_loss": 3.477174400819094, + "tokens_seen": 1877016576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021784353059177533, + "loss": 1.3693, + "theoretical_loss": 3.4771642356693127, + "tokens_seen": 1877082112 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002178335005015045, + "loss": 1.2829, + "theoretical_loss": 3.477154070973797, + "tokens_seen": 1877147648 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021782347041123372, + "loss": 1.3775, + "theoretical_loss": 3.477143906732512, + "tokens_seen": 1877213184 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002178134403209629, + "loss": 1.4163, + "theoretical_loss": 3.477133742945421, + "tokens_seen": 1877278720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021780341023069208, + "loss": 1.2971, + "theoretical_loss": 3.4771235796124884, + "tokens_seen": 1877344256 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021779338014042126, + "loss": 1.3868, + "theoretical_loss": 3.4771134167336775, + "tokens_seen": 1877409792 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021778335005015044, + "loss": 1.4079, + "theoretical_loss": 3.477103254308952, + "tokens_seen": 1877475328 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021777331995987965, + "loss": 1.2999, + "theoretical_loss": 3.4770930923382766, + "tokens_seen": 1877540864 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 684931, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3722288608551025, + "objective/train/theoretical_loss": 3.4770829308216147, + "objective/train/tokens_used": 1898066400, + "theoretical_loss": 3.4770829308216147, + "tokens_seen": 1877606400 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021776328986960883, + "loss": 1.5418, + "theoretical_loss": 3.4770829308216147, + "tokens_seen": 1877606400 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021775325977933802, + "loss": 1.3151, + "theoretical_loss": 3.47707276975893, + "tokens_seen": 1877671936 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002177432296890672, + "loss": 1.3164, + "theoretical_loss": 3.4770626091501864, + "tokens_seen": 1877737472 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002177331995987964, + "loss": 1.456, + "theoretical_loss": 3.477052448995348, + "tokens_seen": 1877803008 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021772316950852559, + "loss": 1.4728, + "theoretical_loss": 3.4770422892943786, + "tokens_seen": 1877868544 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021771313941825477, + "loss": 1.3175, + "theoretical_loss": 3.4770321300472418, + "tokens_seen": 1877934080 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021770310932798395, + "loss": 1.4757, + "theoretical_loss": 3.477021971253902, + "tokens_seen": 1877999616 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021769307923771316, + "loss": 1.4808, + "theoretical_loss": 3.4770118129143226, + "tokens_seen": 1878065152 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021768304914744234, + "loss": 1.3732, + "theoretical_loss": 3.4770016550284684, + "tokens_seen": 1878130688 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021767301905717152, + "loss": 1.3996, + "theoretical_loss": 3.476991497596302, + "tokens_seen": 1878196224 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002176629889669007, + "loss": 1.3752, + "theoretical_loss": 3.476981340617788, + "tokens_seen": 1878261760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021765295887662988, + "loss": 1.3528, + "theoretical_loss": 3.47697118409289, + "tokens_seen": 1878327296 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002176429287863591, + "loss": 1.3991, + "theoretical_loss": 3.476961028021573, + "tokens_seen": 1878392832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021763289869608827, + "loss": 1.365, + "theoretical_loss": 3.4769508724038, + "tokens_seen": 1878458368 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021762286860581745, + "loss": 1.4015, + "theoretical_loss": 3.4769407172395344, + "tokens_seen": 1878523904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021761283851554663, + "loss": 1.4683, + "theoretical_loss": 3.476930562528741, + "tokens_seen": 1878589440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021760280842527584, + "loss": 1.3867, + "theoretical_loss": 3.4769204082713836, + "tokens_seen": 1878654976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021759277833500502, + "loss": 1.4277, + "theoretical_loss": 3.476910254467426, + "tokens_seen": 1878720512 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002175827482447342, + "loss": 1.4007, + "theoretical_loss": 3.476900101116832, + "tokens_seen": 1878786048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021757271815446338, + "loss": 1.3347, + "theoretical_loss": 3.476889948219566, + "tokens_seen": 1878851584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021756268806419257, + "loss": 1.2971, + "theoretical_loss": 3.4768797957755915, + "tokens_seen": 1878917120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021755265797392177, + "loss": 1.3096, + "theoretical_loss": 3.4768696437848723, + "tokens_seen": 1878982656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021754262788365095, + "loss": 1.2447, + "theoretical_loss": 3.4768594922473732, + "tokens_seen": 1879048192 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021753259779338014, + "loss": 1.4163, + "theoretical_loss": 3.4768493411630574, + "tokens_seen": 1879113728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021752256770310932, + "loss": 1.3276, + "theoretical_loss": 3.476839190531889, + "tokens_seen": 1879179264 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021751253761283852, + "loss": 1.394, + "theoretical_loss": 3.476829040353832, + "tokens_seen": 1879244800 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002175025075225677, + "loss": 1.2995, + "theoretical_loss": 3.4768188906288504, + "tokens_seen": 1879310336 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002174924774322969, + "loss": 1.3353, + "theoretical_loss": 3.4768087413569084, + "tokens_seen": 1879375872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021748244734202607, + "loss": 1.3355, + "theoretical_loss": 3.4767985925379703, + "tokens_seen": 1879441408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021747241725175525, + "loss": 1.317, + "theoretical_loss": 3.4767884441719987, + "tokens_seen": 1879506944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021746238716148446, + "loss": 1.3311, + "theoretical_loss": 3.4767782962589586, + "tokens_seen": 1879572480 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021745235707121364, + "loss": 1.3298, + "theoretical_loss": 3.476768148798814, + "tokens_seen": 1879638016 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021744232698094282, + "loss": 1.4049, + "theoretical_loss": 3.476758001791529, + "tokens_seen": 1879703552 + }, + { + "epoch": 0.57, + "learning_rate": 0.000217432296890672, + "loss": 1.3473, + "theoretical_loss": 3.4767478552370674, + "tokens_seen": 1879769088 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002174222668004012, + "loss": 1.3806, + "theoretical_loss": 3.4767377091353926, + "tokens_seen": 1879834624 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002174122367101304, + "loss": 1.353, + "theoretical_loss": 3.4767275634864694, + "tokens_seen": 1879900160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021740220661985957, + "loss": 1.3811, + "theoretical_loss": 3.476717418290262, + "tokens_seen": 1879965696 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021739217652958875, + "loss": 1.3627, + "theoretical_loss": 3.476707273546733, + "tokens_seen": 1880031232 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021738214643931793, + "loss": 1.3307, + "theoretical_loss": 3.4766971292558484, + "tokens_seen": 1880096768 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021737211634904714, + "loss": 1.3865, + "theoretical_loss": 3.476686985417571, + "tokens_seen": 1880162304 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021736208625877632, + "loss": 1.2962, + "theoretical_loss": 3.476676842031865, + "tokens_seen": 1880227840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002173520561685055, + "loss": 1.3598, + "theoretical_loss": 3.4766666990986947, + "tokens_seen": 1880293376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002173420260782347, + "loss": 1.2679, + "theoretical_loss": 3.476656556618024, + "tokens_seen": 1880358912 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021733199598796392, + "loss": 1.2859, + "theoretical_loss": 3.4766464145898164, + "tokens_seen": 1880424448 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002173219658976931, + "loss": 1.329, + "theoretical_loss": 3.4766362730140368, + "tokens_seen": 1880489984 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021731193580742228, + "loss": 1.4241, + "theoretical_loss": 3.476626131890649, + "tokens_seen": 1880555520 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021730190571715146, + "loss": 1.3158, + "theoretical_loss": 3.476615991219617, + "tokens_seen": 1880621056 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021729187562688065, + "loss": 1.3655, + "theoretical_loss": 3.4766058510009046, + "tokens_seen": 1880686592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021728184553660985, + "loss": 1.3169, + "theoretical_loss": 3.476595711234476, + "tokens_seen": 1880752128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021727181544633903, + "loss": 1.2639, + "theoretical_loss": 3.476585571920295, + "tokens_seen": 1880817664 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 686045, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.286769986152649, + "objective/train/theoretical_loss": 3.476575433058327, + "objective/train/tokens_used": 1901343200, + "theoretical_loss": 3.476575433058327, + "tokens_seen": 1880883200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021726178535606822, + "loss": 1.3507, + "theoretical_loss": 3.476575433058327, + "tokens_seen": 1880883200 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002172517552657974, + "loss": 1.4351, + "theoretical_loss": 3.476565294648535, + "tokens_seen": 1880948736 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002172417251755266, + "loss": 1.3324, + "theoretical_loss": 3.4765551566908823, + "tokens_seen": 1881014272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021723169508525579, + "loss": 1.4783, + "theoretical_loss": 3.476545019185334, + "tokens_seen": 1881079808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021722166499498497, + "loss": 1.2795, + "theoretical_loss": 3.4765348821318547, + "tokens_seen": 1881145344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021721163490471415, + "loss": 1.3761, + "theoretical_loss": 3.4765247455304076, + "tokens_seen": 1881210880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021720160481444336, + "loss": 1.3473, + "theoretical_loss": 3.476514609380957, + "tokens_seen": 1881276416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021719157472417254, + "loss": 1.2279, + "theoretical_loss": 3.476504473683467, + "tokens_seen": 1881341952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021718154463390172, + "loss": 1.27, + "theoretical_loss": 3.4764943384379015, + "tokens_seen": 1881407488 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002171715145436309, + "loss": 1.2444, + "theoretical_loss": 3.476484203644225, + "tokens_seen": 1881473024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021716148445336008, + "loss": 1.3403, + "theoretical_loss": 3.4764740693024017, + "tokens_seen": 1881538560 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002171514543630893, + "loss": 1.3504, + "theoretical_loss": 3.4764639354123954, + "tokens_seen": 1881604096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021714142427281847, + "loss": 1.3075, + "theoretical_loss": 3.47645380197417, + "tokens_seen": 1881669632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021713139418254765, + "loss": 1.3207, + "theoretical_loss": 3.4764436689876903, + "tokens_seen": 1881735168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021712136409227683, + "loss": 1.2964, + "theoretical_loss": 3.47643353645292, + "tokens_seen": 1881800704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021711133400200604, + "loss": 1.2978, + "theoretical_loss": 3.476423404369823, + "tokens_seen": 1881866240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021710130391173522, + "loss": 1.3399, + "theoretical_loss": 3.4764132727383643, + "tokens_seen": 1881931776 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002170912738214644, + "loss": 1.4951, + "theoretical_loss": 3.4764031415585066, + "tokens_seen": 1881997312 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021708124373119358, + "loss": 1.3071, + "theoretical_loss": 3.4763930108302157, + "tokens_seen": 1882062848 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021707121364092277, + "loss": 1.4608, + "theoretical_loss": 3.4763828805534547, + "tokens_seen": 1882128384 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021706118355065197, + "loss": 1.2739, + "theoretical_loss": 3.4763727507281876, + "tokens_seen": 1882193920 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021705115346038115, + "loss": 1.3081, + "theoretical_loss": 3.4763626213543795, + "tokens_seen": 1882259456 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021704112337011034, + "loss": 1.3133, + "theoretical_loss": 3.476352492431994, + "tokens_seen": 1882324992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021703109327983952, + "loss": 1.3014, + "theoretical_loss": 3.4763423639609954, + "tokens_seen": 1882390528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021702106318956873, + "loss": 1.3934, + "theoretical_loss": 3.4763322359413475, + "tokens_seen": 1882456064 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002170110330992979, + "loss": 1.4159, + "theoretical_loss": 3.4763221083730147, + "tokens_seen": 1882521600 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002170010030090271, + "loss": 1.33, + "theoretical_loss": 3.4763119812559613, + "tokens_seen": 1882587136 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021699097291875627, + "loss": 1.3942, + "theoretical_loss": 3.4763018545901514, + "tokens_seen": 1882652672 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021698094282848545, + "loss": 1.4399, + "theoretical_loss": 3.476291728375549, + "tokens_seen": 1882718208 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021697091273821466, + "loss": 1.4924, + "theoretical_loss": 3.476281602612119, + "tokens_seen": 1882783744 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021696088264794384, + "loss": 1.3882, + "theoretical_loss": 3.476271477299825, + "tokens_seen": 1882849280 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021695085255767302, + "loss": 1.2819, + "theoretical_loss": 3.476261352438631, + "tokens_seen": 1882914816 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002169408224674022, + "loss": 1.2787, + "theoretical_loss": 3.4762512280285014, + "tokens_seen": 1882980352 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002169307923771314, + "loss": 1.3697, + "theoretical_loss": 3.476241104069401, + "tokens_seen": 1883045888 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002169207622868606, + "loss": 1.3671, + "theoretical_loss": 3.4762309805612928, + "tokens_seen": 1883111424 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021691073219658977, + "loss": 1.2262, + "theoretical_loss": 3.4762208575041424, + "tokens_seen": 1883176960 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021690070210631895, + "loss": 1.3679, + "theoretical_loss": 3.476210734897913, + "tokens_seen": 1883242496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021689067201604813, + "loss": 1.2377, + "theoretical_loss": 3.476200612742569, + "tokens_seen": 1883308032 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021688064192577734, + "loss": 1.3879, + "theoretical_loss": 3.476190491038075, + "tokens_seen": 1883373568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021687061183550652, + "loss": 1.4819, + "theoretical_loss": 3.476180369784395, + "tokens_seen": 1883439104 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002168605817452357, + "loss": 1.2996, + "theoretical_loss": 3.476170248981493, + "tokens_seen": 1883504640 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021685055165496489, + "loss": 1.3362, + "theoretical_loss": 3.476160128629334, + "tokens_seen": 1883570176 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002168405215646941, + "loss": 1.3439, + "theoretical_loss": 3.476150008727881, + "tokens_seen": 1883635712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021683049147442327, + "loss": 1.3719, + "theoretical_loss": 3.4761398892771, + "tokens_seen": 1883701248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021682046138415246, + "loss": 1.3727, + "theoretical_loss": 3.4761297702769536, + "tokens_seen": 1883766784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021681043129388164, + "loss": 1.3694, + "theoretical_loss": 3.476119651727407, + "tokens_seen": 1883832320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021680040120361082, + "loss": 1.3986, + "theoretical_loss": 3.476109533628424, + "tokens_seen": 1883897856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021679037111334003, + "loss": 1.3603, + "theoretical_loss": 3.4760994159799687, + "tokens_seen": 1883963392 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002167803410230692, + "loss": 1.4181, + "theoretical_loss": 3.4760892987820062, + "tokens_seen": 1884028928 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002167703109327984, + "loss": 1.3463, + "theoretical_loss": 3.4760791820345, + "tokens_seen": 1884094464 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 687181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3772673606872559, + "objective/train/theoretical_loss": 3.4760690657374145, + "objective/train/tokens_used": 1904620000, + "theoretical_loss": 3.4760690657374145, + "tokens_seen": 1884160000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021676028084252757, + "loss": 1.4419, + "theoretical_loss": 3.4760690657374145, + "tokens_seen": 1884160000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021675025075225678, + "loss": 1.3179, + "theoretical_loss": 3.476058949890714, + "tokens_seen": 1884225536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021674022066198596, + "loss": 1.4539, + "theoretical_loss": 3.476048834494363, + "tokens_seen": 1884291072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021673019057171514, + "loss": 1.3991, + "theoretical_loss": 3.4760387195483258, + "tokens_seen": 1884356608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021672016048144432, + "loss": 1.337, + "theoretical_loss": 3.4760286050525666, + "tokens_seen": 1884422144 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002167101303911735, + "loss": 1.3634, + "theoretical_loss": 3.47601849100705, + "tokens_seen": 1884487680 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002167001003009027, + "loss": 1.297, + "theoretical_loss": 3.4760083774117394, + "tokens_seen": 1884553216 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002166900702106319, + "loss": 1.4193, + "theoretical_loss": 3.4759982642666003, + "tokens_seen": 1884618752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021668004012036107, + "loss": 1.4418, + "theoretical_loss": 3.475988151571596, + "tokens_seen": 1884684288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021667001003009025, + "loss": 1.3926, + "theoretical_loss": 3.475978039326691, + "tokens_seen": 1884749824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021665997993981946, + "loss": 1.4697, + "theoretical_loss": 3.4759679275318502, + "tokens_seen": 1884815360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021664994984954864, + "loss": 1.4286, + "theoretical_loss": 3.4759578161870373, + "tokens_seen": 1884880896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021663991975927782, + "loss": 1.299, + "theoretical_loss": 3.475947705292217, + "tokens_seen": 1884946432 + }, + { + "epoch": 0.57, + "learning_rate": 0.000216629889669007, + "loss": 1.3514, + "theoretical_loss": 3.4759375948473537, + "tokens_seen": 1885011968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021661985957873621, + "loss": 1.495, + "theoretical_loss": 3.475927484852411, + "tokens_seen": 1885077504 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002166098294884654, + "loss": 1.4296, + "theoretical_loss": 3.475917375307354, + "tokens_seen": 1885143040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021659979939819458, + "loss": 1.4093, + "theoretical_loss": 3.475907266212147, + "tokens_seen": 1885208576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021658976930792378, + "loss": 1.3305, + "theoretical_loss": 3.475897157566754, + "tokens_seen": 1885274112 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021657973921765297, + "loss": 1.4107, + "theoretical_loss": 3.4758870493711393, + "tokens_seen": 1885339648 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021656970912738217, + "loss": 1.2704, + "theoretical_loss": 3.475876941625268, + "tokens_seen": 1885405184 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021655967903711135, + "loss": 1.282, + "theoretical_loss": 3.4758668343291035, + "tokens_seen": 1885470720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021654964894684054, + "loss": 1.3342, + "theoretical_loss": 3.4758567274826104, + "tokens_seen": 1885536256 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021653961885656972, + "loss": 1.3816, + "theoretical_loss": 3.475846621085754, + "tokens_seen": 1885601792 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021652958876629893, + "loss": 1.3332, + "theoretical_loss": 3.4758365151384973, + "tokens_seen": 1885667328 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002165195586760281, + "loss": 1.3197, + "theoretical_loss": 3.4758264096408054, + "tokens_seen": 1885732864 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002165095285857573, + "loss": 1.3942, + "theoretical_loss": 3.4758163045926427, + "tokens_seen": 1885798400 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021649949849548647, + "loss": 1.3158, + "theoretical_loss": 3.475806199993973, + "tokens_seen": 1885863936 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021648946840521565, + "loss": 1.433, + "theoretical_loss": 3.475796095844762, + "tokens_seen": 1885929472 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021647943831494486, + "loss": 1.4195, + "theoretical_loss": 3.4757859921449725, + "tokens_seen": 1885995008 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021646940822467404, + "loss": 1.3864, + "theoretical_loss": 3.47577588889457, + "tokens_seen": 1886060544 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021645937813440322, + "loss": 1.4096, + "theoretical_loss": 3.4757657860935183, + "tokens_seen": 1886126080 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002164493480441324, + "loss": 1.4093, + "theoretical_loss": 3.475755683741782, + "tokens_seen": 1886191616 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002164393179538616, + "loss": 1.3804, + "theoretical_loss": 3.4757455818393255, + "tokens_seen": 1886257152 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002164292878635908, + "loss": 1.3263, + "theoretical_loss": 3.4757354803861134, + "tokens_seen": 1886322688 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021641925777331997, + "loss": 1.331, + "theoretical_loss": 3.4757253793821095, + "tokens_seen": 1886388224 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021640922768304915, + "loss": 1.3225, + "theoretical_loss": 3.475715278827279, + "tokens_seen": 1886453760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021639919759277833, + "loss": 1.3289, + "theoretical_loss": 3.4757051787215865, + "tokens_seen": 1886519296 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021638916750250754, + "loss": 1.4323, + "theoretical_loss": 3.4756950790649954, + "tokens_seen": 1886584832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021637913741223672, + "loss": 1.3138, + "theoretical_loss": 3.4756849798574705, + "tokens_seen": 1886650368 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002163691073219659, + "loss": 1.3683, + "theoretical_loss": 3.4756748810989766, + "tokens_seen": 1886715904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021635907723169509, + "loss": 1.3914, + "theoretical_loss": 3.4756647827894778, + "tokens_seen": 1886781440 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002163490471414243, + "loss": 1.2722, + "theoretical_loss": 3.475654684928939, + "tokens_seen": 1886846976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021633901705115347, + "loss": 1.3209, + "theoretical_loss": 3.475644587517324, + "tokens_seen": 1886912512 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021632898696088266, + "loss": 1.3149, + "theoretical_loss": 3.4756344905545977, + "tokens_seen": 1886978048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021631895687061184, + "loss": 1.3522, + "theoretical_loss": 3.4756243940407243, + "tokens_seen": 1887043584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021630892678034102, + "loss": 1.3178, + "theoretical_loss": 3.4756142979756683, + "tokens_seen": 1887109120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021629889669007023, + "loss": 1.4027, + "theoretical_loss": 3.4756042023593947, + "tokens_seen": 1887174656 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002162888665997994, + "loss": 1.2541, + "theoretical_loss": 3.475594107191867, + "tokens_seen": 1887240192 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002162788365095286, + "loss": 1.3384, + "theoretical_loss": 3.4755840124730506, + "tokens_seen": 1887305728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021626880641925777, + "loss": 1.46, + "theoretical_loss": 3.475573918202909, + "tokens_seen": 1887371264 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 688313, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.499873399734497, + "objective/train/theoretical_loss": 3.4755638243814078, + "objective/train/tokens_used": 1907896800, + "theoretical_loss": 3.4755638243814078, + "tokens_seen": 1887436800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021625877632898698, + "loss": 1.3576, + "theoretical_loss": 3.4755638243814078, + "tokens_seen": 1887436800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021624874623871616, + "loss": 1.4418, + "theoretical_loss": 3.475553731008511, + "tokens_seen": 1887502336 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021623871614844534, + "loss": 1.3939, + "theoretical_loss": 3.475543638084183, + "tokens_seen": 1887567872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021622868605817452, + "loss": 1.3362, + "theoretical_loss": 3.4755335456083873, + "tokens_seen": 1887633408 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002162186559679037, + "loss": 1.4117, + "theoretical_loss": 3.4755234535810904, + "tokens_seen": 1887698944 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002162086258776329, + "loss": 1.3302, + "theoretical_loss": 3.4755133620022556, + "tokens_seen": 1887764480 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002161985957873621, + "loss": 1.304, + "theoretical_loss": 3.4755032708718474, + "tokens_seen": 1887830016 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021618856569709127, + "loss": 1.3879, + "theoretical_loss": 3.4754931801898308, + "tokens_seen": 1887895552 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021617853560682045, + "loss": 1.2443, + "theoretical_loss": 3.4754830899561697, + "tokens_seen": 1887961088 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021616850551654966, + "loss": 1.3058, + "theoretical_loss": 3.4754730001708287, + "tokens_seen": 1888026624 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021615847542627884, + "loss": 1.3039, + "theoretical_loss": 3.475462910833773, + "tokens_seen": 1888092160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021614844533600802, + "loss": 1.3388, + "theoretical_loss": 3.475452821944967, + "tokens_seen": 1888157696 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002161384152457372, + "loss": 1.4096, + "theoretical_loss": 3.4754427335043743, + "tokens_seen": 1888223232 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021612838515546641, + "loss": 1.2844, + "theoretical_loss": 3.47543264551196, + "tokens_seen": 1888288768 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002161183550651956, + "loss": 1.3663, + "theoretical_loss": 3.475422557967689, + "tokens_seen": 1888354304 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021610832497492478, + "loss": 1.3364, + "theoretical_loss": 3.4754124708715257, + "tokens_seen": 1888419840 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021609829488465396, + "loss": 1.3761, + "theoretical_loss": 3.4754023842234343, + "tokens_seen": 1888485376 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021608826479438314, + "loss": 1.1944, + "theoretical_loss": 3.4753922980233796, + "tokens_seen": 1888550912 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021607823470411235, + "loss": 1.3012, + "theoretical_loss": 3.4753822122713256, + "tokens_seen": 1888616448 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021606820461384153, + "loss": 1.2244, + "theoretical_loss": 3.4753721269672373, + "tokens_seen": 1888681984 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002160581745235707, + "loss": 1.2614, + "theoretical_loss": 3.4753620421110796, + "tokens_seen": 1888747520 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002160481444332999, + "loss": 1.3249, + "theoretical_loss": 3.4753519577028165, + "tokens_seen": 1888813056 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002160381143430291, + "loss": 1.2625, + "theoretical_loss": 3.475341873742413, + "tokens_seen": 1888878592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021602808425275828, + "loss": 1.3634, + "theoretical_loss": 3.4753317902298337, + "tokens_seen": 1888944128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021601805416248746, + "loss": 1.3125, + "theoretical_loss": 3.4753217071650426, + "tokens_seen": 1889009664 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021600802407221664, + "loss": 1.2591, + "theoretical_loss": 3.475311624548005, + "tokens_seen": 1889075200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021599799398194582, + "loss": 1.333, + "theoretical_loss": 3.4753015423786846, + "tokens_seen": 1889140736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021598796389167503, + "loss": 1.3821, + "theoretical_loss": 3.475291460657047, + "tokens_seen": 1889206272 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002159779338014042, + "loss": 1.3095, + "theoretical_loss": 3.475281379383056, + "tokens_seen": 1889271808 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002159679037111334, + "loss": 1.435, + "theoretical_loss": 3.4752712985566765, + "tokens_seen": 1889337344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021595787362086257, + "loss": 1.359, + "theoretical_loss": 3.475261218177873, + "tokens_seen": 1889402880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021594784353059178, + "loss": 1.2298, + "theoretical_loss": 3.4752511382466107, + "tokens_seen": 1889468416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021593781344032096, + "loss": 1.3136, + "theoretical_loss": 3.4752410587628533, + "tokens_seen": 1889533952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021592778335005014, + "loss": 1.4008, + "theoretical_loss": 3.475230979726566, + "tokens_seen": 1889599488 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021591775325977933, + "loss": 1.3388, + "theoretical_loss": 3.475220901137713, + "tokens_seen": 1889665024 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002159077231695085, + "loss": 1.3604, + "theoretical_loss": 3.4752108229962593, + "tokens_seen": 1889730560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021589769307923772, + "loss": 1.3266, + "theoretical_loss": 3.4752007453021694, + "tokens_seen": 1889796096 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002158876629889669, + "loss": 1.4167, + "theoretical_loss": 3.475190668055408, + "tokens_seen": 1889861632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021587763289869608, + "loss": 1.3317, + "theoretical_loss": 3.4751805912559393, + "tokens_seen": 1889927168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021586760280842526, + "loss": 1.4112, + "theoretical_loss": 3.475170514903729, + "tokens_seen": 1889992704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021585757271815447, + "loss": 1.4453, + "theoretical_loss": 3.4751604389987403, + "tokens_seen": 1890058240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021584754262788365, + "loss": 1.2077, + "theoretical_loss": 3.475150363540939, + "tokens_seen": 1890123776 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021583751253761286, + "loss": 1.3053, + "theoretical_loss": 3.475140288530289, + "tokens_seen": 1890189312 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021582748244734204, + "loss": 1.3666, + "theoretical_loss": 3.4751302139667555, + "tokens_seen": 1890254848 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021581745235707122, + "loss": 1.3485, + "theoretical_loss": 3.4751201398503024, + "tokens_seen": 1890320384 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021580742226680043, + "loss": 1.2838, + "theoretical_loss": 3.4751100661808954, + "tokens_seen": 1890385920 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002157973921765296, + "loss": 1.383, + "theoretical_loss": 3.4750999929584987, + "tokens_seen": 1890451456 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002157873620862588, + "loss": 1.2557, + "theoretical_loss": 3.475089920183077, + "tokens_seen": 1890516992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021577733199598797, + "loss": 1.4885, + "theoretical_loss": 3.4750798478545946, + "tokens_seen": 1890582528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021576730190571718, + "loss": 1.3049, + "theoretical_loss": 3.4750697759730165, + "tokens_seen": 1890648064 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 689504, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3121987581253052, + "objective/train/theoretical_loss": 3.4750597045383076, + "objective/train/tokens_used": 1911173600, + "theoretical_loss": 3.4750597045383076, + "tokens_seen": 1890713600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021575727181544636, + "loss": 1.2902, + "theoretical_loss": 3.4750597045383076, + "tokens_seen": 1890713600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021574724172517554, + "loss": 1.2779, + "theoretical_loss": 3.475049633550432, + "tokens_seen": 1890779136 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021573721163490472, + "loss": 1.2515, + "theoretical_loss": 3.475039563009355, + "tokens_seen": 1890844672 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002157271815446339, + "loss": 1.2901, + "theoretical_loss": 3.4750294929150405, + "tokens_seen": 1890910208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002157171514543631, + "loss": 1.4063, + "theoretical_loss": 3.4750194232674545, + "tokens_seen": 1890975744 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002157071213640923, + "loss": 1.2715, + "theoretical_loss": 3.4750093540665605, + "tokens_seen": 1891041280 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021569709127382147, + "loss": 1.3124, + "theoretical_loss": 3.4749992853123235, + "tokens_seen": 1891106816 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021568706118355065, + "loss": 1.3241, + "theoretical_loss": 3.4749892170047083, + "tokens_seen": 1891172352 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021567703109327986, + "loss": 1.4295, + "theoretical_loss": 3.47497914914368, + "tokens_seen": 1891237888 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021566700100300904, + "loss": 1.3791, + "theoretical_loss": 3.474969081729203, + "tokens_seen": 1891303424 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021565697091273822, + "loss": 1.3315, + "theoretical_loss": 3.4749590147612417, + "tokens_seen": 1891368960 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002156469408224674, + "loss": 1.3124, + "theoretical_loss": 3.4749489482397613, + "tokens_seen": 1891434496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021563691073219661, + "loss": 1.3262, + "theoretical_loss": 3.474938882164726, + "tokens_seen": 1891500032 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002156268806419258, + "loss": 1.3907, + "theoretical_loss": 3.474928816536101, + "tokens_seen": 1891565568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021561685055165498, + "loss": 1.4082, + "theoretical_loss": 3.4749187513538513, + "tokens_seen": 1891631104 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021560682046138416, + "loss": 1.2903, + "theoretical_loss": 3.4749086866179413, + "tokens_seen": 1891696640 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021559679037111334, + "loss": 1.3269, + "theoretical_loss": 3.474898622328335, + "tokens_seen": 1891762176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021558676028084255, + "loss": 1.306, + "theoretical_loss": 3.4748885584849987, + "tokens_seen": 1891827712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021557673019057173, + "loss": 1.3398, + "theoretical_loss": 3.4748784950878955, + "tokens_seen": 1891893248 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002155667001003009, + "loss": 1.3798, + "theoretical_loss": 3.474868432136991, + "tokens_seen": 1891958784 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002155566700100301, + "loss": 1.3387, + "theoretical_loss": 3.4748583696322504, + "tokens_seen": 1892024320 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002155466399197593, + "loss": 1.3423, + "theoretical_loss": 3.474848307573638, + "tokens_seen": 1892089856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021553660982948848, + "loss": 1.3284, + "theoretical_loss": 3.4748382459611182, + "tokens_seen": 1892155392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021552657973921766, + "loss": 1.3618, + "theoretical_loss": 3.474828184794656, + "tokens_seen": 1892220928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021551654964894684, + "loss": 1.4241, + "theoretical_loss": 3.4748181240742166, + "tokens_seen": 1892286464 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021550651955867602, + "loss": 1.309, + "theoretical_loss": 3.474808063799764, + "tokens_seen": 1892352000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021549648946840523, + "loss": 1.4028, + "theoretical_loss": 3.474798003971264, + "tokens_seen": 1892417536 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002154864593781344, + "loss": 1.3651, + "theoretical_loss": 3.474787944588681, + "tokens_seen": 1892483072 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002154764292878636, + "loss": 1.4097, + "theoretical_loss": 3.4747778856519793, + "tokens_seen": 1892548608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021546639919759277, + "loss": 1.4156, + "theoretical_loss": 3.4747678271611244, + "tokens_seen": 1892614144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021545636910732198, + "loss": 1.321, + "theoretical_loss": 3.4747577691160805, + "tokens_seen": 1892679680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021544633901705116, + "loss": 1.2806, + "theoretical_loss": 3.4747477115168124, + "tokens_seen": 1892745216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021543630892678034, + "loss": 1.367, + "theoretical_loss": 3.4747376543632855, + "tokens_seen": 1892810752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021542627883650953, + "loss": 1.3842, + "theoretical_loss": 3.474727597655464, + "tokens_seen": 1892876288 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002154162487462387, + "loss": 1.2462, + "theoretical_loss": 3.4747175413933133, + "tokens_seen": 1892941824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021540621865596792, + "loss": 1.3008, + "theoretical_loss": 3.4747074855767974, + "tokens_seen": 1893007360 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002153961885656971, + "loss": 1.3303, + "theoretical_loss": 3.4746974302058824, + "tokens_seen": 1893072896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021538615847542628, + "loss": 1.3201, + "theoretical_loss": 3.4746873752805314, + "tokens_seen": 1893138432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021537612838515546, + "loss": 1.3796, + "theoretical_loss": 3.474677320800711, + "tokens_seen": 1893203968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021536609829488467, + "loss": 1.3183, + "theoretical_loss": 3.474667266766385, + "tokens_seen": 1893269504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021535606820461385, + "loss": 1.4123, + "theoretical_loss": 3.4746572131775184, + "tokens_seen": 1893335040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021534603811434303, + "loss": 1.3967, + "theoretical_loss": 3.4746471600340763, + "tokens_seen": 1893400576 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002153360080240722, + "loss": 1.3553, + "theoretical_loss": 3.474637107336023, + "tokens_seen": 1893466112 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002153259779338014, + "loss": 1.3906, + "theoretical_loss": 3.474627055083324, + "tokens_seen": 1893531648 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002153159478435306, + "loss": 1.3288, + "theoretical_loss": 3.4746170032759442, + "tokens_seen": 1893597184 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021530591775325978, + "loss": 1.3236, + "theoretical_loss": 3.474606951913848, + "tokens_seen": 1893662720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021529588766298896, + "loss": 1.3184, + "theoretical_loss": 3.4745969009969997, + "tokens_seen": 1893728256 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021528585757271814, + "loss": 1.4139, + "theoretical_loss": 3.4745868505253656, + "tokens_seen": 1893793792 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021527582748244735, + "loss": 1.2955, + "theoretical_loss": 3.4745768004989097, + "tokens_seen": 1893859328 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021526579739217653, + "loss": 1.3742, + "theoretical_loss": 3.474566750917597, + "tokens_seen": 1893924864 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 690600, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.388242244720459, + "objective/train/theoretical_loss": 3.474556701781392, + "objective/train/tokens_used": 1914450400, + "theoretical_loss": 3.474556701781392, + "tokens_seen": 1893990400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002152557673019057, + "loss": 1.3243, + "theoretical_loss": 3.474556701781392, + "tokens_seen": 1893990400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002152457372116349, + "loss": 1.4105, + "theoretical_loss": 3.4745466530902607, + "tokens_seen": 1894055936 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021523570712136408, + "loss": 1.3609, + "theoretical_loss": 3.474536604844167, + "tokens_seen": 1894121472 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021522567703109328, + "loss": 1.2704, + "theoretical_loss": 3.4745265570430757, + "tokens_seen": 1894187008 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021521564694082247, + "loss": 1.4417, + "theoretical_loss": 3.4745165096869526, + "tokens_seen": 1894252544 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021520561685055165, + "loss": 1.4339, + "theoretical_loss": 3.4745064627757616, + "tokens_seen": 1894318080 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021519558676028083, + "loss": 1.3065, + "theoretical_loss": 3.4744964163094685, + "tokens_seen": 1894383616 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021518555667001004, + "loss": 1.3832, + "theoretical_loss": 3.4744863702880378, + "tokens_seen": 1894449152 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021517552657973922, + "loss": 1.2842, + "theoretical_loss": 3.474476324711434, + "tokens_seen": 1894514688 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002151654964894684, + "loss": 1.4119, + "theoretical_loss": 3.474466279579623, + "tokens_seen": 1894580224 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021515546639919758, + "loss": 1.3519, + "theoretical_loss": 3.4744562348925685, + "tokens_seen": 1894645760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021514543630892676, + "loss": 1.3056, + "theoretical_loss": 3.4744461906502364, + "tokens_seen": 1894711296 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021513540621865597, + "loss": 1.3579, + "theoretical_loss": 3.4744361468525913, + "tokens_seen": 1894776832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021512537612838515, + "loss": 1.3075, + "theoretical_loss": 3.474426103499598, + "tokens_seen": 1894842368 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021511534603811433, + "loss": 1.3377, + "theoretical_loss": 3.4744160605912215, + "tokens_seen": 1894907904 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002151053159478435, + "loss": 1.2391, + "theoretical_loss": 3.4744060181274268, + "tokens_seen": 1894973440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021509528585757272, + "loss": 1.3314, + "theoretical_loss": 3.474395976108179, + "tokens_seen": 1895038976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021508525576730193, + "loss": 1.3978, + "theoretical_loss": 3.474385934533443, + "tokens_seen": 1895104512 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002150752256770311, + "loss": 1.2754, + "theoretical_loss": 3.4743758934031836, + "tokens_seen": 1895170048 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002150651955867603, + "loss": 1.402, + "theoretical_loss": 3.4743658527173658, + "tokens_seen": 1895235584 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002150551654964895, + "loss": 1.3416, + "theoretical_loss": 3.474355812475954, + "tokens_seen": 1895301120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021504513540621868, + "loss": 1.2812, + "theoretical_loss": 3.4743457726789146, + "tokens_seen": 1895366656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021503510531594786, + "loss": 1.3853, + "theoretical_loss": 3.4743357333262113, + "tokens_seen": 1895432192 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021502507522567704, + "loss": 1.3384, + "theoretical_loss": 3.4743256944178094, + "tokens_seen": 1895497728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021501504513540622, + "loss": 1.3132, + "theoretical_loss": 3.4743156559536743, + "tokens_seen": 1895563264 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021500501504513543, + "loss": 1.3769, + "theoretical_loss": 3.47430561793377, + "tokens_seen": 1895628800 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002149949849548646, + "loss": 1.4479, + "theoretical_loss": 3.4742955803580626, + "tokens_seen": 1895694336 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002149849548645938, + "loss": 1.4853, + "theoretical_loss": 3.4742855432265163, + "tokens_seen": 1895759872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021497492477432297, + "loss": 1.3554, + "theoretical_loss": 3.4742755065390964, + "tokens_seen": 1895825408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021496489468405218, + "loss": 1.4348, + "theoretical_loss": 3.4742654702957685, + "tokens_seen": 1895890944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021495486459378136, + "loss": 1.2856, + "theoretical_loss": 3.4742554344964964, + "tokens_seen": 1895956480 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021494483450351055, + "loss": 1.3374, + "theoretical_loss": 3.474245399141246, + "tokens_seen": 1896022016 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021493480441323973, + "loss": 1.2958, + "theoretical_loss": 3.4742353642299815, + "tokens_seen": 1896087552 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002149247743229689, + "loss": 1.4149, + "theoretical_loss": 3.474225329762669, + "tokens_seen": 1896153088 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021491474423269812, + "loss": 1.2716, + "theoretical_loss": 3.4742152957392722, + "tokens_seen": 1896218624 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002149047141424273, + "loss": 1.3256, + "theoretical_loss": 3.4742052621597574, + "tokens_seen": 1896284160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021489468405215648, + "loss": 1.356, + "theoretical_loss": 3.4741952290240894, + "tokens_seen": 1896349696 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021488465396188566, + "loss": 1.3838, + "theoretical_loss": 3.4741851963322326, + "tokens_seen": 1896415232 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021487462387161487, + "loss": 1.4629, + "theoretical_loss": 3.4741751640841523, + "tokens_seen": 1896480768 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021486459378134405, + "loss": 1.3132, + "theoretical_loss": 3.474165132279813, + "tokens_seen": 1896546304 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021485456369107323, + "loss": 1.2846, + "theoretical_loss": 3.474155100919181, + "tokens_seen": 1896611840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002148445336008024, + "loss": 1.3619, + "theoretical_loss": 3.47414507000222, + "tokens_seen": 1896677376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002148345035105316, + "loss": 1.2511, + "theoretical_loss": 3.4741350395288966, + "tokens_seen": 1896742912 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002148244734202608, + "loss": 1.3657, + "theoretical_loss": 3.4741250094991747, + "tokens_seen": 1896808448 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021481444332998998, + "loss": 1.4069, + "theoretical_loss": 3.4741149799130193, + "tokens_seen": 1896873984 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021480441323971916, + "loss": 1.305, + "theoretical_loss": 3.474104950770396, + "tokens_seen": 1896939520 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021479438314944834, + "loss": 1.3134, + "theoretical_loss": 3.474094922071269, + "tokens_seen": 1897005056 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021478435305917755, + "loss": 1.45, + "theoretical_loss": 3.474084893815605, + "tokens_seen": 1897070592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021477432296890673, + "loss": 1.337, + "theoretical_loss": 3.4740748660033676, + "tokens_seen": 1897136128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021476429287863591, + "loss": 1.4609, + "theoretical_loss": 3.474064838634522, + "tokens_seen": 1897201664 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 691800, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.450380563735962, + "objective/train/theoretical_loss": 3.474054811709034, + "objective/train/tokens_used": 1917727200, + "theoretical_loss": 3.474054811709034, + "tokens_seen": 1897267200 + }, + { + "epoch": 0.57, + "learning_rate": 0.0002147542627883651, + "loss": 1.3867, + "theoretical_loss": 3.474054811709034, + "tokens_seen": 1897267200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021474423269809428, + "loss": 1.4208, + "theoretical_loss": 3.4740447852268685, + "tokens_seen": 1897332736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021473420260782348, + "loss": 1.3721, + "theoretical_loss": 3.4740347591879903, + "tokens_seen": 1897398272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00021472417251755267, + "loss": 1.2857, + "theoretical_loss": 3.4740247335923646, + "tokens_seen": 1897463808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021471414242728185, + "loss": 1.3696, + "theoretical_loss": 3.474014708439956, + "tokens_seen": 1897529344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021470411233701103, + "loss": 1.3538, + "theoretical_loss": 3.4740046837307306, + "tokens_seen": 1897594880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021469408224674024, + "loss": 1.3495, + "theoretical_loss": 3.473994659464653, + "tokens_seen": 1897660416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021468405215646942, + "loss": 1.3405, + "theoretical_loss": 3.4739846356416875, + "tokens_seen": 1897725952 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002146740220661986, + "loss": 1.2991, + "theoretical_loss": 3.473974612261801, + "tokens_seen": 1897791488 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021466399197592778, + "loss": 1.2195, + "theoretical_loss": 3.4739645893249573, + "tokens_seen": 1897857024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021465396188565696, + "loss": 1.4212, + "theoretical_loss": 3.4739545668311216, + "tokens_seen": 1897922560 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021464393179538617, + "loss": 1.2297, + "theoretical_loss": 3.4739445447802595, + "tokens_seen": 1897988096 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021463390170511535, + "loss": 1.3483, + "theoretical_loss": 3.473934523172336, + "tokens_seen": 1898053632 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021462387161484453, + "loss": 1.3765, + "theoretical_loss": 3.473924502007316, + "tokens_seen": 1898119168 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002146138415245737, + "loss": 1.3325, + "theoretical_loss": 3.4739144812851643, + "tokens_seen": 1898184704 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021460381143430292, + "loss": 1.217, + "theoretical_loss": 3.473904461005847, + "tokens_seen": 1898250240 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002145937813440321, + "loss": 1.2924, + "theoretical_loss": 3.473894441169329, + "tokens_seen": 1898315776 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021458375125376128, + "loss": 1.3077, + "theoretical_loss": 3.4738844217755744, + "tokens_seen": 1898381312 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021457372116349046, + "loss": 1.3341, + "theoretical_loss": 3.47387440282455, + "tokens_seen": 1898446848 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021456369107321964, + "loss": 1.3386, + "theoretical_loss": 3.4738643843162196, + "tokens_seen": 1898512384 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021455366098294885, + "loss": 1.4443, + "theoretical_loss": 3.4738543662505488, + "tokens_seen": 1898577920 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021454363089267803, + "loss": 1.3473, + "theoretical_loss": 3.473844348627503, + "tokens_seen": 1898643456 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021453360080240721, + "loss": 1.334, + "theoretical_loss": 3.473834331447047, + "tokens_seen": 1898708992 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002145235707121364, + "loss": 1.2504, + "theoretical_loss": 3.4738243147091463, + "tokens_seen": 1898774528 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002145135406218656, + "loss": 1.261, + "theoretical_loss": 3.4738142984137657, + "tokens_seen": 1898840064 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021450351053159479, + "loss": 1.3052, + "theoretical_loss": 3.473804282560871, + "tokens_seen": 1898905600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021449348044132397, + "loss": 1.3263, + "theoretical_loss": 3.473794267150427, + "tokens_seen": 1898971136 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021448345035105315, + "loss": 1.2401, + "theoretical_loss": 3.4737842521823983, + "tokens_seen": 1899036672 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021447342026078236, + "loss": 1.3626, + "theoretical_loss": 3.473774237656751, + "tokens_seen": 1899102208 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021446339017051154, + "loss": 1.305, + "theoretical_loss": 3.47376422357345, + "tokens_seen": 1899167744 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021445336008024072, + "loss": 1.4543, + "theoretical_loss": 3.4737542099324603, + "tokens_seen": 1899233280 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002144433299899699, + "loss": 1.3476, + "theoretical_loss": 3.4737441967337475, + "tokens_seen": 1899298816 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021443329989969908, + "loss": 1.3679, + "theoretical_loss": 3.4737341839772764, + "tokens_seen": 1899364352 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002144232698094283, + "loss": 1.3583, + "theoretical_loss": 3.4737241716630125, + "tokens_seen": 1899429888 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021441323971915747, + "loss": 1.2821, + "theoretical_loss": 3.4737141597909207, + "tokens_seen": 1899495424 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021440320962888665, + "loss": 1.314, + "theoretical_loss": 3.4737041483609663, + "tokens_seen": 1899560960 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021439317953861583, + "loss": 1.2614, + "theoretical_loss": 3.4736941373731147, + "tokens_seen": 1899626496 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021438314944834504, + "loss": 1.3222, + "theoretical_loss": 3.4736841268273313, + "tokens_seen": 1899692032 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021437311935807422, + "loss": 1.4171, + "theoretical_loss": 3.473674116723581, + "tokens_seen": 1899757568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002143630892678034, + "loss": 1.3257, + "theoretical_loss": 3.473664107061829, + "tokens_seen": 1899823104 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021435305917753258, + "loss": 1.2759, + "theoretical_loss": 3.4736540978420405, + "tokens_seen": 1899888640 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021434302908726176, + "loss": 1.3075, + "theoretical_loss": 3.473644089064181, + "tokens_seen": 1899954176 + }, + { + "epoch": 0.58, + "learning_rate": 0.000214332998996991, + "loss": 1.4134, + "theoretical_loss": 3.473634080728216, + "tokens_seen": 1900019712 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021432296890672018, + "loss": 1.327, + "theoretical_loss": 3.47362407283411, + "tokens_seen": 1900085248 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021431293881644936, + "loss": 1.3882, + "theoretical_loss": 3.4736140653818284, + "tokens_seen": 1900150784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021430290872617854, + "loss": 1.3576, + "theoretical_loss": 3.473604058371337, + "tokens_seen": 1900216320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021429287863590775, + "loss": 1.3166, + "theoretical_loss": 3.4735940518026007, + "tokens_seen": 1900281856 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021428284854563693, + "loss": 1.339, + "theoretical_loss": 3.473584045675585, + "tokens_seen": 1900347392 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021427281845536611, + "loss": 1.3958, + "theoretical_loss": 3.473574039990255, + "tokens_seen": 1900412928 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002142627883650953, + "loss": 1.2866, + "theoretical_loss": 3.473564034746576, + "tokens_seen": 1900478464 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 693007, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2872929573059082, + "objective/train/theoretical_loss": 3.4735540299445127, + "objective/train/tokens_used": 1921004000, + "theoretical_loss": 3.4735540299445127, + "tokens_seen": 1900544000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021425275827482448, + "loss": 1.3148, + "theoretical_loss": 3.4735540299445127, + "tokens_seen": 1900544000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021424272818455368, + "loss": 1.2842, + "theoretical_loss": 3.4735440255840313, + "tokens_seen": 1900609536 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021423269809428287, + "loss": 1.2593, + "theoretical_loss": 3.4735340216650967, + "tokens_seen": 1900675072 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021422266800401205, + "loss": 1.348, + "theoretical_loss": 3.4735240181876743, + "tokens_seen": 1900740608 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021421263791374123, + "loss": 1.3733, + "theoretical_loss": 3.473514015151729, + "tokens_seen": 1900806144 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021420260782347044, + "loss": 1.3967, + "theoretical_loss": 3.4735040125572265, + "tokens_seen": 1900871680 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021419257773319962, + "loss": 1.3448, + "theoretical_loss": 3.4734940104041323, + "tokens_seen": 1900937216 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002141825476429288, + "loss": 1.2998, + "theoretical_loss": 3.473484008692411, + "tokens_seen": 1901002752 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021417251755265798, + "loss": 1.2868, + "theoretical_loss": 3.4734740074220283, + "tokens_seen": 1901068288 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021416248746238716, + "loss": 1.3566, + "theoretical_loss": 3.4734640065929496, + "tokens_seen": 1901133824 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021415245737211637, + "loss": 1.3264, + "theoretical_loss": 3.47345400620514, + "tokens_seen": 1901199360 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021414242728184555, + "loss": 1.356, + "theoretical_loss": 3.473444006258565, + "tokens_seen": 1901264896 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021413239719157473, + "loss": 1.3326, + "theoretical_loss": 3.47343400675319, + "tokens_seen": 1901330432 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002141223671013039, + "loss": 1.345, + "theoretical_loss": 3.47342400768898, + "tokens_seen": 1901395968 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021411233701103312, + "loss": 1.3949, + "theoretical_loss": 3.473414009065901, + "tokens_seen": 1901461504 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002141023069207623, + "loss": 1.3759, + "theoretical_loss": 3.4734040108839173, + "tokens_seen": 1901527040 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021409227683049148, + "loss": 1.3234, + "theoretical_loss": 3.473394013142995, + "tokens_seen": 1901592576 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021408224674022066, + "loss": 1.2921, + "theoretical_loss": 3.4733840158430995, + "tokens_seen": 1901658112 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021407221664994984, + "loss": 1.3942, + "theoretical_loss": 3.4733740189841953, + "tokens_seen": 1901723648 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021406218655967905, + "loss": 1.3367, + "theoretical_loss": 3.4733640225662485, + "tokens_seen": 1901789184 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021405215646940823, + "loss": 1.4175, + "theoretical_loss": 3.4733540265892247, + "tokens_seen": 1901854720 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021404212637913742, + "loss": 1.3696, + "theoretical_loss": 3.473344031053089, + "tokens_seen": 1901920256 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002140320962888666, + "loss": 1.3548, + "theoretical_loss": 3.473334035957806, + "tokens_seen": 1901985792 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002140220661985958, + "loss": 1.4047, + "theoretical_loss": 3.473324041303342, + "tokens_seen": 1902051328 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021401203610832499, + "loss": 1.3349, + "theoretical_loss": 3.4733140470896617, + "tokens_seen": 1902116864 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021400200601805417, + "loss": 1.3984, + "theoretical_loss": 3.473304053316731, + "tokens_seen": 1902182400 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021399197592778335, + "loss": 1.2565, + "theoretical_loss": 3.4732940599845152, + "tokens_seen": 1902247936 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021398194583751256, + "loss": 1.4081, + "theoretical_loss": 3.4732840670929797, + "tokens_seen": 1902313472 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021397191574724174, + "loss": 1.2926, + "theoretical_loss": 3.47327407464209, + "tokens_seen": 1902379008 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021396188565697092, + "loss": 1.4563, + "theoretical_loss": 3.4732640826318106, + "tokens_seen": 1902444544 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002139518555667001, + "loss": 1.4159, + "theoretical_loss": 3.473254091062108, + "tokens_seen": 1902510080 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021394182547642928, + "loss": 1.3747, + "theoretical_loss": 3.473244099932947, + "tokens_seen": 1902575616 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002139317953861585, + "loss": 1.3479, + "theoretical_loss": 3.4732341092442933, + "tokens_seen": 1902641152 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021392176529588767, + "loss": 1.3338, + "theoretical_loss": 3.4732241189961117, + "tokens_seen": 1902706688 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021391173520561685, + "loss": 1.3035, + "theoretical_loss": 3.473214129188369, + "tokens_seen": 1902772224 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021390170511534603, + "loss": 1.3345, + "theoretical_loss": 3.473204139821029, + "tokens_seen": 1902837760 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021389167502507524, + "loss": 1.3572, + "theoretical_loss": 3.473194150894058, + "tokens_seen": 1902903296 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021388164493480442, + "loss": 1.4477, + "theoretical_loss": 3.473184162407421, + "tokens_seen": 1902968832 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002138716148445336, + "loss": 1.342, + "theoretical_loss": 3.473174174361084, + "tokens_seen": 1903034368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021386158475426278, + "loss": 1.2402, + "theoretical_loss": 3.473164186755012, + "tokens_seen": 1903099904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021385155466399196, + "loss": 1.3925, + "theoretical_loss": 3.47315419958917, + "tokens_seen": 1903165440 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021384152457372117, + "loss": 1.299, + "theoretical_loss": 3.4731442128635246, + "tokens_seen": 1903230976 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021383149448345035, + "loss": 1.4969, + "theoretical_loss": 3.4731342265780407, + "tokens_seen": 1903296512 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021382146439317954, + "loss": 1.4231, + "theoretical_loss": 3.473124240732683, + "tokens_seen": 1903362048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021381143430290872, + "loss": 1.4431, + "theoretical_loss": 3.473114255327418, + "tokens_seen": 1903427584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021380140421263792, + "loss": 1.2664, + "theoretical_loss": 3.473104270362211, + "tokens_seen": 1903493120 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002137913741223671, + "loss": 1.4412, + "theoretical_loss": 3.4730942858370266, + "tokens_seen": 1903558656 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002137813440320963, + "loss": 1.3569, + "theoretical_loss": 3.473084301751831, + "tokens_seen": 1903624192 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021377131394182547, + "loss": 1.296, + "theoretical_loss": 3.47307431810659, + "tokens_seen": 1903689728 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021376128385155465, + "loss": 1.3779, + "theoretical_loss": 3.4730643349012684, + "tokens_seen": 1903755264 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 694210, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3409249782562256, + "objective/train/theoretical_loss": 3.4730543521358315, + "objective/train/tokens_used": 1924280800, + "theoretical_loss": 3.4730543521358315, + "tokens_seen": 1903820800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021375125376128386, + "loss": 1.3548, + "theoretical_loss": 3.4730543521358315, + "tokens_seen": 1903820800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021374122367101304, + "loss": 1.3955, + "theoretical_loss": 3.4730443698102453, + "tokens_seen": 1903886336 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021373119358074222, + "loss": 1.3501, + "theoretical_loss": 3.473034387924475, + "tokens_seen": 1903951872 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002137211634904714, + "loss": 1.2338, + "theoretical_loss": 3.473024406478487, + "tokens_seen": 1904017408 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002137111334002006, + "loss": 1.3649, + "theoretical_loss": 3.473014425472245, + "tokens_seen": 1904082944 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002137011033099298, + "loss": 1.3358, + "theoretical_loss": 3.4730044449057162, + "tokens_seen": 1904148480 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021369107321965897, + "loss": 1.4045, + "theoretical_loss": 3.472994464778865, + "tokens_seen": 1904214016 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021368104312938815, + "loss": 1.3671, + "theoretical_loss": 3.472984485091657, + "tokens_seen": 1904279552 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021367101303911733, + "loss": 1.4048, + "theoretical_loss": 3.4729745058440584, + "tokens_seen": 1904345088 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021366098294884654, + "loss": 1.275, + "theoretical_loss": 3.4729645270360345, + "tokens_seen": 1904410624 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021365095285857572, + "loss": 1.3499, + "theoretical_loss": 3.47295454866755, + "tokens_seen": 1904476160 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002136409227683049, + "loss": 1.2978, + "theoretical_loss": 3.472944570738571, + "tokens_seen": 1904541696 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021363089267803409, + "loss": 1.369, + "theoretical_loss": 3.4729345932490636, + "tokens_seen": 1904607232 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002136208625877633, + "loss": 1.441, + "theoretical_loss": 3.472924616198993, + "tokens_seen": 1904672768 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021361083249749247, + "loss": 1.3001, + "theoretical_loss": 3.472914639588324, + "tokens_seen": 1904738304 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021360080240722166, + "loss": 1.3228, + "theoretical_loss": 3.4729046634170224, + "tokens_seen": 1904803840 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021359077231695084, + "loss": 1.4209, + "theoretical_loss": 3.4728946876850544, + "tokens_seen": 1904869376 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021358074222668004, + "loss": 1.3993, + "theoretical_loss": 3.4728847123923847, + "tokens_seen": 1904934912 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021357071213640925, + "loss": 1.245, + "theoretical_loss": 3.4728747375389792, + "tokens_seen": 1905000448 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021356068204613843, + "loss": 1.3969, + "theoretical_loss": 3.4728647631248037, + "tokens_seen": 1905065984 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021355065195586762, + "loss": 1.3192, + "theoretical_loss": 3.4728547891498236, + "tokens_seen": 1905131520 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002135406218655968, + "loss": 1.4041, + "theoretical_loss": 3.4728448156140046, + "tokens_seen": 1905197056 + }, + { + "epoch": 0.58, + "learning_rate": 0.000213530591775326, + "loss": 1.3288, + "theoretical_loss": 3.4728348425173117, + "tokens_seen": 1905262592 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021352056168505519, + "loss": 1.312, + "theoretical_loss": 3.4728248698597106, + "tokens_seen": 1905328128 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021351053159478437, + "loss": 1.3423, + "theoretical_loss": 3.4728148976411677, + "tokens_seen": 1905393664 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021350050150451355, + "loss": 1.4117, + "theoretical_loss": 3.472804925861648, + "tokens_seen": 1905459200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021349047141424276, + "loss": 1.3983, + "theoretical_loss": 3.472794954521116, + "tokens_seen": 1905524736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021348044132397194, + "loss": 1.408, + "theoretical_loss": 3.472784983619539, + "tokens_seen": 1905590272 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021347041123370112, + "loss": 1.3137, + "theoretical_loss": 3.4727750131568818, + "tokens_seen": 1905655808 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002134603811434303, + "loss": 1.3498, + "theoretical_loss": 3.47276504313311, + "tokens_seen": 1905721344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021345035105315948, + "loss": 1.4528, + "theoretical_loss": 3.472755073548189, + "tokens_seen": 1905786880 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002134403209628887, + "loss": 1.3492, + "theoretical_loss": 3.472745104402085, + "tokens_seen": 1905852416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021343029087261787, + "loss": 1.3401, + "theoretical_loss": 3.472735135694763, + "tokens_seen": 1905917952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021342026078234705, + "loss": 1.2804, + "theoretical_loss": 3.472725167426189, + "tokens_seen": 1905983488 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021341023069207623, + "loss": 1.299, + "theoretical_loss": 3.472715199596329, + "tokens_seen": 1906049024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021340020060180544, + "loss": 1.4426, + "theoretical_loss": 3.472705232205147, + "tokens_seen": 1906114560 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021339017051153462, + "loss": 1.3919, + "theoretical_loss": 3.4726952652526104, + "tokens_seen": 1906180096 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002133801404212638, + "loss": 1.384, + "theoretical_loss": 3.472685298738684, + "tokens_seen": 1906245632 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021337011033099298, + "loss": 1.3765, + "theoretical_loss": 3.472675332663333, + "tokens_seen": 1906311168 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021336008024072216, + "loss": 1.3712, + "theoretical_loss": 3.472665367026524, + "tokens_seen": 1906376704 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021335005015045137, + "loss": 1.2638, + "theoretical_loss": 3.472655401828222, + "tokens_seen": 1906442240 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021334002006018055, + "loss": 1.2552, + "theoretical_loss": 3.4726454370683926, + "tokens_seen": 1906507776 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021332998996990974, + "loss": 1.3356, + "theoretical_loss": 3.472635472747002, + "tokens_seen": 1906573312 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021331995987963892, + "loss": 1.4591, + "theoretical_loss": 3.4726255088640148, + "tokens_seen": 1906638848 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021330992978936812, + "loss": 1.2745, + "theoretical_loss": 3.472615545419398, + "tokens_seen": 1906704384 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002132998996990973, + "loss": 1.2872, + "theoretical_loss": 3.472605582413116, + "tokens_seen": 1906769920 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002132898696088265, + "loss": 1.4284, + "theoretical_loss": 3.4725956198451353, + "tokens_seen": 1906835456 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021327983951855567, + "loss": 1.4176, + "theoretical_loss": 3.4725856577154213, + "tokens_seen": 1906900992 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021326980942828485, + "loss": 1.3347, + "theoretical_loss": 3.47257569602394, + "tokens_seen": 1906966528 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021325977933801406, + "loss": 1.3481, + "theoretical_loss": 3.4725657347706558, + "tokens_seen": 1907032064 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 695522, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3152399063110352, + "objective/train/theoretical_loss": 3.4725557739555355, + "objective/train/tokens_used": 1927557600, + "theoretical_loss": 3.4725557739555355, + "tokens_seen": 1907097600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021324974924774324, + "loss": 1.3092, + "theoretical_loss": 3.4725557739555355, + "tokens_seen": 1907097600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021323971915747242, + "loss": 1.2988, + "theoretical_loss": 3.4725458135785447, + "tokens_seen": 1907163136 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002132296890672016, + "loss": 1.2503, + "theoretical_loss": 3.4725358536396485, + "tokens_seen": 1907228672 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002132196589769308, + "loss": 1.4069, + "theoretical_loss": 3.4725258941388137, + "tokens_seen": 1907294208 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021320962888666, + "loss": 1.3004, + "theoretical_loss": 3.4725159350760046, + "tokens_seen": 1907359744 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021319959879638917, + "loss": 1.318, + "theoretical_loss": 3.472505976451188, + "tokens_seen": 1907425280 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021318956870611835, + "loss": 1.2649, + "theoretical_loss": 3.472496018264329, + "tokens_seen": 1907490816 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021317953861584753, + "loss": 1.3652, + "theoretical_loss": 3.4724860605153935, + "tokens_seen": 1907556352 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021316950852557674, + "loss": 1.2779, + "theoretical_loss": 3.4724761032043467, + "tokens_seen": 1907621888 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021315947843530592, + "loss": 1.3598, + "theoretical_loss": 3.472466146331155, + "tokens_seen": 1907687424 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002131494483450351, + "loss": 1.3593, + "theoretical_loss": 3.4724561898957838, + "tokens_seen": 1907752960 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021313941825476429, + "loss": 1.2505, + "theoretical_loss": 3.4724462338981983, + "tokens_seen": 1907818496 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002131293881644935, + "loss": 1.3692, + "theoretical_loss": 3.4724362783383658, + "tokens_seen": 1907884032 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021311935807422267, + "loss": 1.4142, + "theoretical_loss": 3.47242632321625, + "tokens_seen": 1907949568 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021310932798395186, + "loss": 1.254, + "theoretical_loss": 3.4724163685318183, + "tokens_seen": 1908015104 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021309929789368104, + "loss": 1.3492, + "theoretical_loss": 3.472406414285035, + "tokens_seen": 1908080640 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021308926780341022, + "loss": 1.3191, + "theoretical_loss": 3.472396460475867, + "tokens_seen": 1908146176 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021307923771313943, + "loss": 1.3348, + "theoretical_loss": 3.4723865071042797, + "tokens_seen": 1908211712 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002130692076228686, + "loss": 1.283, + "theoretical_loss": 3.4723765541702383, + "tokens_seen": 1908277248 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002130591775325978, + "loss": 1.43, + "theoretical_loss": 3.472366601673709, + "tokens_seen": 1908342784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021304914744232697, + "loss": 1.2465, + "theoretical_loss": 3.4723566496146576, + "tokens_seen": 1908408320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021303911735205618, + "loss": 1.3611, + "theoretical_loss": 3.47234669799305, + "tokens_seen": 1908473856 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021302908726178536, + "loss": 1.3382, + "theoretical_loss": 3.4723367468088515, + "tokens_seen": 1908539392 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021301905717151454, + "loss": 1.3041, + "theoretical_loss": 3.4723267960620277, + "tokens_seen": 1908604928 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021300902708124372, + "loss": 1.4197, + "theoretical_loss": 3.4723168457525446, + "tokens_seen": 1908670464 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002129989969909729, + "loss": 1.2405, + "theoretical_loss": 3.4723068958803687, + "tokens_seen": 1908736000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002129889669007021, + "loss": 1.3297, + "theoretical_loss": 3.4722969464454647, + "tokens_seen": 1908801536 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002129789368104313, + "loss": 1.4466, + "theoretical_loss": 3.472286997447799, + "tokens_seen": 1908867072 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021296890672016047, + "loss": 1.3323, + "theoretical_loss": 3.472277048887337, + "tokens_seen": 1908932608 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021295887662988965, + "loss": 1.384, + "theoretical_loss": 3.4722671007640447, + "tokens_seen": 1908998144 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021294884653961886, + "loss": 1.3467, + "theoretical_loss": 3.472257153077888, + "tokens_seen": 1909063680 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021293881644934804, + "loss": 1.2836, + "theoretical_loss": 3.4722472058288325, + "tokens_seen": 1909129216 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021292878635907722, + "loss": 1.3955, + "theoretical_loss": 3.472237259016844, + "tokens_seen": 1909194752 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002129187562688064, + "loss": 1.2953, + "theoretical_loss": 3.472227312641888, + "tokens_seen": 1909260288 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002129087261785356, + "loss": 1.3973, + "theoretical_loss": 3.472217366703931, + "tokens_seen": 1909325824 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002128986960882648, + "loss": 1.3892, + "theoretical_loss": 3.4722074212029383, + "tokens_seen": 1909391360 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021288866599799398, + "loss": 1.4083, + "theoretical_loss": 3.4721974761388754, + "tokens_seen": 1909456896 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021287863590772316, + "loss": 1.4575, + "theoretical_loss": 3.4721875315117092, + "tokens_seen": 1909522432 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021286860581745234, + "loss": 1.3176, + "theoretical_loss": 3.4721775873214042, + "tokens_seen": 1909587968 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021285857572718155, + "loss": 1.3385, + "theoretical_loss": 3.472167643567927, + "tokens_seen": 1909653504 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021284854563691073, + "loss": 1.3341, + "theoretical_loss": 3.472157700251244, + "tokens_seen": 1909719040 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002128385155466399, + "loss": 1.4032, + "theoretical_loss": 3.472147757371319, + "tokens_seen": 1909784576 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021282848545636912, + "loss": 1.2737, + "theoretical_loss": 3.47213781492812, + "tokens_seen": 1909850112 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021281845536609832, + "loss": 1.4034, + "theoretical_loss": 3.472127872921612, + "tokens_seen": 1909915648 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002128084252758275, + "loss": 1.4137, + "theoretical_loss": 3.4721179313517605, + "tokens_seen": 1909981184 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002127983951855567, + "loss": 1.3336, + "theoretical_loss": 3.4721079902185314, + "tokens_seen": 1910046720 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021278836509528587, + "loss": 1.2634, + "theoretical_loss": 3.4720980495218914, + "tokens_seen": 1910112256 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021277833500501505, + "loss": 1.3093, + "theoretical_loss": 3.4720881092618052, + "tokens_seen": 1910177792 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021276830491474426, + "loss": 1.4021, + "theoretical_loss": 3.472078169438239, + "tokens_seen": 1910243328 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021275827482447344, + "loss": 1.3303, + "theoretical_loss": 3.4720682300511596, + "tokens_seen": 1910308864 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 696705, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2651183605194092, + "objective/train/theoretical_loss": 3.4720582911005318, + "objective/train/tokens_used": 1930834400, + "theoretical_loss": 3.4720582911005318, + "tokens_seen": 1910374400 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021274824473420262, + "loss": 1.2936, + "theoretical_loss": 3.4720582911005318, + "tokens_seen": 1910374400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002127382146439318, + "loss": 1.3455, + "theoretical_loss": 3.4720483525863215, + "tokens_seen": 1910439936 + }, + { + "epoch": 0.58, + "learning_rate": 0.000212728184553661, + "loss": 1.3662, + "theoretical_loss": 3.472038414508495, + "tokens_seen": 1910505472 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002127181544633902, + "loss": 1.3784, + "theoretical_loss": 3.4720284768670178, + "tokens_seen": 1910571008 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021270812437311937, + "loss": 1.2853, + "theoretical_loss": 3.4720185396618564, + "tokens_seen": 1910636544 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021269809428284855, + "loss": 1.2227, + "theoretical_loss": 3.472008602892976, + "tokens_seen": 1910702080 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021268806419257773, + "loss": 1.3105, + "theoretical_loss": 3.4719986665603426, + "tokens_seen": 1910767616 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021267803410230694, + "loss": 1.4032, + "theoretical_loss": 3.471988730663922, + "tokens_seen": 1910833152 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021266800401203612, + "loss": 1.4713, + "theoretical_loss": 3.4719787952036807, + "tokens_seen": 1910898688 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002126579739217653, + "loss": 1.3601, + "theoretical_loss": 3.471968860179584, + "tokens_seen": 1910964224 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021264794383149449, + "loss": 1.3869, + "theoretical_loss": 3.4719589255915984, + "tokens_seen": 1911029760 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002126379137412237, + "loss": 1.4518, + "theoretical_loss": 3.471948991439689, + "tokens_seen": 1911095296 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021262788365095287, + "loss": 1.3985, + "theoretical_loss": 3.471939057723822, + "tokens_seen": 1911160832 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021261785356068206, + "loss": 1.3293, + "theoretical_loss": 3.471929124443964, + "tokens_seen": 1911226368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021260782347041124, + "loss": 1.3944, + "theoretical_loss": 3.4719191916000796, + "tokens_seen": 1911291904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021259779338014042, + "loss": 1.3841, + "theoretical_loss": 3.471909259192136, + "tokens_seen": 1911357440 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021258776328986963, + "loss": 1.3498, + "theoretical_loss": 3.471899327220098, + "tokens_seen": 1911422976 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002125777331995988, + "loss": 1.3916, + "theoretical_loss": 3.4718893956839327, + "tokens_seen": 1911488512 + }, + { + "epoch": 0.58, + "learning_rate": 0.000212567703109328, + "loss": 1.2997, + "theoretical_loss": 3.471879464583605, + "tokens_seen": 1911554048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021255767301905717, + "loss": 1.3817, + "theoretical_loss": 3.4718695339190813, + "tokens_seen": 1911619584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021254764292878638, + "loss": 1.2221, + "theoretical_loss": 3.4718596036903273, + "tokens_seen": 1911685120 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021253761283851556, + "loss": 1.3373, + "theoretical_loss": 3.4718496738973097, + "tokens_seen": 1911750656 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021252758274824474, + "loss": 1.305, + "theoretical_loss": 3.4718397445399933, + "tokens_seen": 1911816192 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021251755265797392, + "loss": 1.3249, + "theoretical_loss": 3.471829815618345, + "tokens_seen": 1911881728 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002125075225677031, + "loss": 1.3349, + "theoretical_loss": 3.4718198871323303, + "tokens_seen": 1911947264 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002124974924774323, + "loss": 1.3394, + "theoretical_loss": 3.471809959081915, + "tokens_seen": 1912012800 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002124874623871615, + "loss": 1.536, + "theoretical_loss": 3.4718000314670654, + "tokens_seen": 1912078336 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021247743229689067, + "loss": 1.2669, + "theoretical_loss": 3.4717901042877477, + "tokens_seen": 1912143872 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021246740220661985, + "loss": 1.2887, + "theoretical_loss": 3.471780177543927, + "tokens_seen": 1912209408 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021245737211634906, + "loss": 1.4342, + "theoretical_loss": 3.47177025123557, + "tokens_seen": 1912274944 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021244734202607824, + "loss": 1.4605, + "theoretical_loss": 3.4717603253626423, + "tokens_seen": 1912340480 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021243731193580742, + "loss": 1.2995, + "theoretical_loss": 3.47175039992511, + "tokens_seen": 1912406016 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002124272818455366, + "loss": 1.4855, + "theoretical_loss": 3.4717404749229397, + "tokens_seen": 1912471552 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002124172517552658, + "loss": 1.315, + "theoretical_loss": 3.471730550356096, + "tokens_seen": 1912537088 + }, + { + "epoch": 0.58, + "learning_rate": 0.000212407221664995, + "loss": 1.352, + "theoretical_loss": 3.4717206262245464, + "tokens_seen": 1912602624 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021239719157472418, + "loss": 1.2283, + "theoretical_loss": 3.471710702528256, + "tokens_seen": 1912668160 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021238716148445336, + "loss": 1.4019, + "theoretical_loss": 3.4717007792671906, + "tokens_seen": 1912733696 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021237713139418254, + "loss": 1.3678, + "theoretical_loss": 3.4716908564413167, + "tokens_seen": 1912799232 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021236710130391175, + "loss": 1.3253, + "theoretical_loss": 3.4716809340506, + "tokens_seen": 1912864768 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021235707121364093, + "loss": 1.3347, + "theoretical_loss": 3.4716710120950074, + "tokens_seen": 1912930304 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002123470411233701, + "loss": 1.2811, + "theoretical_loss": 3.4716610905745036, + "tokens_seen": 1912995840 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002123370110330993, + "loss": 1.4127, + "theoretical_loss": 3.4716511694890553, + "tokens_seen": 1913061376 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002123269809428285, + "loss": 1.3407, + "theoretical_loss": 3.471641248838629, + "tokens_seen": 1913126912 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021231695085255768, + "loss": 1.3494, + "theoretical_loss": 3.4716313286231895, + "tokens_seen": 1913192448 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021230692076228686, + "loss": 1.15, + "theoretical_loss": 3.4716214088427035, + "tokens_seen": 1913257984 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021229689067201604, + "loss": 1.278, + "theoretical_loss": 3.4716114894971373, + "tokens_seen": 1913323520 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021228686058174522, + "loss": 1.4025, + "theoretical_loss": 3.471601570586456, + "tokens_seen": 1913389056 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021227683049147443, + "loss": 1.3745, + "theoretical_loss": 3.471591652110627, + "tokens_seen": 1913454592 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002122668004012036, + "loss": 1.4792, + "theoretical_loss": 3.471581734069616, + "tokens_seen": 1913520128 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002122567703109328, + "loss": 1.3528, + "theoretical_loss": 3.471571816463388, + "tokens_seen": 1913585664 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 697951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4669880867004395, + "objective/train/theoretical_loss": 3.4715618992919097, + "objective/train/tokens_used": 1934111200, + "theoretical_loss": 3.4715618992919097, + "tokens_seen": 1913651200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021224674022066197, + "loss": 1.3922, + "theoretical_loss": 3.4715618992919097, + "tokens_seen": 1913651200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021223671013039118, + "loss": 1.2825, + "theoretical_loss": 3.4715519825551477, + "tokens_seen": 1913716736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021222668004012036, + "loss": 1.3387, + "theoretical_loss": 3.4715420662530674, + "tokens_seen": 1913782272 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021221664994984954, + "loss": 1.3778, + "theoretical_loss": 3.471532150385635, + "tokens_seen": 1913847808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021220661985957873, + "loss": 1.3735, + "theoretical_loss": 3.4715222349528165, + "tokens_seen": 1913913344 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002121965897693079, + "loss": 1.3543, + "theoretical_loss": 3.471512319954578, + "tokens_seen": 1913978880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021218655967903712, + "loss": 1.3459, + "theoretical_loss": 3.4715024053908854, + "tokens_seen": 1914044416 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002121765295887663, + "loss": 1.3446, + "theoretical_loss": 3.4714924912617056, + "tokens_seen": 1914109952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021216649949849548, + "loss": 1.4007, + "theoretical_loss": 3.471482577567004, + "tokens_seen": 1914175488 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021215646940822466, + "loss": 1.2691, + "theoretical_loss": 3.4714726643067464, + "tokens_seen": 1914241024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021214643931795387, + "loss": 1.2435, + "theoretical_loss": 3.4714627514808996, + "tokens_seen": 1914306560 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021213640922768305, + "loss": 1.4177, + "theoretical_loss": 3.4714528390894293, + "tokens_seen": 1914372096 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021212637913741223, + "loss": 1.439, + "theoretical_loss": 3.4714429271323013, + "tokens_seen": 1914437632 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002121163490471414, + "loss": 1.4491, + "theoretical_loss": 3.471433015609482, + "tokens_seen": 1914503168 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002121063189568706, + "loss": 1.3131, + "theoretical_loss": 3.4714231045209383, + "tokens_seen": 1914568704 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002120962888665998, + "loss": 1.3323, + "theoretical_loss": 3.471413193866635, + "tokens_seen": 1914634240 + }, + { + "epoch": 0.58, + "learning_rate": 0.000212086258776329, + "loss": 1.2046, + "theoretical_loss": 3.4714032836465387, + "tokens_seen": 1914699776 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002120762286860582, + "loss": 1.4183, + "theoretical_loss": 3.4713933738606157, + "tokens_seen": 1914765312 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021206619859578737, + "loss": 1.3912, + "theoretical_loss": 3.4713834645088317, + "tokens_seen": 1914830848 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021205616850551658, + "loss": 1.3888, + "theoretical_loss": 3.471373555591154, + "tokens_seen": 1914896384 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021204613841524576, + "loss": 1.4248, + "theoretical_loss": 3.4713636471075473, + "tokens_seen": 1914961920 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021203610832497494, + "loss": 1.3626, + "theoretical_loss": 3.471353739057978, + "tokens_seen": 1915027456 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021202607823470412, + "loss": 1.3012, + "theoretical_loss": 3.4713438314424128, + "tokens_seen": 1915092992 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002120160481444333, + "loss": 1.3358, + "theoretical_loss": 3.4713339242608177, + "tokens_seen": 1915158528 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002120060180541625, + "loss": 1.4374, + "theoretical_loss": 3.471324017513159, + "tokens_seen": 1915224064 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002119959879638917, + "loss": 1.407, + "theoretical_loss": 3.471314111199402, + "tokens_seen": 1915289600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021198595787362087, + "loss": 1.3534, + "theoretical_loss": 3.4713042053195133, + "tokens_seen": 1915355136 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021197592778335005, + "loss": 1.3546, + "theoretical_loss": 3.4712942998734597, + "tokens_seen": 1915420672 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021196589769307926, + "loss": 1.3482, + "theoretical_loss": 3.4712843948612067, + "tokens_seen": 1915486208 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021195586760280844, + "loss": 1.3009, + "theoretical_loss": 3.47127449028272, + "tokens_seen": 1915551744 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021194583751253762, + "loss": 1.3777, + "theoretical_loss": 3.4712645861379667, + "tokens_seen": 1915617280 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002119358074222668, + "loss": 1.3091, + "theoretical_loss": 3.4712546824269124, + "tokens_seen": 1915682816 + }, + { + "epoch": 0.58, + "learning_rate": 0.000211925777331996, + "loss": 1.3922, + "theoretical_loss": 3.471244779149524, + "tokens_seen": 1915748352 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002119157472417252, + "loss": 1.3677, + "theoretical_loss": 3.4712348763057665, + "tokens_seen": 1915813888 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021190571715145438, + "loss": 1.3396, + "theoretical_loss": 3.471224973895607, + "tokens_seen": 1915879424 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021189568706118356, + "loss": 1.3173, + "theoretical_loss": 3.471215071919012, + "tokens_seen": 1915944960 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021188565697091274, + "loss": 1.2666, + "theoretical_loss": 3.4712051703759466, + "tokens_seen": 1916010496 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021187562688064195, + "loss": 1.1941, + "theoretical_loss": 3.4711952692663774, + "tokens_seen": 1916076032 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021186559679037113, + "loss": 1.3774, + "theoretical_loss": 3.471185368590271, + "tokens_seen": 1916141568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002118555667001003, + "loss": 1.3354, + "theoretical_loss": 3.4711754683475933, + "tokens_seen": 1916207104 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002118455366098295, + "loss": 1.2814, + "theoretical_loss": 3.47116556853831, + "tokens_seen": 1916272640 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002118355065195587, + "loss": 1.4224, + "theoretical_loss": 3.471155669162388, + "tokens_seen": 1916338176 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021182547642928788, + "loss": 1.3547, + "theoretical_loss": 3.4711457702197936, + "tokens_seen": 1916403712 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021181544633901706, + "loss": 1.3034, + "theoretical_loss": 3.4711358717104925, + "tokens_seen": 1916469248 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021180541624874624, + "loss": 1.393, + "theoretical_loss": 3.471125973634451, + "tokens_seen": 1916534784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021179538615847542, + "loss": 1.208, + "theoretical_loss": 3.4711160759916355, + "tokens_seen": 1916600320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021178535606820463, + "loss": 1.3082, + "theoretical_loss": 3.471106178782012, + "tokens_seen": 1916665856 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002117753259779338, + "loss": 1.3613, + "theoretical_loss": 3.4710962820055475, + "tokens_seen": 1916731392 + }, + { + "epoch": 0.58, + "learning_rate": 0.000211765295887663, + "loss": 1.4163, + "theoretical_loss": 3.4710863856622076, + "tokens_seen": 1916796928 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021175526579739217, + "loss": 1.2597, + "theoretical_loss": 3.471076489751958, + "tokens_seen": 1916862464 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 699035, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2874658107757568, + "objective/train/theoretical_loss": 3.471066594274766, + "objective/train/tokens_used": 1937388000, + "theoretical_loss": 3.471066594274766, + "tokens_seen": 1916928000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021174523570712138, + "loss": 1.3714, + "theoretical_loss": 3.471066594274766, + "tokens_seen": 1916928000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021173520561685056, + "loss": 1.3745, + "theoretical_loss": 3.4710566992305973, + "tokens_seen": 1916993536 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021172517552657974, + "loss": 1.3507, + "theoretical_loss": 3.4710468046194185, + "tokens_seen": 1917059072 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021171514543630893, + "loss": 1.4415, + "theoretical_loss": 3.471036910441195, + "tokens_seen": 1917124608 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002117051153460381, + "loss": 1.2858, + "theoretical_loss": 3.471027016695894, + "tokens_seen": 1917190144 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021169508525576732, + "loss": 1.2562, + "theoretical_loss": 3.471017123383481, + "tokens_seen": 1917255680 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002116850551654965, + "loss": 1.3388, + "theoretical_loss": 3.4710072305039232, + "tokens_seen": 1917321216 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021167502507522568, + "loss": 1.3508, + "theoretical_loss": 3.470997338057186, + "tokens_seen": 1917386752 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021166499498495486, + "loss": 1.3893, + "theoretical_loss": 3.470987446043236, + "tokens_seen": 1917452288 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021165496489468407, + "loss": 1.3551, + "theoretical_loss": 3.4709775544620394, + "tokens_seen": 1917517824 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021164493480441325, + "loss": 1.3189, + "theoretical_loss": 3.4709676633135627, + "tokens_seen": 1917583360 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021163490471414243, + "loss": 1.3242, + "theoretical_loss": 3.470957772597772, + "tokens_seen": 1917648896 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002116248746238716, + "loss": 1.4647, + "theoretical_loss": 3.4709478823146336, + "tokens_seen": 1917714432 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002116148445336008, + "loss": 1.3606, + "theoretical_loss": 3.4709379924641137, + "tokens_seen": 1917779968 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021160481444333, + "loss": 1.4518, + "theoretical_loss": 3.470928103046179, + "tokens_seen": 1917845504 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021159478435305918, + "loss": 1.3808, + "theoretical_loss": 3.470918214060795, + "tokens_seen": 1917911040 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021158475426278836, + "loss": 1.332, + "theoretical_loss": 3.4709083255079287, + "tokens_seen": 1917976576 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021157472417251754, + "loss": 1.2877, + "theoretical_loss": 3.4708984373875467, + "tokens_seen": 1918042112 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021156469408224675, + "loss": 1.243, + "theoretical_loss": 3.4708885496996142, + "tokens_seen": 1918107648 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021155466399197593, + "loss": 1.4339, + "theoretical_loss": 3.4708786624440986, + "tokens_seen": 1918173184 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002115446339017051, + "loss": 1.3245, + "theoretical_loss": 3.470868775620965, + "tokens_seen": 1918238720 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002115346038114343, + "loss": 1.2517, + "theoretical_loss": 3.470858889230181, + "tokens_seen": 1918304256 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021152457372116348, + "loss": 1.4537, + "theoretical_loss": 3.470849003271712, + "tokens_seen": 1918369792 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021151454363089268, + "loss": 1.2972, + "theoretical_loss": 3.470839117745525, + "tokens_seen": 1918435328 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021150451354062186, + "loss": 1.3714, + "theoretical_loss": 3.470829232651586, + "tokens_seen": 1918500864 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021149448345035105, + "loss": 1.2849, + "theoretical_loss": 3.4708193479898615, + "tokens_seen": 1918566400 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021148445336008023, + "loss": 1.311, + "theoretical_loss": 3.4708094637603177, + "tokens_seen": 1918631936 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021147442326980944, + "loss": 1.3939, + "theoretical_loss": 3.4707995799629208, + "tokens_seen": 1918697472 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021146439317953862, + "loss": 1.4049, + "theoretical_loss": 3.4707896965976373, + "tokens_seen": 1918763008 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002114543630892678, + "loss": 1.32, + "theoretical_loss": 3.470779813664434, + "tokens_seen": 1918828544 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021144433299899698, + "loss": 1.3041, + "theoretical_loss": 3.4707699311632765, + "tokens_seen": 1918894080 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021143430290872616, + "loss": 1.3492, + "theoretical_loss": 3.4707600490941313, + "tokens_seen": 1918959616 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021142427281845537, + "loss": 1.3581, + "theoretical_loss": 3.470750167456965, + "tokens_seen": 1919025152 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021141424272818455, + "loss": 1.3438, + "theoretical_loss": 3.4707402862517442, + "tokens_seen": 1919090688 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021140421263791373, + "loss": 1.2852, + "theoretical_loss": 3.4707304054784345, + "tokens_seen": 1919156224 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002113941825476429, + "loss": 1.2656, + "theoretical_loss": 3.470720525137003, + "tokens_seen": 1919221760 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021138415245737212, + "loss": 1.3048, + "theoretical_loss": 3.4707106452274155, + "tokens_seen": 1919287296 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002113741223671013, + "loss": 1.3331, + "theoretical_loss": 3.4707007657496387, + "tokens_seen": 1919352832 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021136409227683048, + "loss": 1.3129, + "theoretical_loss": 3.4706908867036397, + "tokens_seen": 1919418368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021135406218655966, + "loss": 1.3055, + "theoretical_loss": 3.4706810080893833, + "tokens_seen": 1919483904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021134403209628884, + "loss": 1.4652, + "theoretical_loss": 3.4706711299068376, + "tokens_seen": 1919549440 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021133400200601808, + "loss": 1.3061, + "theoretical_loss": 3.4706612521559674, + "tokens_seen": 1919614976 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021132397191574726, + "loss": 1.3579, + "theoretical_loss": 3.47065137483674, + "tokens_seen": 1919680512 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021131394182547644, + "loss": 1.2352, + "theoretical_loss": 3.4706414979491216, + "tokens_seen": 1919746048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021130391173520562, + "loss": 1.3549, + "theoretical_loss": 3.4706316214930792, + "tokens_seen": 1919811584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021129388164493483, + "loss": 1.3732, + "theoretical_loss": 3.470621745468578, + "tokens_seen": 1919877120 + }, + { + "epoch": 0.58, + "learning_rate": 0.000211283851554664, + "loss": 1.398, + "theoretical_loss": 3.4706118698755857, + "tokens_seen": 1919942656 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002112738214643932, + "loss": 1.3515, + "theoretical_loss": 3.470601994714068, + "tokens_seen": 1920008192 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021126379137412237, + "loss": 1.3649, + "theoretical_loss": 3.4705921199839906, + "tokens_seen": 1920073728 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021125376128385158, + "loss": 1.3722, + "theoretical_loss": 3.4705822456853213, + "tokens_seen": 1920139264 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 700307, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4499931335449219, + "objective/train/theoretical_loss": 3.4705723718180264, + "objective/train/tokens_used": 1940664800, + "theoretical_loss": 3.4705723718180264, + "tokens_seen": 1920204800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021124373119358076, + "loss": 1.4546, + "theoretical_loss": 3.4705723718180264, + "tokens_seen": 1920204800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021123370110330994, + "loss": 1.3496, + "theoretical_loss": 3.4705624983820718, + "tokens_seen": 1920270336 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021122367101303913, + "loss": 1.3427, + "theoretical_loss": 3.4705526253774233, + "tokens_seen": 1920335872 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002112136409227683, + "loss": 1.3578, + "theoretical_loss": 3.470542752804049, + "tokens_seen": 1920401408 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021120361083249752, + "loss": 1.3496, + "theoretical_loss": 3.4705328806619136, + "tokens_seen": 1920466944 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002111935807422267, + "loss": 1.3733, + "theoretical_loss": 3.470523008950985, + "tokens_seen": 1920532480 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021118355065195588, + "loss": 1.2646, + "theoretical_loss": 3.470513137671229, + "tokens_seen": 1920598016 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021117352056168506, + "loss": 1.2597, + "theoretical_loss": 3.470503266822612, + "tokens_seen": 1920663552 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021116349047141427, + "loss": 1.4385, + "theoretical_loss": 3.4704933964051, + "tokens_seen": 1920729088 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021115346038114345, + "loss": 1.2598, + "theoretical_loss": 3.470483526418661, + "tokens_seen": 1920794624 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021114343029087263, + "loss": 1.2523, + "theoretical_loss": 3.47047365686326, + "tokens_seen": 1920860160 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002111334002006018, + "loss": 1.4006, + "theoretical_loss": 3.470463787738864, + "tokens_seen": 1920925696 + }, + { + "epoch": 0.58, + "learning_rate": 0.000211123370110331, + "loss": 1.3235, + "theoretical_loss": 3.470453919045439, + "tokens_seen": 1920991232 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002111133400200602, + "loss": 1.3918, + "theoretical_loss": 3.4704440507829526, + "tokens_seen": 1921056768 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021110330992978938, + "loss": 1.1956, + "theoretical_loss": 3.4704341829513696, + "tokens_seen": 1921122304 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021109327983951856, + "loss": 1.4225, + "theoretical_loss": 3.4704243155506584, + "tokens_seen": 1921187840 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021108324974924774, + "loss": 1.4437, + "theoretical_loss": 3.470414448580784, + "tokens_seen": 1921253376 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021107321965897695, + "loss": 1.2672, + "theoretical_loss": 3.4704045820417138, + "tokens_seen": 1921318912 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021106318956870613, + "loss": 1.3123, + "theoretical_loss": 3.470394715933414, + "tokens_seen": 1921384448 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002110531594784353, + "loss": 1.2863, + "theoretical_loss": 3.4703848502558508, + "tokens_seen": 1921449984 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002110431293881645, + "loss": 1.3701, + "theoretical_loss": 3.4703749850089913, + "tokens_seen": 1921515520 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021103309929789368, + "loss": 1.4364, + "theoretical_loss": 3.4703651201928016, + "tokens_seen": 1921581056 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021102306920762288, + "loss": 1.2219, + "theoretical_loss": 3.470355255807248, + "tokens_seen": 1921646592 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021101303911735207, + "loss": 1.3562, + "theoretical_loss": 3.4703453918522973, + "tokens_seen": 1921712128 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021100300902708125, + "loss": 1.2571, + "theoretical_loss": 3.470335528327916, + "tokens_seen": 1921777664 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021099297893681043, + "loss": 1.3353, + "theoretical_loss": 3.470325665234071, + "tokens_seen": 1921843200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021098294884653964, + "loss": 1.3391, + "theoretical_loss": 3.4703158025707284, + "tokens_seen": 1921908736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021097291875626882, + "loss": 1.3521, + "theoretical_loss": 3.4703059403378544, + "tokens_seen": 1921974272 + }, + { + "epoch": 0.58, + "learning_rate": 0.000210962888665998, + "loss": 1.3236, + "theoretical_loss": 3.4702960785354158, + "tokens_seen": 1922039808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021095285857572718, + "loss": 1.3909, + "theoretical_loss": 3.4702862171633795, + "tokens_seen": 1922105344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021094282848545636, + "loss": 1.3965, + "theoretical_loss": 3.4702763562217123, + "tokens_seen": 1922170880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021093279839518557, + "loss": 1.225, + "theoretical_loss": 3.47026649571038, + "tokens_seen": 1922236416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021092276830491475, + "loss": 1.3167, + "theoretical_loss": 3.470256635629349, + "tokens_seen": 1922301952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021091273821464393, + "loss": 1.367, + "theoretical_loss": 3.470246775978586, + "tokens_seen": 1922367488 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002109027081243731, + "loss": 1.3701, + "theoretical_loss": 3.4702369167580587, + "tokens_seen": 1922433024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021089267803410232, + "loss": 1.3462, + "theoretical_loss": 3.4702270579677323, + "tokens_seen": 1922498560 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002108826479438315, + "loss": 1.3126, + "theoretical_loss": 3.470217199607574, + "tokens_seen": 1922564096 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021087261785356068, + "loss": 1.3796, + "theoretical_loss": 3.4702073416775496, + "tokens_seen": 1922629632 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021086258776328986, + "loss": 1.4246, + "theoretical_loss": 3.470197484177627, + "tokens_seen": 1922695168 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021085255767301904, + "loss": 1.4165, + "theoretical_loss": 3.470187627107772, + "tokens_seen": 1922760704 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021084252758274825, + "loss": 1.3853, + "theoretical_loss": 3.470177770467951, + "tokens_seen": 1922826240 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021083249749247743, + "loss": 1.4093, + "theoretical_loss": 3.4701679142581305, + "tokens_seen": 1922891776 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021082246740220661, + "loss": 1.2591, + "theoretical_loss": 3.470158058478278, + "tokens_seen": 1922957312 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002108124373119358, + "loss": 1.3472, + "theoretical_loss": 3.4701482031283595, + "tokens_seen": 1923022848 + }, + { + "epoch": 0.58, + "learning_rate": 0.000210802407221665, + "loss": 1.3173, + "theoretical_loss": 3.4701383482083408, + "tokens_seen": 1923088384 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021079237713139419, + "loss": 1.3226, + "theoretical_loss": 3.4701284937181898, + "tokens_seen": 1923153920 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021078234704112337, + "loss": 1.3909, + "theoretical_loss": 3.4701186396578727, + "tokens_seen": 1923219456 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021077231695085255, + "loss": 1.3509, + "theoretical_loss": 3.470108786027356, + "tokens_seen": 1923284992 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021076228686058176, + "loss": 1.5008, + "theoretical_loss": 3.4700989328266063, + "tokens_seen": 1923350528 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021075225677031094, + "loss": 1.361, + "theoretical_loss": 3.4700890800555895, + "tokens_seen": 1923416064 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 701572, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.334864854812622, + "objective/train/theoretical_loss": 3.470079227714274, + "objective/train/tokens_used": 1943941600, + "theoretical_loss": 3.470079227714274, + "tokens_seen": 1923481600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021074222668004012, + "loss": 1.3418, + "theoretical_loss": 3.470079227714274, + "tokens_seen": 1923481600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002107321965897693, + "loss": 1.2662, + "theoretical_loss": 3.4700693758026246, + "tokens_seen": 1923547136 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021072216649949848, + "loss": 1.3647, + "theoretical_loss": 3.470059524320609, + "tokens_seen": 1923612672 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002107121364092277, + "loss": 1.3438, + "theoretical_loss": 3.4700496732681936, + "tokens_seen": 1923678208 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021070210631895687, + "loss": 1.3147, + "theoretical_loss": 3.4700398226453446, + "tokens_seen": 1923743744 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021069207622868605, + "loss": 1.3505, + "theoretical_loss": 3.470029972452029, + "tokens_seen": 1923809280 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021068204613841523, + "loss": 1.542, + "theoretical_loss": 3.4700201226882132, + "tokens_seen": 1923874816 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021067201604814444, + "loss": 1.2685, + "theoretical_loss": 3.4700102733538642, + "tokens_seen": 1923940352 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021066198595787362, + "loss": 1.3568, + "theoretical_loss": 3.470000424448949, + "tokens_seen": 1924005888 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002106519558676028, + "loss": 1.3805, + "theoretical_loss": 3.4699905759734335, + "tokens_seen": 1924071424 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021064192577733198, + "loss": 1.3073, + "theoretical_loss": 3.4699807279272843, + "tokens_seen": 1924136960 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021063189568706116, + "loss": 1.3622, + "theoretical_loss": 3.4699708803104685, + "tokens_seen": 1924202496 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021062186559679037, + "loss": 1.2602, + "theoretical_loss": 3.469961033122953, + "tokens_seen": 1924268032 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021061183550651955, + "loss": 1.2279, + "theoretical_loss": 3.4699511863647032, + "tokens_seen": 1924333568 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021060180541624873, + "loss": 1.316, + "theoretical_loss": 3.469941340035687, + "tokens_seen": 1924399104 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021059177532597792, + "loss": 1.2918, + "theoretical_loss": 3.4699314941358708, + "tokens_seen": 1924464640 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021058174523570715, + "loss": 1.2883, + "theoretical_loss": 3.4699216486652213, + "tokens_seen": 1924530176 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021057171514543633, + "loss": 1.3112, + "theoretical_loss": 3.469911803623705, + "tokens_seen": 1924595712 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021056168505516551, + "loss": 1.411, + "theoretical_loss": 3.469901959011289, + "tokens_seen": 1924661248 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002105516549648947, + "loss": 1.1528, + "theoretical_loss": 3.4698921148279394, + "tokens_seen": 1924726784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021054162487462388, + "loss": 1.3487, + "theoretical_loss": 3.4698822710736232, + "tokens_seen": 1924792320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021053159478435308, + "loss": 1.3064, + "theoretical_loss": 3.4698724277483066, + "tokens_seen": 1924857856 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021052156469408227, + "loss": 1.3114, + "theoretical_loss": 3.469862584851957, + "tokens_seen": 1924923392 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021051153460381145, + "loss": 1.3406, + "theoretical_loss": 3.469852742384541, + "tokens_seen": 1924988928 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021050150451354063, + "loss": 1.41, + "theoretical_loss": 3.469842900346025, + "tokens_seen": 1925054464 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021049147442326984, + "loss": 1.3049, + "theoretical_loss": 3.469833058736376, + "tokens_seen": 1925120000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021048144433299902, + "loss": 1.3669, + "theoretical_loss": 3.4698232175555606, + "tokens_seen": 1925185536 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002104714142427282, + "loss": 1.3185, + "theoretical_loss": 3.469813376803545, + "tokens_seen": 1925251072 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021046138415245738, + "loss": 1.2739, + "theoretical_loss": 3.469803536480297, + "tokens_seen": 1925316608 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021045135406218656, + "loss": 1.3226, + "theoretical_loss": 3.469793696585782, + "tokens_seen": 1925382144 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021044132397191577, + "loss": 1.4093, + "theoretical_loss": 3.4697838571199684, + "tokens_seen": 1925447680 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021043129388164495, + "loss": 1.2752, + "theoretical_loss": 3.469774018082821, + "tokens_seen": 1925513216 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021042126379137413, + "loss": 1.2948, + "theoretical_loss": 3.4697641794743084, + "tokens_seen": 1925578752 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002104112337011033, + "loss": 1.3017, + "theoretical_loss": 3.469754341294396, + "tokens_seen": 1925644288 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021040120361083252, + "loss": 1.3026, + "theoretical_loss": 3.469744503543051, + "tokens_seen": 1925709824 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002103911735205617, + "loss": 1.3447, + "theoretical_loss": 3.46973466622024, + "tokens_seen": 1925775360 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021038114343029088, + "loss": 1.3611, + "theoretical_loss": 3.4697248293259304, + "tokens_seen": 1925840896 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021037111334002006, + "loss": 1.407, + "theoretical_loss": 3.469714992860088, + "tokens_seen": 1925906432 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021036108324974924, + "loss": 1.3847, + "theoretical_loss": 3.4697051568226804, + "tokens_seen": 1925971968 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021035105315947845, + "loss": 1.3503, + "theoretical_loss": 3.4696953212136736, + "tokens_seen": 1926037504 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021034102306920763, + "loss": 1.2697, + "theoretical_loss": 3.469685486033035, + "tokens_seen": 1926103040 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021033099297893681, + "loss": 1.3265, + "theoretical_loss": 3.469675651280731, + "tokens_seen": 1926168576 + }, + { + "epoch": 0.58, + "learning_rate": 0.000210320962888666, + "loss": 1.2776, + "theoretical_loss": 3.4696658169567285, + "tokens_seen": 1926234112 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002103109327983952, + "loss": 1.3526, + "theoretical_loss": 3.469655983060994, + "tokens_seen": 1926299648 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021030090270812439, + "loss": 1.3251, + "theoretical_loss": 3.469646149593495, + "tokens_seen": 1926365184 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021029087261785357, + "loss": 1.3099, + "theoretical_loss": 3.4696363165541975, + "tokens_seen": 1926430720 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021028084252758275, + "loss": 1.3518, + "theoretical_loss": 3.4696264839430686, + "tokens_seen": 1926496256 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021027081243731196, + "loss": 1.4349, + "theoretical_loss": 3.4696166517600755, + "tokens_seen": 1926561792 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021026078234704114, + "loss": 1.3353, + "theoretical_loss": 3.4696068200051844, + "tokens_seen": 1926627328 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021025075225677032, + "loss": 1.3936, + "theoretical_loss": 3.469596988678362, + "tokens_seen": 1926692864 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 702709, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.545248031616211, + "objective/train/theoretical_loss": 3.469587157779576, + "objective/train/tokens_used": 1947218400, + "theoretical_loss": 3.469587157779576, + "tokens_seen": 1926758400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002102407221664995, + "loss": 1.3877, + "theoretical_loss": 3.469587157779576, + "tokens_seen": 1926758400 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021023069207622868, + "loss": 1.3365, + "theoretical_loss": 3.4695773273087918, + "tokens_seen": 1926823936 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002102206619859579, + "loss": 1.3321, + "theoretical_loss": 3.4695674972659774, + "tokens_seen": 1926889472 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021021063189568707, + "loss": 1.2141, + "theoretical_loss": 3.469557667651099, + "tokens_seen": 1926955008 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021020060180541625, + "loss": 1.3536, + "theoretical_loss": 3.4695478384641243, + "tokens_seen": 1927020544 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021019057171514543, + "loss": 1.3886, + "theoretical_loss": 3.469538009705019, + "tokens_seen": 1927086080 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021018054162487464, + "loss": 1.2471, + "theoretical_loss": 3.46952818137375, + "tokens_seen": 1927151616 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021017051153460382, + "loss": 1.4203, + "theoretical_loss": 3.4695183534702854, + "tokens_seen": 1927217152 + }, + { + "epoch": 0.58, + "learning_rate": 0.000210160481444333, + "loss": 1.396, + "theoretical_loss": 3.4695085259945904, + "tokens_seen": 1927282688 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021015045135406218, + "loss": 1.4124, + "theoretical_loss": 3.4694986989466328, + "tokens_seen": 1927348224 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021014042126379136, + "loss": 1.3843, + "theoretical_loss": 3.4694888723263793, + "tokens_seen": 1927413760 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021013039117352057, + "loss": 1.3975, + "theoretical_loss": 3.469479046133797, + "tokens_seen": 1927479296 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021012036108324975, + "loss": 1.3763, + "theoretical_loss": 3.469469220368852, + "tokens_seen": 1927544832 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021011033099297894, + "loss": 1.2459, + "theoretical_loss": 3.4694593950315116, + "tokens_seen": 1927610368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021010030090270812, + "loss": 1.3637, + "theoretical_loss": 3.4694495701217427, + "tokens_seen": 1927675904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021009027081243732, + "loss": 1.3099, + "theoretical_loss": 3.4694397456395123, + "tokens_seen": 1927741440 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002100802407221665, + "loss": 1.4176, + "theoretical_loss": 3.469429921584787, + "tokens_seen": 1927806976 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002100702106318957, + "loss": 1.3287, + "theoretical_loss": 3.469420097957533, + "tokens_seen": 1927872512 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021006018054162487, + "loss": 1.3719, + "theoretical_loss": 3.4694102747577187, + "tokens_seen": 1927938048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021005015045135405, + "loss": 1.3257, + "theoretical_loss": 3.46940045198531, + "tokens_seen": 1928003584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021004012036108326, + "loss": 1.2705, + "theoretical_loss": 3.4693906296402734, + "tokens_seen": 1928069120 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021003009027081244, + "loss": 1.3227, + "theoretical_loss": 3.469380807722577, + "tokens_seen": 1928134656 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021002006018054162, + "loss": 1.4709, + "theoretical_loss": 3.469370986232186, + "tokens_seen": 1928200192 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002100100300902708, + "loss": 1.3556, + "theoretical_loss": 3.4693611651690692, + "tokens_seen": 1928265728 + }, + { + "epoch": 0.58, + "learning_rate": 0.00021, + "loss": 1.3848, + "theoretical_loss": 3.4693513445331923, + "tokens_seen": 1928331264 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002099899699097292, + "loss": 1.3346, + "theoretical_loss": 3.4693415243245225, + "tokens_seen": 1928396800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020997993981945837, + "loss": 1.2714, + "theoretical_loss": 3.4693317045430265, + "tokens_seen": 1928462336 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020996990972918755, + "loss": 1.4184, + "theoretical_loss": 3.4693218851886716, + "tokens_seen": 1928527872 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020995987963891673, + "loss": 1.3132, + "theoretical_loss": 3.4693120662614243, + "tokens_seen": 1928593408 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020994984954864594, + "loss": 1.4693, + "theoretical_loss": 3.4693022477612514, + "tokens_seen": 1928658944 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020993981945837512, + "loss": 1.4133, + "theoretical_loss": 3.4692924296881205, + "tokens_seen": 1928724480 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002099297893681043, + "loss": 1.4162, + "theoretical_loss": 3.469282612041998, + "tokens_seen": 1928790016 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020991975927783348, + "loss": 1.3306, + "theoretical_loss": 3.469272794822851, + "tokens_seen": 1928855552 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002099097291875627, + "loss": 1.4182, + "theoretical_loss": 3.4692629780306463, + "tokens_seen": 1928921088 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020989969909729187, + "loss": 1.3409, + "theoretical_loss": 3.4692531616653506, + "tokens_seen": 1928986624 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020988966900702106, + "loss": 1.3426, + "theoretical_loss": 3.469243345726931, + "tokens_seen": 1929052160 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020987963891675024, + "loss": 1.2682, + "theoretical_loss": 3.469233530215355, + "tokens_seen": 1929117696 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020986960882647942, + "loss": 1.2526, + "theoretical_loss": 3.4692237151305894, + "tokens_seen": 1929183232 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020985957873620863, + "loss": 1.268, + "theoretical_loss": 3.4692139004726004, + "tokens_seen": 1929248768 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002098495486459378, + "loss": 1.3511, + "theoretical_loss": 3.469204086241355, + "tokens_seen": 1929314304 + }, + { + "epoch": 0.58, + "learning_rate": 0.000209839518555667, + "loss": 1.3609, + "theoretical_loss": 3.4691942724368214, + "tokens_seen": 1929379840 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002098294884653962, + "loss": 1.4881, + "theoretical_loss": 3.469184459058965, + "tokens_seen": 1929445376 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002098194583751254, + "loss": 1.4279, + "theoretical_loss": 3.4691746461077537, + "tokens_seen": 1929510912 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020980942828485459, + "loss": 1.2968, + "theoretical_loss": 3.4691648335831538, + "tokens_seen": 1929576448 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020979939819458377, + "loss": 1.4145, + "theoretical_loss": 3.4691550214851334, + "tokens_seen": 1929641984 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020978936810431295, + "loss": 1.3377, + "theoretical_loss": 3.469145209813658, + "tokens_seen": 1929707520 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020977933801404216, + "loss": 1.3827, + "theoretical_loss": 3.469135398568696, + "tokens_seen": 1929773056 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020976930792377134, + "loss": 1.424, + "theoretical_loss": 3.469125587750213, + "tokens_seen": 1929838592 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020975927783350052, + "loss": 1.2849, + "theoretical_loss": 3.469115777358177, + "tokens_seen": 1929904128 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002097492477432297, + "loss": 1.2802, + "theoretical_loss": 3.4691059673925544, + "tokens_seen": 1929969664 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 703903, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3830986022949219, + "objective/train/theoretical_loss": 3.4690961578533126, + "objective/train/tokens_used": 1950495200, + "theoretical_loss": 3.4690961578533126, + "tokens_seen": 1930035200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020973921765295888, + "loss": 1.3531, + "theoretical_loss": 3.4690961578533126, + "tokens_seen": 1930035200 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002097291875626881, + "loss": 1.3185, + "theoretical_loss": 3.4690863487404187, + "tokens_seen": 1930100736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020971915747241727, + "loss": 1.4348, + "theoretical_loss": 3.4690765400538393, + "tokens_seen": 1930166272 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020970912738214645, + "loss": 1.3919, + "theoretical_loss": 3.469066731793541, + "tokens_seen": 1930231808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020969909729187563, + "loss": 1.1954, + "theoretical_loss": 3.4690569239594917, + "tokens_seen": 1930297344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020968906720160484, + "loss": 1.3693, + "theoretical_loss": 3.469047116551658, + "tokens_seen": 1930362880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00020967903711133402, + "loss": 1.3333, + "theoretical_loss": 3.469037309570007, + "tokens_seen": 1930428416 + }, + { + "epoch": 0.58, + "learning_rate": 0.0002096690070210632, + "loss": 1.3433, + "theoretical_loss": 3.4690275030145052, + "tokens_seen": 1930493952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020965897693079238, + "loss": 1.3928, + "theoretical_loss": 3.4690176968851203, + "tokens_seen": 1930559488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020964894684052156, + "loss": 1.3025, + "theoretical_loss": 3.4690078911818194, + "tokens_seen": 1930625024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020963891675025077, + "loss": 1.3023, + "theoretical_loss": 3.468998085904569, + "tokens_seen": 1930690560 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020962888665997995, + "loss": 1.401, + "theoretical_loss": 3.4689882810533366, + "tokens_seen": 1930756096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020961885656970914, + "loss": 1.3668, + "theoretical_loss": 3.4689784766280884, + "tokens_seen": 1930821632 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020960882647943832, + "loss": 1.3061, + "theoretical_loss": 3.4689686726287925, + "tokens_seen": 1930887168 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020959879638916752, + "loss": 1.3908, + "theoretical_loss": 3.468958869055415, + "tokens_seen": 1930952704 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002095887662988967, + "loss": 1.3554, + "theoretical_loss": 3.468949065907924, + "tokens_seen": 1931018240 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002095787362086259, + "loss": 1.2253, + "theoretical_loss": 3.468939263186285, + "tokens_seen": 1931083776 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020956870611835507, + "loss": 1.3842, + "theoretical_loss": 3.4689294608904664, + "tokens_seen": 1931149312 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020955867602808425, + "loss": 1.2874, + "theoretical_loss": 3.468919659020435, + "tokens_seen": 1931214848 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020954864593781346, + "loss": 1.358, + "theoretical_loss": 3.4689098575761577, + "tokens_seen": 1931280384 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020953861584754264, + "loss": 1.3721, + "theoretical_loss": 3.4689000565576014, + "tokens_seen": 1931345920 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020952858575727182, + "loss": 1.3814, + "theoretical_loss": 3.468890255964733, + "tokens_seen": 1931411456 + }, + { + "epoch": 0.59, + "learning_rate": 0.000209518555667001, + "loss": 1.3469, + "theoretical_loss": 3.46888045579752, + "tokens_seen": 1931476992 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002095085255767302, + "loss": 1.3702, + "theoretical_loss": 3.46887065605593, + "tokens_seen": 1931542528 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002094984954864594, + "loss": 1.3826, + "theoretical_loss": 3.468860856739929, + "tokens_seen": 1931608064 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020948846539618857, + "loss": 1.3868, + "theoretical_loss": 3.468851057849484, + "tokens_seen": 1931673600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020947843530591775, + "loss": 1.3548, + "theoretical_loss": 3.468841259384563, + "tokens_seen": 1931739136 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020946840521564693, + "loss": 1.4351, + "theoretical_loss": 3.4688314613451325, + "tokens_seen": 1931804672 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020945837512537614, + "loss": 1.3068, + "theoretical_loss": 3.46882166373116, + "tokens_seen": 1931870208 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020944834503510532, + "loss": 1.3973, + "theoretical_loss": 3.4688118665426124, + "tokens_seen": 1931935744 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002094383149448345, + "loss": 1.3657, + "theoretical_loss": 3.4688020697794566, + "tokens_seen": 1932001280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020942828485456368, + "loss": 1.3685, + "theoretical_loss": 3.46879227344166, + "tokens_seen": 1932066816 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002094182547642929, + "loss": 1.3613, + "theoretical_loss": 3.4687824775291896, + "tokens_seen": 1932132352 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020940822467402207, + "loss": 1.2756, + "theoretical_loss": 3.4687726820420117, + "tokens_seen": 1932197888 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020939819458375126, + "loss": 1.2819, + "theoretical_loss": 3.4687628869800946, + "tokens_seen": 1932263424 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020938816449348044, + "loss": 1.2758, + "theoretical_loss": 3.4687530923434053, + "tokens_seen": 1932328960 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020937813440320962, + "loss": 1.3416, + "theoretical_loss": 3.46874329813191, + "tokens_seen": 1932394496 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020936810431293883, + "loss": 1.3723, + "theoretical_loss": 3.468733504345577, + "tokens_seen": 1932460032 + }, + { + "epoch": 0.59, + "learning_rate": 0.000209358074222668, + "loss": 1.3837, + "theoretical_loss": 3.468723710984373, + "tokens_seen": 1932525568 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002093480441323972, + "loss": 1.3091, + "theoretical_loss": 3.4687139180482647, + "tokens_seen": 1932591104 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020933801404212637, + "loss": 1.2893, + "theoretical_loss": 3.4687041255372195, + "tokens_seen": 1932656640 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020932798395185558, + "loss": 1.3826, + "theoretical_loss": 3.4686943334512046, + "tokens_seen": 1932722176 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020931795386158476, + "loss": 1.4034, + "theoretical_loss": 3.4686845417901866, + "tokens_seen": 1932787712 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020930792377131394, + "loss": 1.3918, + "theoretical_loss": 3.4686747505541335, + "tokens_seen": 1932853248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020929789368104312, + "loss": 1.395, + "theoretical_loss": 3.4686649597430126, + "tokens_seen": 1932918784 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002092878635907723, + "loss": 1.4823, + "theoretical_loss": 3.46865516935679, + "tokens_seen": 1932984320 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002092778335005015, + "loss": 1.1854, + "theoretical_loss": 3.4686453793954337, + "tokens_seen": 1933049856 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002092678034102307, + "loss": 1.233, + "theoretical_loss": 3.46863558985891, + "tokens_seen": 1933115392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020925777331995987, + "loss": 1.2519, + "theoretical_loss": 3.468625800747187, + "tokens_seen": 1933180928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020924774322968905, + "loss": 1.3871, + "theoretical_loss": 3.4686160120602314, + "tokens_seen": 1933246464 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 705162, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3234174251556396, + "objective/train/theoretical_loss": 3.4686062237980106, + "objective/train/tokens_used": 1953772000, + "theoretical_loss": 3.4686062237980106, + "tokens_seen": 1933312000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020923771313941826, + "loss": 1.3385, + "theoretical_loss": 3.4686062237980106, + "tokens_seen": 1933312000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020922768304914744, + "loss": 1.1521, + "theoretical_loss": 3.468596435960491, + "tokens_seen": 1933377536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020921765295887662, + "loss": 1.2861, + "theoretical_loss": 3.468586648547641, + "tokens_seen": 1933443072 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002092076228686058, + "loss": 1.3558, + "theoretical_loss": 3.468576861559427, + "tokens_seen": 1933508608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020919759277833499, + "loss": 1.4589, + "theoretical_loss": 3.468567074995817, + "tokens_seen": 1933574144 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002091875626880642, + "loss": 1.3664, + "theoretical_loss": 3.4685572888567764, + "tokens_seen": 1933639680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020917753259779338, + "loss": 1.4864, + "theoretical_loss": 3.4685475031422746, + "tokens_seen": 1933705216 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020916750250752256, + "loss": 1.3918, + "theoretical_loss": 3.468537717852277, + "tokens_seen": 1933770752 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020915747241725174, + "loss": 1.3922, + "theoretical_loss": 3.4685279329867518, + "tokens_seen": 1933836288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020914744232698095, + "loss": 1.256, + "theoretical_loss": 3.468518148545666, + "tokens_seen": 1933901824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020913741223671013, + "loss": 1.3675, + "theoretical_loss": 3.468508364528987, + "tokens_seen": 1933967360 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002091273821464393, + "loss": 1.3245, + "theoretical_loss": 3.4684985809366813, + "tokens_seen": 1934032896 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002091173520561685, + "loss": 1.222, + "theoretical_loss": 3.4684887977687167, + "tokens_seen": 1934098432 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002091073219658977, + "loss": 1.3679, + "theoretical_loss": 3.4684790150250606, + "tokens_seen": 1934163968 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020909729187562688, + "loss": 1.3264, + "theoretical_loss": 3.4684692327056794, + "tokens_seen": 1934229504 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020908726178535606, + "loss": 1.3933, + "theoretical_loss": 3.468459450810541, + "tokens_seen": 1934295040 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020907723169508527, + "loss": 1.4107, + "theoretical_loss": 3.4684496693396127, + "tokens_seen": 1934360576 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020906720160481445, + "loss": 1.2902, + "theoretical_loss": 3.4684398882928615, + "tokens_seen": 1934426112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020905717151454366, + "loss": 1.3827, + "theoretical_loss": 3.4684301076702546, + "tokens_seen": 1934491648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020904714142427284, + "loss": 1.3455, + "theoretical_loss": 3.468420327471759, + "tokens_seen": 1934557184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020903711133400202, + "loss": 1.3815, + "theoretical_loss": 3.4684105476973426, + "tokens_seen": 1934622720 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002090270812437312, + "loss": 1.3244, + "theoretical_loss": 3.4684007683469718, + "tokens_seen": 1934688256 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002090170511534604, + "loss": 1.3244, + "theoretical_loss": 3.468390989420615, + "tokens_seen": 1934753792 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002090070210631896, + "loss": 1.3188, + "theoretical_loss": 3.4683812109182384, + "tokens_seen": 1934819328 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020899699097291877, + "loss": 1.4065, + "theoretical_loss": 3.4683714328398096, + "tokens_seen": 1934884864 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020898696088264795, + "loss": 1.4013, + "theoretical_loss": 3.4683616551852956, + "tokens_seen": 1934950400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020897693079237713, + "loss": 1.3869, + "theoretical_loss": 3.4683518779546647, + "tokens_seen": 1935015936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020896690070210634, + "loss": 1.379, + "theoretical_loss": 3.4683421011478828, + "tokens_seen": 1935081472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020895687061183552, + "loss": 1.4013, + "theoretical_loss": 3.468332324764918, + "tokens_seen": 1935147008 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002089468405215647, + "loss": 1.2784, + "theoretical_loss": 3.4683225488057374, + "tokens_seen": 1935212544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020893681043129389, + "loss": 1.3014, + "theoretical_loss": 3.468312773270308, + "tokens_seen": 1935278080 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002089267803410231, + "loss": 1.3881, + "theoretical_loss": 3.4683029981585975, + "tokens_seen": 1935343616 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020891675025075227, + "loss": 1.3588, + "theoretical_loss": 3.468293223470573, + "tokens_seen": 1935409152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020890672016048146, + "loss": 1.2958, + "theoretical_loss": 3.4682834492062016, + "tokens_seen": 1935474688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020889669007021064, + "loss": 1.3335, + "theoretical_loss": 3.4682736753654515, + "tokens_seen": 1935540224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020888665997993982, + "loss": 1.2233, + "theoretical_loss": 3.468263901948289, + "tokens_seen": 1935605760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020887662988966903, + "loss": 1.2561, + "theoretical_loss": 3.468254128954681, + "tokens_seen": 1935671296 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002088665997993982, + "loss": 1.2984, + "theoretical_loss": 3.468244356384596, + "tokens_seen": 1935736832 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002088565697091274, + "loss": 1.3058, + "theoretical_loss": 3.468234584238001, + "tokens_seen": 1935802368 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020884653961885657, + "loss": 1.3899, + "theoretical_loss": 3.468224812514863, + "tokens_seen": 1935867904 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020883650952858578, + "loss": 1.2685, + "theoretical_loss": 3.4682150412151493, + "tokens_seen": 1935933440 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020882647943831496, + "loss": 1.3508, + "theoretical_loss": 3.468205270338828, + "tokens_seen": 1935998976 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020881644934804414, + "loss": 1.3047, + "theoretical_loss": 3.468195499885865, + "tokens_seen": 1936064512 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020880641925777332, + "loss": 1.3473, + "theoretical_loss": 3.468185729856229, + "tokens_seen": 1936130048 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002087963891675025, + "loss": 1.4184, + "theoretical_loss": 3.4681759602498863, + "tokens_seen": 1936195584 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002087863590772317, + "loss": 1.453, + "theoretical_loss": 3.4681661910668047, + "tokens_seen": 1936261120 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002087763289869609, + "loss": 1.3907, + "theoretical_loss": 3.4681564223069516, + "tokens_seen": 1936326656 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020876629889669007, + "loss": 1.5468, + "theoretical_loss": 3.4681466539702948, + "tokens_seen": 1936392192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020875626880641925, + "loss": 1.4061, + "theoretical_loss": 3.4681368860568, + "tokens_seen": 1936457728 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020874623871614846, + "loss": 1.4385, + "theoretical_loss": 3.4681271185664366, + "tokens_seen": 1936523264 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 706032, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2811541557312012, + "objective/train/theoretical_loss": 3.4681173514991706, + "objective/train/tokens_used": 1957048800, + "theoretical_loss": 3.4681173514991706, + "tokens_seen": 1936588800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020873620862587764, + "loss": 1.3258, + "theoretical_loss": 3.4681173514991706, + "tokens_seen": 1936588800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020872617853560682, + "loss": 1.3161, + "theoretical_loss": 3.4681075848549696, + "tokens_seen": 1936654336 + }, + { + "epoch": 0.59, + "learning_rate": 0.000208716148445336, + "loss": 1.342, + "theoretical_loss": 3.4680978186338014, + "tokens_seen": 1936719872 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020870611835506519, + "loss": 1.3797, + "theoretical_loss": 3.4680880528356326, + "tokens_seen": 1936785408 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002086960882647944, + "loss": 1.3988, + "theoretical_loss": 3.4680782874604317, + "tokens_seen": 1936850944 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020868605817452358, + "loss": 1.3542, + "theoretical_loss": 3.4680685225081653, + "tokens_seen": 1936916480 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020867602808425276, + "loss": 1.3145, + "theoretical_loss": 3.4680587579788007, + "tokens_seen": 1936982016 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020866599799398194, + "loss": 1.3107, + "theoretical_loss": 3.468048993872306, + "tokens_seen": 1937047552 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020865596790371115, + "loss": 1.3119, + "theoretical_loss": 3.468039230188647, + "tokens_seen": 1937113088 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020864593781344033, + "loss": 1.3185, + "theoretical_loss": 3.468029466927793, + "tokens_seen": 1937178624 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002086359077231695, + "loss": 1.3843, + "theoretical_loss": 3.46801970408971, + "tokens_seen": 1937244160 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002086258776328987, + "loss": 1.3454, + "theoretical_loss": 3.468009941674366, + "tokens_seen": 1937309696 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002086158475426279, + "loss": 1.3929, + "theoretical_loss": 3.468000179681728, + "tokens_seen": 1937375232 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020860581745235708, + "loss": 1.3456, + "theoretical_loss": 3.4679904181117642, + "tokens_seen": 1937440768 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020859578736208626, + "loss": 1.3611, + "theoretical_loss": 3.467980656964442, + "tokens_seen": 1937506304 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020858575727181544, + "loss": 1.41, + "theoretical_loss": 3.467970896239727, + "tokens_seen": 1937571840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020857572718154462, + "loss": 1.3569, + "theoretical_loss": 3.467961135937589, + "tokens_seen": 1937637376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020856569709127383, + "loss": 1.3891, + "theoretical_loss": 3.467951376057994, + "tokens_seen": 1937702912 + }, + { + "epoch": 0.59, + "learning_rate": 0.000208555667001003, + "loss": 1.3442, + "theoretical_loss": 3.4679416166009096, + "tokens_seen": 1937768448 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002085456369107322, + "loss": 1.3372, + "theoretical_loss": 3.467931857566304, + "tokens_seen": 1937833984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020853560682046137, + "loss": 1.2657, + "theoretical_loss": 3.4679220989541433, + "tokens_seen": 1937899520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020852557673019058, + "loss": 1.3661, + "theoretical_loss": 3.4679123407643955, + "tokens_seen": 1937965056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020851554663991976, + "loss": 1.3544, + "theoretical_loss": 3.467902582997029, + "tokens_seen": 1938030592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020850551654964894, + "loss": 1.3225, + "theoretical_loss": 3.46789282565201, + "tokens_seen": 1938096128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020849548645937813, + "loss": 1.4552, + "theoretical_loss": 3.467883068729306, + "tokens_seen": 1938161664 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002084854563691073, + "loss": 1.3162, + "theoretical_loss": 3.467873312228885, + "tokens_seen": 1938227200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020847542627883651, + "loss": 1.3742, + "theoretical_loss": 3.4678635561507143, + "tokens_seen": 1938292736 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002084653961885657, + "loss": 1.3515, + "theoretical_loss": 3.467853800494761, + "tokens_seen": 1938358272 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020845536609829488, + "loss": 1.3566, + "theoretical_loss": 3.467844045260993, + "tokens_seen": 1938423808 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020844533600802406, + "loss": 1.3177, + "theoretical_loss": 3.4678342904493777, + "tokens_seen": 1938489344 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020843530591775327, + "loss": 1.3264, + "theoretical_loss": 3.467824536059882, + "tokens_seen": 1938554880 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020842527582748245, + "loss": 1.3569, + "theoretical_loss": 3.4678147820924745, + "tokens_seen": 1938620416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020841524573721163, + "loss": 1.2876, + "theoretical_loss": 3.4678050285471214, + "tokens_seen": 1938685952 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002084052156469408, + "loss": 1.3088, + "theoretical_loss": 3.467795275423791, + "tokens_seen": 1938751488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020839518555667, + "loss": 1.3719, + "theoretical_loss": 3.4677855227224508, + "tokens_seen": 1938817024 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002083851554663992, + "loss": 1.3953, + "theoretical_loss": 3.4677757704430676, + "tokens_seen": 1938882560 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020837512537612838, + "loss": 1.257, + "theoretical_loss": 3.4677660185856096, + "tokens_seen": 1938948096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020836509528585756, + "loss": 1.3396, + "theoretical_loss": 3.467756267150044, + "tokens_seen": 1939013632 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020835506519558674, + "loss": 1.3605, + "theoretical_loss": 3.467746516136337, + "tokens_seen": 1939079168 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020834503510531595, + "loss": 1.2694, + "theoretical_loss": 3.467736765544459, + "tokens_seen": 1939144704 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020833500501504513, + "loss": 1.3541, + "theoretical_loss": 3.467727015374375, + "tokens_seen": 1939210240 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020832497492477434, + "loss": 1.3588, + "theoretical_loss": 3.467717265626054, + "tokens_seen": 1939275776 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020831494483450352, + "loss": 1.3975, + "theoretical_loss": 3.4677075162994617, + "tokens_seen": 1939341312 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002083049147442327, + "loss": 1.3593, + "theoretical_loss": 3.4676977673945677, + "tokens_seen": 1939406848 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002082948846539619, + "loss": 1.3904, + "theoretical_loss": 3.467688018911338, + "tokens_seen": 1939472384 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002082848545636911, + "loss": 1.3566, + "theoretical_loss": 3.467678270849741, + "tokens_seen": 1939537920 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020827482447342027, + "loss": 1.3079, + "theoretical_loss": 3.467668523209744, + "tokens_seen": 1939603456 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020826479438314945, + "loss": 1.3552, + "theoretical_loss": 3.467658775991315, + "tokens_seen": 1939668992 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020825476429287866, + "loss": 1.2902, + "theoretical_loss": 3.46764902919442, + "tokens_seen": 1939734528 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020824473420260784, + "loss": 1.3411, + "theoretical_loss": 3.4676392828190274, + "tokens_seen": 1939800064 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 707259, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.380761742591858, + "objective/train/theoretical_loss": 3.4676295368651053, + "objective/train/tokens_used": 1960325600, + "theoretical_loss": 3.4676295368651053, + "tokens_seen": 1939865600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020823470411233702, + "loss": 1.3667, + "theoretical_loss": 3.4676295368651053, + "tokens_seen": 1939865600 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002082246740220662, + "loss": 1.3137, + "theoretical_loss": 3.4676197913326208, + "tokens_seen": 1939931136 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002082146439317954, + "loss": 1.3332, + "theoretical_loss": 3.467610046221541, + "tokens_seen": 1939996672 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002082046138415246, + "loss": 1.3979, + "theoretical_loss": 3.467600301531834, + "tokens_seen": 1940062208 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020819458375125378, + "loss": 1.3223, + "theoretical_loss": 3.467590557263467, + "tokens_seen": 1940127744 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020818455366098296, + "loss": 1.2769, + "theoretical_loss": 3.467580813416408, + "tokens_seen": 1940193280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020817452357071214, + "loss": 1.374, + "theoretical_loss": 3.4675710699906244, + "tokens_seen": 1940258816 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020816449348044135, + "loss": 1.3293, + "theoretical_loss": 3.4675613269860834, + "tokens_seen": 1940324352 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020815446339017053, + "loss": 1.3602, + "theoretical_loss": 3.4675515844027522, + "tokens_seen": 1940389888 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002081444332998997, + "loss": 1.3917, + "theoretical_loss": 3.4675418422405997, + "tokens_seen": 1940455424 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002081344032096289, + "loss": 1.399, + "theoretical_loss": 3.467532100499593, + "tokens_seen": 1940520960 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002081243731193581, + "loss": 1.3219, + "theoretical_loss": 3.4675223591796986, + "tokens_seen": 1940586496 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020811434302908728, + "loss": 1.3316, + "theoretical_loss": 3.4675126182808853, + "tokens_seen": 1940652032 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020810431293881646, + "loss": 1.3384, + "theoretical_loss": 3.4675028778031205, + "tokens_seen": 1940717568 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020809428284854564, + "loss": 1.3214, + "theoretical_loss": 3.467493137746371, + "tokens_seen": 1940783104 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020808425275827482, + "loss": 1.2316, + "theoretical_loss": 3.4674833981106055, + "tokens_seen": 1940848640 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020807422266800403, + "loss": 1.3031, + "theoretical_loss": 3.4674736588957904, + "tokens_seen": 1940914176 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002080641925777332, + "loss": 1.3295, + "theoretical_loss": 3.4674639201018946, + "tokens_seen": 1940979712 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002080541624874624, + "loss": 1.3713, + "theoretical_loss": 3.4674541817288844, + "tokens_seen": 1941045248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020804413239719157, + "loss": 1.3658, + "theoretical_loss": 3.4674444437767282, + "tokens_seen": 1941110784 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020803410230692078, + "loss": 1.3644, + "theoretical_loss": 3.4674347062453936, + "tokens_seen": 1941176320 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020802407221664996, + "loss": 1.3151, + "theoretical_loss": 3.467424969134848, + "tokens_seen": 1941241856 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020801404212637914, + "loss": 1.2845, + "theoretical_loss": 3.4674152324450587, + "tokens_seen": 1941307392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020800401203610833, + "loss": 1.3705, + "theoretical_loss": 3.467405496175994, + "tokens_seen": 1941372928 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002079939819458375, + "loss": 1.2531, + "theoretical_loss": 3.467395760327621, + "tokens_seen": 1941438464 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020798395185556671, + "loss": 1.451, + "theoretical_loss": 3.4673860248999073, + "tokens_seen": 1941504000 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002079739217652959, + "loss": 1.3738, + "theoretical_loss": 3.467376289892821, + "tokens_seen": 1941569536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020796389167502508, + "loss": 1.4041, + "theoretical_loss": 3.4673665553063295, + "tokens_seen": 1941635072 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020795386158475426, + "loss": 1.3255, + "theoretical_loss": 3.4673568211404002, + "tokens_seen": 1941700608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020794383149448347, + "loss": 1.2156, + "theoretical_loss": 3.4673470873950007, + "tokens_seen": 1941766144 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020793380140421265, + "loss": 1.2264, + "theoretical_loss": 3.467337354070099, + "tokens_seen": 1941831680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020792377131394183, + "loss": 1.2983, + "theoretical_loss": 3.467327621165663, + "tokens_seen": 1941897216 + }, + { + "epoch": 0.59, + "learning_rate": 0.000207913741223671, + "loss": 1.3403, + "theoretical_loss": 3.46731788868166, + "tokens_seen": 1941962752 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002079037111334002, + "loss": 1.332, + "theoretical_loss": 3.4673081566180572, + "tokens_seen": 1942028288 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002078936810431294, + "loss": 1.2874, + "theoretical_loss": 3.467298424974823, + "tokens_seen": 1942093824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020788365095285858, + "loss": 1.306, + "theoretical_loss": 3.4672886937519243, + "tokens_seen": 1942159360 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020787362086258776, + "loss": 1.2774, + "theoretical_loss": 3.467278962949329, + "tokens_seen": 1942224896 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020786359077231694, + "loss": 1.362, + "theoretical_loss": 3.4672692325670056, + "tokens_seen": 1942290432 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020785356068204615, + "loss": 1.2915, + "theoretical_loss": 3.4672595026049207, + "tokens_seen": 1942355968 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020784353059177533, + "loss": 1.2883, + "theoretical_loss": 3.467249773063042, + "tokens_seen": 1942421504 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002078335005015045, + "loss": 1.3314, + "theoretical_loss": 3.467240043941338, + "tokens_seen": 1942487040 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002078234704112337, + "loss": 1.4241, + "theoretical_loss": 3.4672303152397763, + "tokens_seen": 1942552576 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020781344032096288, + "loss": 1.3587, + "theoretical_loss": 3.4672205869583235, + "tokens_seen": 1942618112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020780341023069208, + "loss": 1.4118, + "theoretical_loss": 3.467210859096948, + "tokens_seen": 1942683648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020779338014042126, + "loss": 1.2801, + "theoretical_loss": 3.467201131655618, + "tokens_seen": 1942749184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020778335005015045, + "loss": 1.3687, + "theoretical_loss": 3.467191404634301, + "tokens_seen": 1942814720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020777331995987963, + "loss": 1.3486, + "theoretical_loss": 3.467181678032963, + "tokens_seen": 1942880256 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020776328986960884, + "loss": 1.2756, + "theoretical_loss": 3.4671719518515745, + "tokens_seen": 1942945792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020775325977933802, + "loss": 1.274, + "theoretical_loss": 3.4671622260901014, + "tokens_seen": 1943011328 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002077432296890672, + "loss": 1.4291, + "theoretical_loss": 3.4671525007485116, + "tokens_seen": 1943076864 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 708379, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3127143383026123, + "objective/train/theoretical_loss": 3.4671427758267725, + "objective/train/tokens_used": 1963602400, + "theoretical_loss": 3.4671427758267725, + "tokens_seen": 1943142400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020773319959879638, + "loss": 1.2619, + "theoretical_loss": 3.4671427758267725, + "tokens_seen": 1943142400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020772316950852556, + "loss": 1.4222, + "theoretical_loss": 3.4671330513248533, + "tokens_seen": 1943207936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020771313941825477, + "loss": 1.414, + "theoretical_loss": 3.46712332724272, + "tokens_seen": 1943273472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020770310932798395, + "loss": 1.3883, + "theoretical_loss": 3.4671136035803416, + "tokens_seen": 1943339008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020769307923771313, + "loss": 1.2996, + "theoretical_loss": 3.4671038803376852, + "tokens_seen": 1943404544 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002076830491474423, + "loss": 1.3213, + "theoretical_loss": 3.4670941575147185, + "tokens_seen": 1943470080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020767301905717152, + "loss": 1.3219, + "theoretical_loss": 3.4670844351114094, + "tokens_seen": 1943535616 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002076629889669007, + "loss": 1.3637, + "theoretical_loss": 3.4670747131277255, + "tokens_seen": 1943601152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020765295887662988, + "loss": 1.3767, + "theoretical_loss": 3.467064991563635, + "tokens_seen": 1943666688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020764292878635906, + "loss": 1.3689, + "theoretical_loss": 3.4670552704191047, + "tokens_seen": 1943732224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020763289869608824, + "loss": 1.4724, + "theoretical_loss": 3.4670455496941033, + "tokens_seen": 1943797760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020762286860581745, + "loss": 1.3218, + "theoretical_loss": 3.467035829388598, + "tokens_seen": 1943863296 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020761283851554663, + "loss": 1.2839, + "theoretical_loss": 3.467026109502557, + "tokens_seen": 1943928832 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020760280842527581, + "loss": 1.3651, + "theoretical_loss": 3.4670163900359476, + "tokens_seen": 1943994368 + }, + { + "epoch": 0.59, + "learning_rate": 0.000207592778335005, + "loss": 1.4292, + "theoretical_loss": 3.4670066709887375, + "tokens_seen": 1944059904 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002075827482447342, + "loss": 1.3114, + "theoretical_loss": 3.466996952360895, + "tokens_seen": 1944125440 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002075727181544634, + "loss": 1.2728, + "theoretical_loss": 3.4669872341523877, + "tokens_seen": 1944190976 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002075626880641926, + "loss": 1.125, + "theoretical_loss": 3.4669775163631833, + "tokens_seen": 1944256512 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020755265797392177, + "loss": 1.4505, + "theoretical_loss": 3.466967798993249, + "tokens_seen": 1944322048 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020754262788365098, + "loss": 1.3166, + "theoretical_loss": 3.466958082042553, + "tokens_seen": 1944387584 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020753259779338016, + "loss": 1.3195, + "theoretical_loss": 3.466948365511064, + "tokens_seen": 1944453120 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020752256770310934, + "loss": 1.4781, + "theoretical_loss": 3.466938649398749, + "tokens_seen": 1944518656 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020751253761283853, + "loss": 1.3434, + "theoretical_loss": 3.4669289337055753, + "tokens_seen": 1944584192 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002075025075225677, + "loss": 1.3754, + "theoretical_loss": 3.466919218431511, + "tokens_seen": 1944649728 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020749247743229692, + "loss": 1.3275, + "theoretical_loss": 3.4669095035765247, + "tokens_seen": 1944715264 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002074824473420261, + "loss": 1.3675, + "theoretical_loss": 3.466899789140583, + "tokens_seen": 1944780800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020747241725175528, + "loss": 1.3339, + "theoretical_loss": 3.466890075123654, + "tokens_seen": 1944846336 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020746238716148446, + "loss": 1.3965, + "theoretical_loss": 3.466880361525707, + "tokens_seen": 1944911872 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020745235707121367, + "loss": 1.3234, + "theoretical_loss": 3.466870648346707, + "tokens_seen": 1944977408 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020744232698094285, + "loss": 1.3008, + "theoretical_loss": 3.4668609355866247, + "tokens_seen": 1945042944 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020743229689067203, + "loss": 1.3118, + "theoretical_loss": 3.466851223245426, + "tokens_seen": 1945108480 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002074222668004012, + "loss": 1.3769, + "theoretical_loss": 3.4668415113230795, + "tokens_seen": 1945174016 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002074122367101304, + "loss": 1.3912, + "theoretical_loss": 3.466831799819553, + "tokens_seen": 1945239552 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002074022066198596, + "loss": 1.1949, + "theoretical_loss": 3.466822088734814, + "tokens_seen": 1945305088 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020739217652958878, + "loss": 1.378, + "theoretical_loss": 3.4668123780688305, + "tokens_seen": 1945370624 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020738214643931796, + "loss": 1.3613, + "theoretical_loss": 3.4668026678215704, + "tokens_seen": 1945436160 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020737211634904714, + "loss": 1.3176, + "theoretical_loss": 3.4667929579930012, + "tokens_seen": 1945501696 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020736208625877635, + "loss": 1.3311, + "theoretical_loss": 3.466783248583092, + "tokens_seen": 1945567232 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020735205616850553, + "loss": 1.2906, + "theoretical_loss": 3.466773539591809, + "tokens_seen": 1945632768 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002073420260782347, + "loss": 1.2003, + "theoretical_loss": 3.4667638310191204, + "tokens_seen": 1945698304 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002073319959879639, + "loss": 1.4048, + "theoretical_loss": 3.4667541228649954, + "tokens_seen": 1945763840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020732196589769308, + "loss": 1.343, + "theoretical_loss": 3.4667444151294005, + "tokens_seen": 1945829376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020731193580742228, + "loss": 1.2626, + "theoretical_loss": 3.4667347078123036, + "tokens_seen": 1945894912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020730190571715146, + "loss": 1.3899, + "theoretical_loss": 3.466725000913673, + "tokens_seen": 1945960448 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020729187562688065, + "loss": 1.363, + "theoretical_loss": 3.4667152944334765, + "tokens_seen": 1946025984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020728184553660983, + "loss": 1.3086, + "theoretical_loss": 3.466705588371682, + "tokens_seen": 1946091520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020727181544633904, + "loss": 1.2545, + "theoretical_loss": 3.4666958827282572, + "tokens_seen": 1946157056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020726178535606822, + "loss": 1.2988, + "theoretical_loss": 3.46668617750317, + "tokens_seen": 1946222592 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002072517552657974, + "loss": 1.4157, + "theoretical_loss": 3.4666764726963883, + "tokens_seen": 1946288128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020724172517552658, + "loss": 1.3439, + "theoretical_loss": 3.4666667683078805, + "tokens_seen": 1946353664 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 709628, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.346217393875122, + "objective/train/theoretical_loss": 3.466657064337614, + "objective/train/tokens_used": 1966879200, + "theoretical_loss": 3.466657064337614, + "tokens_seen": 1946419200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020723169508525576, + "loss": 1.2991, + "theoretical_loss": 3.466657064337614, + "tokens_seen": 1946419200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020722166499498497, + "loss": 1.3147, + "theoretical_loss": 3.466647360785556, + "tokens_seen": 1946484736 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020721163490471415, + "loss": 1.4264, + "theoretical_loss": 3.4666376576516758, + "tokens_seen": 1946550272 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020720160481444333, + "loss": 1.2602, + "theoretical_loss": 3.4666279549359404, + "tokens_seen": 1946615808 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002071915747241725, + "loss": 1.3486, + "theoretical_loss": 3.4666182526383182, + "tokens_seen": 1946681344 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020718154463390172, + "loss": 1.2247, + "theoretical_loss": 3.4666085507587767, + "tokens_seen": 1946746880 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002071715145436309, + "loss": 1.4221, + "theoretical_loss": 3.466598849297284, + "tokens_seen": 1946812416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020716148445336008, + "loss": 1.4173, + "theoretical_loss": 3.466589148253808, + "tokens_seen": 1946877952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020715145436308926, + "loss": 1.3901, + "theoretical_loss": 3.466579447628316, + "tokens_seen": 1946943488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020714142427281844, + "loss": 1.4437, + "theoretical_loss": 3.4665697474207775, + "tokens_seen": 1947009024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020713139418254765, + "loss": 1.3714, + "theoretical_loss": 3.466560047631159, + "tokens_seen": 1947074560 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020712136409227683, + "loss": 1.2671, + "theoretical_loss": 3.466550348259428, + "tokens_seen": 1947140096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020711133400200601, + "loss": 1.3764, + "theoretical_loss": 3.4665406493055544, + "tokens_seen": 1947205632 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002071013039117352, + "loss": 1.2859, + "theoretical_loss": 3.4665309507695046, + "tokens_seen": 1947271168 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002070912738214644, + "loss": 1.4219, + "theoretical_loss": 3.466521252651247, + "tokens_seen": 1947336704 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020708124373119358, + "loss": 1.2875, + "theoretical_loss": 3.46651155495075, + "tokens_seen": 1947402240 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020707121364092277, + "loss": 1.2431, + "theoretical_loss": 3.4665018576679802, + "tokens_seen": 1947467776 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020706118355065195, + "loss": 1.3203, + "theoretical_loss": 3.466492160802907, + "tokens_seen": 1947533312 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020705115346038116, + "loss": 1.3779, + "theoretical_loss": 3.466482464355498, + "tokens_seen": 1947598848 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020704112337011034, + "loss": 1.3647, + "theoretical_loss": 3.4664727683257204, + "tokens_seen": 1947664384 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020703109327983952, + "loss": 1.3191, + "theoretical_loss": 3.466463072713543, + "tokens_seen": 1947729920 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002070210631895687, + "loss": 1.3345, + "theoretical_loss": 3.466453377518933, + "tokens_seen": 1947795456 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020701103309929788, + "loss": 1.2966, + "theoretical_loss": 3.466443682741859, + "tokens_seen": 1947860992 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002070010030090271, + "loss": 1.3604, + "theoretical_loss": 3.466433988382289, + "tokens_seen": 1947926528 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020699097291875627, + "loss": 1.2948, + "theoretical_loss": 3.4664242944401904, + "tokens_seen": 1947992064 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020698094282848545, + "loss": 1.3023, + "theoretical_loss": 3.466414600915532, + "tokens_seen": 1948057600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020697091273821463, + "loss": 1.3877, + "theoretical_loss": 3.466404907808281, + "tokens_seen": 1948123136 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020696088264794384, + "loss": 1.3604, + "theoretical_loss": 3.466395215118406, + "tokens_seen": 1948188672 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020695085255767302, + "loss": 1.4198, + "theoretical_loss": 3.466385522845874, + "tokens_seen": 1948254208 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002069408224674022, + "loss": 1.339, + "theoretical_loss": 3.4663758309906543, + "tokens_seen": 1948319744 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020693079237713138, + "loss": 1.3152, + "theoretical_loss": 3.466366139552714, + "tokens_seen": 1948385280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020692076228686056, + "loss": 1.3772, + "theoretical_loss": 3.4663564485320215, + "tokens_seen": 1948450816 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020691073219658977, + "loss": 1.3742, + "theoretical_loss": 3.466346757928545, + "tokens_seen": 1948516352 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020690070210631895, + "loss": 1.3104, + "theoretical_loss": 3.466337067742252, + "tokens_seen": 1948581888 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020689067201604813, + "loss": 1.2396, + "theoretical_loss": 3.4663273779731103, + "tokens_seen": 1948647424 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020688064192577732, + "loss": 1.4651, + "theoretical_loss": 3.4663176886210887, + "tokens_seen": 1948712960 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020687061183550652, + "loss": 1.3664, + "theoretical_loss": 3.4663079996861548, + "tokens_seen": 1948778496 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002068605817452357, + "loss": 1.2383, + "theoretical_loss": 3.4662983111682766, + "tokens_seen": 1948844032 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020685055165496489, + "loss": 1.3932, + "theoretical_loss": 3.466288623067422, + "tokens_seen": 1948909568 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020684052156469407, + "loss": 1.4339, + "theoretical_loss": 3.4662789353835595, + "tokens_seen": 1948975104 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020683049147442328, + "loss": 1.4146, + "theoretical_loss": 3.4662692481166566, + "tokens_seen": 1949040640 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020682046138415248, + "loss": 1.2693, + "theoretical_loss": 3.466259561266682, + "tokens_seen": 1949106176 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020681043129388166, + "loss": 1.2448, + "theoretical_loss": 3.466249874833603, + "tokens_seen": 1949171712 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020680040120361085, + "loss": 1.2847, + "theoretical_loss": 3.4662401888173875, + "tokens_seen": 1949237248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020679037111334003, + "loss": 1.2111, + "theoretical_loss": 3.4662305032180045, + "tokens_seen": 1949302784 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020678034102306924, + "loss": 1.454, + "theoretical_loss": 3.4662208180354215, + "tokens_seen": 1949368320 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020677031093279842, + "loss": 1.3634, + "theoretical_loss": 3.4662111332696064, + "tokens_seen": 1949433856 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002067602808425276, + "loss": 1.3286, + "theoretical_loss": 3.4662014489205277, + "tokens_seen": 1949499392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020675025075225678, + "loss": 1.2135, + "theoretical_loss": 3.466191764988153, + "tokens_seen": 1949564928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020674022066198596, + "loss": 1.4734, + "theoretical_loss": 3.466182081472451, + "tokens_seen": 1949630464 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 710798, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3653513193130493, + "objective/train/theoretical_loss": 3.4661723983733888, + "objective/train/tokens_used": 1970156000, + "theoretical_loss": 3.4661723983733888, + "tokens_seen": 1949696000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020673019057171517, + "loss": 1.3522, + "theoretical_loss": 3.4661723983733888, + "tokens_seen": 1949696000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020672016048144435, + "loss": 1.389, + "theoretical_loss": 3.4661627156909356, + "tokens_seen": 1949761536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020671013039117353, + "loss": 1.3758, + "theoretical_loss": 3.466153033425058, + "tokens_seen": 1949827072 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002067001003009027, + "loss": 1.4411, + "theoretical_loss": 3.4661433515757256, + "tokens_seen": 1949892608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020669007021063192, + "loss": 1.2862, + "theoretical_loss": 3.4661336701429057, + "tokens_seen": 1949958144 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002066800401203611, + "loss": 1.3148, + "theoretical_loss": 3.4661239891265665, + "tokens_seen": 1950023680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020667001003009028, + "loss": 1.1443, + "theoretical_loss": 3.4661143085266763, + "tokens_seen": 1950089216 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020665997993981946, + "loss": 1.3936, + "theoretical_loss": 3.466104628343203, + "tokens_seen": 1950154752 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020664994984954864, + "loss": 1.3064, + "theoretical_loss": 3.4660949485761146, + "tokens_seen": 1950220288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020663991975927785, + "loss": 1.3547, + "theoretical_loss": 3.466085269225379, + "tokens_seen": 1950285824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020662988966900703, + "loss": 1.3557, + "theoretical_loss": 3.4660755902909646, + "tokens_seen": 1950351360 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020661985957873621, + "loss": 1.4493, + "theoretical_loss": 3.46606591177284, + "tokens_seen": 1950416896 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002066098294884654, + "loss": 1.4097, + "theoretical_loss": 3.4660562336709724, + "tokens_seen": 1950482432 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002065997993981946, + "loss": 1.3422, + "theoretical_loss": 3.4660465559853306, + "tokens_seen": 1950547968 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020658976930792379, + "loss": 1.3352, + "theoretical_loss": 3.466036878715882, + "tokens_seen": 1950613504 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020657973921765297, + "loss": 1.4057, + "theoretical_loss": 3.4660272018625955, + "tokens_seen": 1950679040 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020656970912738215, + "loss": 1.2667, + "theoretical_loss": 3.466017525425439, + "tokens_seen": 1950744576 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020655967903711136, + "loss": 1.2819, + "theoretical_loss": 3.4660078494043804, + "tokens_seen": 1950810112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020654964894684054, + "loss": 1.2972, + "theoretical_loss": 3.465998173799388, + "tokens_seen": 1950875648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020653961885656972, + "loss": 1.3157, + "theoretical_loss": 3.4659884986104297, + "tokens_seen": 1950941184 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002065295887662989, + "loss": 1.3159, + "theoretical_loss": 3.4659788238374736, + "tokens_seen": 1951006720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020651955867602808, + "loss": 1.4341, + "theoretical_loss": 3.4659691494804887, + "tokens_seen": 1951072256 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002065095285857573, + "loss": 1.4255, + "theoretical_loss": 3.4659594755394423, + "tokens_seen": 1951137792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020649949849548647, + "loss": 1.3715, + "theoretical_loss": 3.4659498020143023, + "tokens_seen": 1951203328 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020648946840521565, + "loss": 1.3764, + "theoretical_loss": 3.4659401289050376, + "tokens_seen": 1951268864 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020647943831494483, + "loss": 1.2289, + "theoretical_loss": 3.4659304562116158, + "tokens_seen": 1951334400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020646940822467404, + "loss": 1.2734, + "theoretical_loss": 3.4659207839340054, + "tokens_seen": 1951399936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020645937813440322, + "loss": 1.2811, + "theoretical_loss": 3.465911112072175, + "tokens_seen": 1951465472 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002064493480441324, + "loss": 1.3208, + "theoretical_loss": 3.4659014406260913, + "tokens_seen": 1951531008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020643931795386158, + "loss": 1.3227, + "theoretical_loss": 3.465891769595724, + "tokens_seen": 1951596544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020642928786359076, + "loss": 1.3638, + "theoretical_loss": 3.465882098981041, + "tokens_seen": 1951662080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020641925777331997, + "loss": 1.3886, + "theoretical_loss": 3.465872428782009, + "tokens_seen": 1951727616 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020640922768304915, + "loss": 1.4408, + "theoretical_loss": 3.465862758998598, + "tokens_seen": 1951793152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020639919759277833, + "loss": 1.2929, + "theoretical_loss": 3.4658530896307758, + "tokens_seen": 1951858688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020638916750250752, + "loss": 1.2962, + "theoretical_loss": 3.46584342067851, + "tokens_seen": 1951924224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020637913741223672, + "loss": 1.3006, + "theoretical_loss": 3.465833752141769, + "tokens_seen": 1951989760 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002063691073219659, + "loss": 1.3456, + "theoretical_loss": 3.465824084020521, + "tokens_seen": 1952055296 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020635907723169509, + "loss": 1.329, + "theoretical_loss": 3.465814416314734, + "tokens_seen": 1952120832 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020634904714142427, + "loss": 1.3111, + "theoretical_loss": 3.465804749024377, + "tokens_seen": 1952186368 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020633901705115345, + "loss": 1.3116, + "theoretical_loss": 3.465795082149418, + "tokens_seen": 1952251904 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020632898696088266, + "loss": 1.3631, + "theoretical_loss": 3.4657854156898242, + "tokens_seen": 1952317440 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020631895687061184, + "loss": 1.4591, + "theoretical_loss": 3.4657757496455646, + "tokens_seen": 1952382976 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020630892678034102, + "loss": 1.3938, + "theoretical_loss": 3.4657660840166074, + "tokens_seen": 1952448512 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002062988966900702, + "loss": 1.3416, + "theoretical_loss": 3.4657564188029206, + "tokens_seen": 1952514048 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002062888665997994, + "loss": 1.3788, + "theoretical_loss": 3.465746754004473, + "tokens_seen": 1952579584 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002062788365095286, + "loss": 1.3575, + "theoretical_loss": 3.465737089621232, + "tokens_seen": 1952645120 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020626880641925777, + "loss": 1.3285, + "theoretical_loss": 3.4657274256531663, + "tokens_seen": 1952710656 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020625877632898695, + "loss": 1.2946, + "theoretical_loss": 3.465717762100244, + "tokens_seen": 1952776192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020624874623871613, + "loss": 1.3317, + "theoretical_loss": 3.465708098962433, + "tokens_seen": 1952841728 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020623871614844534, + "loss": 1.5069, + "theoretical_loss": 3.465698436239702, + "tokens_seen": 1952907264 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 712021, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3026814460754395, + "objective/train/theoretical_loss": 3.4656887739320195, + "objective/train/tokens_used": 1973432800, + "theoretical_loss": 3.4656887739320195, + "tokens_seen": 1952972800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020622868605817452, + "loss": 1.2863, + "theoretical_loss": 3.4656887739320195, + "tokens_seen": 1952972800 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002062186559679037, + "loss": 1.3737, + "theoretical_loss": 3.465679112039353, + "tokens_seen": 1953038336 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020620862587763288, + "loss": 1.3165, + "theoretical_loss": 3.4656694505616716, + "tokens_seen": 1953103872 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002061985957873621, + "loss": 1.3059, + "theoretical_loss": 3.4656597894989427, + "tokens_seen": 1953169408 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020618856569709127, + "loss": 1.3187, + "theoretical_loss": 3.465650128851135, + "tokens_seen": 1953234944 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020617853560682046, + "loss": 1.2685, + "theoretical_loss": 3.4656404686182167, + "tokens_seen": 1953300480 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020616850551654964, + "loss": 1.3629, + "theoretical_loss": 3.465630808800156, + "tokens_seen": 1953366016 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020615847542627882, + "loss": 1.2535, + "theoretical_loss": 3.4656211493969216, + "tokens_seen": 1953431552 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020614844533600803, + "loss": 1.3294, + "theoretical_loss": 3.465611490408481, + "tokens_seen": 1953497088 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002061384152457372, + "loss": 1.3925, + "theoretical_loss": 3.465601831834803, + "tokens_seen": 1953562624 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002061283851554664, + "loss": 1.3901, + "theoretical_loss": 3.465592173675856, + "tokens_seen": 1953628160 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020611835506519557, + "loss": 1.3159, + "theoretical_loss": 3.465582515931608, + "tokens_seen": 1953693696 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020610832497492478, + "loss": 1.3651, + "theoretical_loss": 3.4655728586020267, + "tokens_seen": 1953759232 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020609829488465396, + "loss": 1.4299, + "theoretical_loss": 3.4655632016870817, + "tokens_seen": 1953824768 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020608826479438314, + "loss": 1.2646, + "theoretical_loss": 3.46555354518674, + "tokens_seen": 1953890304 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020607823470411235, + "loss": 1.3904, + "theoretical_loss": 3.465543889100971, + "tokens_seen": 1953955840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020606820461384156, + "loss": 1.4271, + "theoretical_loss": 3.465534233429742, + "tokens_seen": 1954021376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020605817452357074, + "loss": 1.2773, + "theoretical_loss": 3.4655245781730226, + "tokens_seen": 1954086912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020604814443329992, + "loss": 1.3786, + "theoretical_loss": 3.4655149233307796, + "tokens_seen": 1954152448 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002060381143430291, + "loss": 1.4491, + "theoretical_loss": 3.465505268902982, + "tokens_seen": 1954217984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020602808425275828, + "loss": 1.2902, + "theoretical_loss": 3.4654956148895986, + "tokens_seen": 1954283520 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002060180541624875, + "loss": 1.2653, + "theoretical_loss": 3.4654859612905966, + "tokens_seen": 1954349056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020600802407221667, + "loss": 1.361, + "theoretical_loss": 3.465476308105945, + "tokens_seen": 1954414592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020599799398194585, + "loss": 1.3504, + "theoretical_loss": 3.4654666553356126, + "tokens_seen": 1954480128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020598796389167503, + "loss": 1.4047, + "theoretical_loss": 3.4654570029795666, + "tokens_seen": 1954545664 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020597793380140424, + "loss": 1.3098, + "theoretical_loss": 3.465447351037776, + "tokens_seen": 1954611200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020596790371113342, + "loss": 1.3758, + "theoretical_loss": 3.4654376995102094, + "tokens_seen": 1954676736 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002059578736208626, + "loss": 1.4649, + "theoretical_loss": 3.4654280483968347, + "tokens_seen": 1954742272 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020594784353059178, + "loss": 1.3295, + "theoretical_loss": 3.4654183976976203, + "tokens_seen": 1954807808 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020593781344032096, + "loss": 1.3037, + "theoretical_loss": 3.4654087474125346, + "tokens_seen": 1954873344 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020592778335005017, + "loss": 1.4205, + "theoretical_loss": 3.4653990975415456, + "tokens_seen": 1954938880 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020591775325977935, + "loss": 1.3748, + "theoretical_loss": 3.4653894480846223, + "tokens_seen": 1955004416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020590772316950853, + "loss": 1.3515, + "theoretical_loss": 3.4653797990417328, + "tokens_seen": 1955069952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020589769307923772, + "loss": 1.3411, + "theoretical_loss": 3.465370150412845, + "tokens_seen": 1955135488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020588766298896692, + "loss": 1.3934, + "theoretical_loss": 3.465360502197928, + "tokens_seen": 1955201024 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002058776328986961, + "loss": 1.3404, + "theoretical_loss": 3.4653508543969496, + "tokens_seen": 1955266560 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002058676028084253, + "loss": 1.3152, + "theoretical_loss": 3.465341207009878, + "tokens_seen": 1955332096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020585757271815447, + "loss": 1.4353, + "theoretical_loss": 3.4653315600366827, + "tokens_seen": 1955397632 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020584754262788365, + "loss": 1.3789, + "theoretical_loss": 3.465321913477331, + "tokens_seen": 1955463168 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020583751253761286, + "loss": 1.3757, + "theoretical_loss": 3.4653122673317913, + "tokens_seen": 1955528704 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020582748244734204, + "loss": 1.3337, + "theoretical_loss": 3.4653026216000327, + "tokens_seen": 1955594240 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020581745235707122, + "loss": 1.4491, + "theoretical_loss": 3.465292976282023, + "tokens_seen": 1955659776 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002058074222668004, + "loss": 1.2665, + "theoretical_loss": 3.465283331377731, + "tokens_seen": 1955725312 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002057973921765296, + "loss": 1.3223, + "theoretical_loss": 3.465273686887124, + "tokens_seen": 1955790848 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002057873620862588, + "loss": 1.3371, + "theoretical_loss": 3.4652640428101718, + "tokens_seen": 1955856384 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020577733199598797, + "loss": 1.3642, + "theoretical_loss": 3.4652543991468425, + "tokens_seen": 1955921920 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020576730190571715, + "loss": 1.3513, + "theoretical_loss": 3.4652447558971042, + "tokens_seen": 1955987456 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020575727181544633, + "loss": 1.3482, + "theoretical_loss": 3.465235113060925, + "tokens_seen": 1956052992 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020574724172517554, + "loss": 1.3563, + "theoretical_loss": 3.4652254706382735, + "tokens_seen": 1956118528 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020573721163490472, + "loss": 1.3735, + "theoretical_loss": 3.4652158286291184, + "tokens_seen": 1956184064 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 713197, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.288440465927124, + "objective/train/theoretical_loss": 3.465206187033428, + "objective/train/tokens_used": 1976709600, + "theoretical_loss": 3.465206187033428, + "tokens_seen": 1956249600 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002057271815446339, + "loss": 1.3084, + "theoretical_loss": 3.465206187033428, + "tokens_seen": 1956249600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020571715145436308, + "loss": 1.3772, + "theoretical_loss": 3.465196545851171, + "tokens_seen": 1956315136 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002057071213640923, + "loss": 1.3889, + "theoretical_loss": 3.465186905082315, + "tokens_seen": 1956380672 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020569709127382147, + "loss": 1.3859, + "theoretical_loss": 3.4651772647268295, + "tokens_seen": 1956446208 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020568706118355066, + "loss": 1.232, + "theoretical_loss": 3.465167624784682, + "tokens_seen": 1956511744 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020567703109327984, + "loss": 1.4754, + "theoretical_loss": 3.4651579852558414, + "tokens_seen": 1956577280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020566700100300902, + "loss": 1.3664, + "theoretical_loss": 3.465148346140276, + "tokens_seen": 1956642816 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020565697091273823, + "loss": 1.2843, + "theoretical_loss": 3.4651387074379545, + "tokens_seen": 1956708352 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002056469408224674, + "loss": 1.3371, + "theoretical_loss": 3.465129069148845, + "tokens_seen": 1956773888 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002056369107321966, + "loss": 1.4323, + "theoretical_loss": 3.4651194312729157, + "tokens_seen": 1956839424 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020562688064192577, + "loss": 1.297, + "theoretical_loss": 3.465109793810136, + "tokens_seen": 1956904960 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020561685055165498, + "loss": 1.386, + "theoretical_loss": 3.4651001567604736, + "tokens_seen": 1956970496 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020560682046138416, + "loss": 1.3909, + "theoretical_loss": 3.465090520123897, + "tokens_seen": 1957036032 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020559679037111334, + "loss": 1.4286, + "theoretical_loss": 3.465080883900375, + "tokens_seen": 1957101568 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020558676028084252, + "loss": 1.3133, + "theoretical_loss": 3.4650712480898758, + "tokens_seen": 1957167104 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002055767301905717, + "loss": 1.3309, + "theoretical_loss": 3.4650616126923675, + "tokens_seen": 1957232640 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002055667001003009, + "loss": 1.382, + "theoretical_loss": 3.4650519777078195, + "tokens_seen": 1957298176 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002055566700100301, + "loss": 1.2649, + "theoretical_loss": 3.4650423431362, + "tokens_seen": 1957363712 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020554663991975927, + "loss": 1.2872, + "theoretical_loss": 3.4650327089774766, + "tokens_seen": 1957429248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020553660982948845, + "loss": 1.5066, + "theoretical_loss": 3.4650230752316187, + "tokens_seen": 1957494784 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020552657973921766, + "loss": 1.3976, + "theoretical_loss": 3.465013441898595, + "tokens_seen": 1957560320 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020551654964894684, + "loss": 1.398, + "theoretical_loss": 3.4650038089783726, + "tokens_seen": 1957625856 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020550651955867602, + "loss": 1.3605, + "theoretical_loss": 3.4649941764709213, + "tokens_seen": 1957691392 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002054964894684052, + "loss": 1.4739, + "theoretical_loss": 3.4649845443762093, + "tokens_seen": 1957756928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020548645937813439, + "loss": 1.4391, + "theoretical_loss": 3.4649749126942053, + "tokens_seen": 1957822464 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002054764292878636, + "loss": 1.3329, + "theoretical_loss": 3.4649652814248775, + "tokens_seen": 1957888000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020546639919759278, + "loss": 1.2891, + "theoretical_loss": 3.4649556505681938, + "tokens_seen": 1957953536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020545636910732196, + "loss": 1.4088, + "theoretical_loss": 3.4649460201241236, + "tokens_seen": 1958019072 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020544633901705114, + "loss": 1.315, + "theoretical_loss": 3.4649363900926353, + "tokens_seen": 1958084608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020543630892678035, + "loss": 1.3077, + "theoretical_loss": 3.4649267604736966, + "tokens_seen": 1958150144 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020542627883650953, + "loss": 1.381, + "theoretical_loss": 3.4649171312672777, + "tokens_seen": 1958215680 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002054162487462387, + "loss": 1.3322, + "theoretical_loss": 3.4649075024733453, + "tokens_seen": 1958281216 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002054062186559679, + "loss": 1.3304, + "theoretical_loss": 3.464897874091869, + "tokens_seen": 1958346752 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002053961885656971, + "loss": 1.2834, + "theoretical_loss": 3.4648882461228174, + "tokens_seen": 1958412288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020538615847542628, + "loss": 1.3489, + "theoretical_loss": 3.464878618566158, + "tokens_seen": 1958477824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020537612838515546, + "loss": 1.3333, + "theoretical_loss": 3.4648689914218602, + "tokens_seen": 1958543360 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020536609829488464, + "loss": 1.4099, + "theoretical_loss": 3.4648593646898926, + "tokens_seen": 1958608896 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020535606820461382, + "loss": 1.3364, + "theoretical_loss": 3.4648497383702237, + "tokens_seen": 1958674432 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020534603811434303, + "loss": 1.4536, + "theoretical_loss": 3.464840112462821, + "tokens_seen": 1958739968 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002053360080240722, + "loss": 1.2568, + "theoretical_loss": 3.4648304869676547, + "tokens_seen": 1958805504 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020532597793380142, + "loss": 1.3505, + "theoretical_loss": 3.464820861884692, + "tokens_seen": 1958871040 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002053159478435306, + "loss": 1.3354, + "theoretical_loss": 3.464811237213903, + "tokens_seen": 1958936576 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002053059177532598, + "loss": 1.3799, + "theoretical_loss": 3.464801612955254, + "tokens_seen": 1959002112 + }, + { + "epoch": 0.59, + "learning_rate": 0.000205295887662989, + "loss": 1.351, + "theoretical_loss": 3.4647919891087158, + "tokens_seen": 1959067648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020528585757271817, + "loss": 1.3938, + "theoretical_loss": 3.4647823656742553, + "tokens_seen": 1959133184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020527582748244735, + "loss": 1.3697, + "theoretical_loss": 3.464772742651842, + "tokens_seen": 1959198720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020526579739217653, + "loss": 1.3771, + "theoretical_loss": 3.464763120041445, + "tokens_seen": 1959264256 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020525576730190574, + "loss": 1.3024, + "theoretical_loss": 3.4647534978430308, + "tokens_seen": 1959329792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020524573721163492, + "loss": 1.3616, + "theoretical_loss": 3.4647438760565703, + "tokens_seen": 1959395328 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002052357071213641, + "loss": 1.2793, + "theoretical_loss": 3.464734254682031, + "tokens_seen": 1959460864 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 714425, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.421403169631958, + "objective/train/theoretical_loss": 3.4647246337193813, + "objective/train/tokens_used": 1979986400, + "theoretical_loss": 3.4647246337193813, + "tokens_seen": 1959526400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020522567703109328, + "loss": 1.4015, + "theoretical_loss": 3.4647246337193813, + "tokens_seen": 1959526400 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002052156469408225, + "loss": 1.3835, + "theoretical_loss": 3.4647150131685898, + "tokens_seen": 1959591936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020520561685055167, + "loss": 1.3344, + "theoretical_loss": 3.464705393029626, + "tokens_seen": 1959657472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020519558676028086, + "loss": 1.3978, + "theoretical_loss": 3.4646957733024575, + "tokens_seen": 1959723008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020518555667001004, + "loss": 1.371, + "theoretical_loss": 3.4646861539870537, + "tokens_seen": 1959788544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020517552657973922, + "loss": 1.3421, + "theoretical_loss": 3.4646765350833824, + "tokens_seen": 1959854080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020516549648946843, + "loss": 1.287, + "theoretical_loss": 3.464666916591413, + "tokens_seen": 1959919616 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002051554663991976, + "loss": 1.261, + "theoretical_loss": 3.4646572985111126, + "tokens_seen": 1959985152 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002051454363089268, + "loss": 1.4233, + "theoretical_loss": 3.464647680842452, + "tokens_seen": 1960050688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020513540621865597, + "loss": 1.2878, + "theoretical_loss": 3.464638063585398, + "tokens_seen": 1960116224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020512537612838518, + "loss": 1.3707, + "theoretical_loss": 3.464628446739921, + "tokens_seen": 1960181760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020511534603811436, + "loss": 1.4208, + "theoretical_loss": 3.464618830305988, + "tokens_seen": 1960247296 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020510531594784354, + "loss": 1.4447, + "theoretical_loss": 3.464609214283568, + "tokens_seen": 1960312832 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020509528585757272, + "loss": 1.3644, + "theoretical_loss": 3.4645995986726303, + "tokens_seen": 1960378368 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002050852557673019, + "loss": 1.387, + "theoretical_loss": 3.464589983473143, + "tokens_seen": 1960443904 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002050752256770311, + "loss": 1.4481, + "theoretical_loss": 3.4645803686850747, + "tokens_seen": 1960509440 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002050651955867603, + "loss": 1.3827, + "theoretical_loss": 3.464570754308394, + "tokens_seen": 1960574976 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020505516549648947, + "loss": 1.2769, + "theoretical_loss": 3.4645611403430703, + "tokens_seen": 1960640512 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020504513540621865, + "loss": 1.5119, + "theoretical_loss": 3.464551526789071, + "tokens_seen": 1960706048 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020503510531594786, + "loss": 1.3515, + "theoretical_loss": 3.464541913646366, + "tokens_seen": 1960771584 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020502507522567704, + "loss": 1.3888, + "theoretical_loss": 3.4645323009149234, + "tokens_seen": 1960837120 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020501504513540622, + "loss": 1.345, + "theoretical_loss": 3.4645226885947116, + "tokens_seen": 1960902656 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002050050150451354, + "loss": 1.4272, + "theoretical_loss": 3.4645130766856993, + "tokens_seen": 1960968192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020499498495486459, + "loss": 1.3089, + "theoretical_loss": 3.464503465187856, + "tokens_seen": 1961033728 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002049849548645938, + "loss": 1.3794, + "theoretical_loss": 3.464493854101149, + "tokens_seen": 1961099264 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020497492477432298, + "loss": 1.3445, + "theoretical_loss": 3.464484243425548, + "tokens_seen": 1961164800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020496489468405216, + "loss": 1.4615, + "theoretical_loss": 3.4644746331610214, + "tokens_seen": 1961230336 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020495486459378134, + "loss": 1.3498, + "theoretical_loss": 3.464465023307538, + "tokens_seen": 1961295872 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020494483450351055, + "loss": 1.3054, + "theoretical_loss": 3.4644554138650667, + "tokens_seen": 1961361408 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020493480441323973, + "loss": 1.3348, + "theoretical_loss": 3.4644458048335753, + "tokens_seen": 1961426944 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002049247743229689, + "loss": 1.333, + "theoretical_loss": 3.4644361962130334, + "tokens_seen": 1961492480 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002049147442326981, + "loss": 1.4344, + "theoretical_loss": 3.464426588003409, + "tokens_seen": 1961558016 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002049047141424273, + "loss": 1.4037, + "theoretical_loss": 3.4644169802046716, + "tokens_seen": 1961623552 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020489468405215648, + "loss": 1.4186, + "theoretical_loss": 3.4644073728167895, + "tokens_seen": 1961689088 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020488465396188566, + "loss": 1.316, + "theoretical_loss": 3.4643977658397307, + "tokens_seen": 1961754624 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020487462387161484, + "loss": 1.2951, + "theoretical_loss": 3.464388159273465, + "tokens_seen": 1961820160 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020486459378134402, + "loss": 1.3788, + "theoretical_loss": 3.4643785531179607, + "tokens_seen": 1961885696 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020485456369107323, + "loss": 1.3524, + "theoretical_loss": 3.4643689473731865, + "tokens_seen": 1961951232 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002048445336008024, + "loss": 1.2546, + "theoretical_loss": 3.464359342039111, + "tokens_seen": 1962016768 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002048345035105316, + "loss": 1.2948, + "theoretical_loss": 3.464349737115703, + "tokens_seen": 1962082304 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020482447342026077, + "loss": 1.3275, + "theoretical_loss": 3.4643401326029313, + "tokens_seen": 1962147840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020481444332998998, + "loss": 1.2906, + "theoretical_loss": 3.464330528500765, + "tokens_seen": 1962213376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020480441323971916, + "loss": 1.3487, + "theoretical_loss": 3.464320924809172, + "tokens_seen": 1962278912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020479438314944834, + "loss": 1.3337, + "theoretical_loss": 3.464311321528122, + "tokens_seen": 1962344448 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020478435305917753, + "loss": 1.3711, + "theoretical_loss": 3.464301718657582, + "tokens_seen": 1962409984 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002047743229689067, + "loss": 1.3917, + "theoretical_loss": 3.464292116197523, + "tokens_seen": 1962475520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020476429287863591, + "loss": 1.3836, + "theoretical_loss": 3.4642825141479126, + "tokens_seen": 1962541056 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002047542627883651, + "loss": 1.3174, + "theoretical_loss": 3.4642729125087195, + "tokens_seen": 1962606592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020474423269809428, + "loss": 1.3527, + "theoretical_loss": 3.464263311279912, + "tokens_seen": 1962672128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020473420260782346, + "loss": 1.3627, + "theoretical_loss": 3.4642537104614606, + "tokens_seen": 1962737664 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 715564, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2140882015228271, + "objective/train/theoretical_loss": 3.464244110053332, + "objective/train/tokens_used": 1983263200, + "theoretical_loss": 3.464244110053332, + "tokens_seen": 1962803200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020472417251755267, + "loss": 1.342, + "theoretical_loss": 3.464244110053332, + "tokens_seen": 1962803200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020471414242728185, + "loss": 1.4226, + "theoretical_loss": 3.4642345100554963, + "tokens_seen": 1962868736 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020470411233701103, + "loss": 1.3578, + "theoretical_loss": 3.4642249104679217, + "tokens_seen": 1962934272 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002046940822467402, + "loss": 1.3409, + "theoretical_loss": 3.464215311290577, + "tokens_seen": 1962999808 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002046840521564694, + "loss": 1.2912, + "theoretical_loss": 3.4642057125234316, + "tokens_seen": 1963065344 + }, + { + "epoch": 0.59, + "learning_rate": 0.0002046740220661986, + "loss": 1.4602, + "theoretical_loss": 3.4641961141664535, + "tokens_seen": 1963130880 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020466399197592778, + "loss": 1.4438, + "theoretical_loss": 3.4641865162196117, + "tokens_seen": 1963196416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020465396188565696, + "loss": 1.4308, + "theoretical_loss": 3.464176918682875, + "tokens_seen": 1963261952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020464393179538614, + "loss": 1.3859, + "theoretical_loss": 3.464167321556212, + "tokens_seen": 1963327488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020463390170511535, + "loss": 1.3806, + "theoretical_loss": 3.464157724839592, + "tokens_seen": 1963393024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00020462387161484453, + "loss": 1.4022, + "theoretical_loss": 3.4641481285329836, + "tokens_seen": 1963458560 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002046138415245737, + "loss": 1.3851, + "theoretical_loss": 3.4641385326363556, + "tokens_seen": 1963524096 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002046038114343029, + "loss": 1.3738, + "theoretical_loss": 3.4641289371496766, + "tokens_seen": 1963589632 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020459378134403207, + "loss": 1.4176, + "theoretical_loss": 3.464119342072915, + "tokens_seen": 1963655168 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020458375125376128, + "loss": 1.3629, + "theoretical_loss": 3.4641097474060407, + "tokens_seen": 1963720704 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002045737211634905, + "loss": 1.2904, + "theoretical_loss": 3.4641001531490216, + "tokens_seen": 1963786240 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020456369107321967, + "loss": 1.2983, + "theoretical_loss": 3.464090559301827, + "tokens_seen": 1963851776 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020455366098294885, + "loss": 1.3545, + "theoretical_loss": 3.464080965864426, + "tokens_seen": 1963917312 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020454363089267806, + "loss": 1.3582, + "theoretical_loss": 3.4640713728367865, + "tokens_seen": 1963982848 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020453360080240724, + "loss": 1.3518, + "theoretical_loss": 3.464061780218878, + "tokens_seen": 1964048384 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020452357071213642, + "loss": 1.3435, + "theoretical_loss": 3.464052188010669, + "tokens_seen": 1964113920 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002045135406218656, + "loss": 1.3637, + "theoretical_loss": 3.4640425962121286, + "tokens_seen": 1964179456 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020450351053159479, + "loss": 1.2851, + "theoretical_loss": 3.464033004823225, + "tokens_seen": 1964244992 + }, + { + "epoch": 0.6, + "learning_rate": 0.000204493480441324, + "loss": 1.3238, + "theoretical_loss": 3.464023413843928, + "tokens_seen": 1964310528 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020448345035105318, + "loss": 1.3557, + "theoretical_loss": 3.4640138232742066, + "tokens_seen": 1964376064 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020447342026078236, + "loss": 1.3809, + "theoretical_loss": 3.464004233114028, + "tokens_seen": 1964441600 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020446339017051154, + "loss": 1.3085, + "theoretical_loss": 3.4639946433633626, + "tokens_seen": 1964507136 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020445336008024075, + "loss": 1.3998, + "theoretical_loss": 3.463985054022179, + "tokens_seen": 1964572672 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020444332998996993, + "loss": 1.2752, + "theoretical_loss": 3.4639754650904453, + "tokens_seen": 1964638208 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002044332998996991, + "loss": 1.3297, + "theoretical_loss": 3.463965876568131, + "tokens_seen": 1964703744 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002044232698094283, + "loss": 1.4139, + "theoretical_loss": 3.463956288455205, + "tokens_seen": 1964769280 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002044132397191575, + "loss": 1.2938, + "theoretical_loss": 3.463946700751636, + "tokens_seen": 1964834816 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020440320962888668, + "loss": 1.3178, + "theoretical_loss": 3.4639371134573924, + "tokens_seen": 1964900352 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020439317953861586, + "loss": 1.2804, + "theoretical_loss": 3.463927526572444, + "tokens_seen": 1964965888 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020438314944834504, + "loss": 1.38, + "theoretical_loss": 3.463917940096759, + "tokens_seen": 1965031424 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020437311935807422, + "loss": 1.3274, + "theoretical_loss": 3.4639083540303064, + "tokens_seen": 1965096960 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020436308926780343, + "loss": 1.3205, + "theoretical_loss": 3.4638987683730553, + "tokens_seen": 1965162496 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002043530591775326, + "loss": 1.4059, + "theoretical_loss": 3.4638891831249747, + "tokens_seen": 1965228032 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002043430290872618, + "loss": 1.3417, + "theoretical_loss": 3.463879598286033, + "tokens_seen": 1965293568 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020433299899699097, + "loss": 1.4128, + "theoretical_loss": 3.4638700138561993, + "tokens_seen": 1965359104 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020432296890672018, + "loss": 1.2242, + "theoretical_loss": 3.4638604298354427, + "tokens_seen": 1965424640 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020431293881644936, + "loss": 1.336, + "theoretical_loss": 3.463850846223732, + "tokens_seen": 1965490176 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020430290872617854, + "loss": 1.403, + "theoretical_loss": 3.4638412630210356, + "tokens_seen": 1965555712 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020429287863590773, + "loss": 1.4091, + "theoretical_loss": 3.4638316802273232, + "tokens_seen": 1965621248 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002042828485456369, + "loss": 1.2832, + "theoretical_loss": 3.463822097842564, + "tokens_seen": 1965686784 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020427281845536611, + "loss": 1.3442, + "theoretical_loss": 3.4638125158667252, + "tokens_seen": 1965752320 + }, + { + "epoch": 0.6, + "learning_rate": 0.0002042627883650953, + "loss": 1.4381, + "theoretical_loss": 3.4638029342997774, + "tokens_seen": 1965817856 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020425275827482448, + "loss": 1.3466, + "theoretical_loss": 3.4637933531416882, + "tokens_seen": 1965883392 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020424272818455366, + "loss": 1.339, + "theoretical_loss": 3.4637837723924276, + "tokens_seen": 1965948928 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020423269809428287, + "loss": 1.2322, + "theoretical_loss": 3.4637741920519645, + "tokens_seen": 1966014464 + }, + { + "debugging/distinct-1-grams": 0.7394139103740143, + "debugging/entropy-1-grams": 5.1329642330953185, + "debugging/length": 486.55555555555554, + "debugging/num_segments": 9, + "debugging/score": 0.00915515707107717, + "debugging/score_std": 0.01661134296838666, + "epoch": 0.6, + "objective/train/docs_used": 716760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4679943323135376, + "objective/train/theoretical_loss": 3.4637646121202676, + "objective/train/tokens_used": 1986540000, + "theoretical_loss": 3.4637646121202676, + "tokens_seen": 1966080000 + }, + { + "epoch": 0.6, + "learning_rate": 0.00020422266800401205, + "loss": 1.4388, + "theoretical_loss": 3.4637646121202676, + "tokens_seen": 1966080000 + } + ], + "max_steps": 50354, + "num_train_epochs": 9223372036854775807, + "total_flos": 1.00336140288e+18, + "trial_name": null, + "trial_params": null +}