diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,8084 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.0, + "eval_steps": 500, + "global_step": 4024, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.002485089463220676, + "grad_norm": 1.48516430932384, + "learning_rate": 9.99997561945331e-06, + "loss": 0.7547443389892579, + "num_input_tokens_seen": 526856, + "step": 5, + "train_runtime": 72.4525, + "train_tokens_per_second": 7271.743 + }, + { + "epoch": 0.004970178926441352, + "grad_norm": 1.0499373091515534, + "learning_rate": 9.999876573889877e-06, + "loss": 0.6137888908386231, + "num_input_tokens_seen": 1092088, + "step": 10, + "train_runtime": 136.4629, + "train_tokens_per_second": 8002.82 + }, + { + "epoch": 0.007455268389662028, + "grad_norm": 0.7435657135826963, + "learning_rate": 9.99970134103361e-06, + "loss": 0.568207836151123, + "num_input_tokens_seen": 1650032, + "step": 15, + "train_runtime": 203.4614, + "train_tokens_per_second": 8109.804 + }, + { + "epoch": 0.009940357852882704, + "grad_norm": 0.8076060280076128, + "learning_rate": 9.99944992355468e-06, + "loss": 0.526887321472168, + "num_input_tokens_seen": 2163080, + "step": 20, + "train_runtime": 278.4156, + "train_tokens_per_second": 7769.248 + }, + { + "epoch": 0.01242544731610338, + "grad_norm": 0.9031987244014792, + "learning_rate": 9.999122325284146e-06, + "loss": 0.5160922050476074, + "num_input_tokens_seen": 2714856, + "step": 25, + "train_runtime": 349.361, + "train_tokens_per_second": 7770.919 + }, + { + "epoch": 0.014910536779324055, + "grad_norm": 0.6662354297886746, + "learning_rate": 9.998718551213894e-06, + "loss": 0.5206570625305176, + "num_input_tokens_seen": 3271808, + "step": 30, + "train_runtime": 426.1933, + "train_tokens_per_second": 7676.818 + }, + { + "epoch": 0.01739562624254473, + "grad_norm": 0.7499687523197902, + "learning_rate": 9.998238607496569e-06, + "loss": 0.5006282806396485, + "num_input_tokens_seen": 3805000, + "step": 35, + "train_runtime": 502.4142, + "train_tokens_per_second": 7573.433 + }, + { + "epoch": 0.019880715705765408, + "grad_norm": 0.7001249778043209, + "learning_rate": 9.99768250144547e-06, + "loss": 0.4811150550842285, + "num_input_tokens_seen": 4325816, + "step": 40, + "train_runtime": 570.5961, + "train_tokens_per_second": 7581.223 + }, + { + "epoch": 0.022365805168986085, + "grad_norm": 0.7014623312711038, + "learning_rate": 9.997050241534455e-06, + "loss": 0.4966543197631836, + "num_input_tokens_seen": 4860112, + "step": 45, + "train_runtime": 636.028, + "train_tokens_per_second": 7641.349 + }, + { + "epoch": 0.02485089463220676, + "grad_norm": 0.7924304172544439, + "learning_rate": 9.99634183739779e-06, + "loss": 0.4881586074829102, + "num_input_tokens_seen": 5401472, + "step": 50, + "train_runtime": 710.8997, + "train_tokens_per_second": 7598.079 + }, + { + "epoch": 0.027335984095427434, + "grad_norm": 0.7302555340482748, + "learning_rate": 9.995557299830024e-06, + "loss": 0.49920997619628904, + "num_input_tokens_seen": 5916648, + "step": 55, + "train_runtime": 787.1096, + "train_tokens_per_second": 7516.931 + }, + { + "epoch": 0.02982107355864811, + "grad_norm": 0.7470311250225292, + "learning_rate": 9.99469664078581e-06, + "loss": 0.48033866882324217, + "num_input_tokens_seen": 6503416, + "step": 60, + "train_runtime": 856.9915, + "train_tokens_per_second": 7588.659 + }, + { + "epoch": 0.03230616302186879, + "grad_norm": 0.7243488236400842, + "learning_rate": 9.993759873379728e-06, + "loss": 0.47772536277770994, + "num_input_tokens_seen": 7049000, + "step": 65, + "train_runtime": 926.4752, + "train_tokens_per_second": 7608.406 + }, + { + "epoch": 0.03479125248508946, + "grad_norm": 0.7105041057933622, + "learning_rate": 9.992747011886083e-06, + "loss": 0.4617642879486084, + "num_input_tokens_seen": 7605432, + "step": 70, + "train_runtime": 985.1402, + "train_tokens_per_second": 7720.152 + }, + { + "epoch": 0.037276341948310136, + "grad_norm": 0.6817164427173915, + "learning_rate": 9.991658071738696e-06, + "loss": 0.4472812652587891, + "num_input_tokens_seen": 8170688, + "step": 75, + "train_runtime": 1062.2492, + "train_tokens_per_second": 7691.875 + }, + { + "epoch": 0.039761431411530816, + "grad_norm": 0.6822320863803008, + "learning_rate": 9.990493069530652e-06, + "loss": 0.45863828659057615, + "num_input_tokens_seen": 8779096, + "step": 80, + "train_runtime": 1125.2207, + "train_tokens_per_second": 7802.11 + }, + { + "epoch": 0.04224652087475149, + "grad_norm": 0.7320483930776894, + "learning_rate": 9.989252023014064e-06, + "loss": 0.4660465717315674, + "num_input_tokens_seen": 9301592, + "step": 85, + "train_runtime": 1201.0781, + "train_tokens_per_second": 7744.369 + }, + { + "epoch": 0.04473161033797217, + "grad_norm": 0.6477582273980964, + "learning_rate": 9.987934951099794e-06, + "loss": 0.47597427368164064, + "num_input_tokens_seen": 9842296, + "step": 90, + "train_runtime": 1263.0055, + "train_tokens_per_second": 7792.758 + }, + { + "epoch": 0.04721669980119284, + "grad_norm": 0.6692134288771309, + "learning_rate": 9.986541873857167e-06, + "loss": 0.4624770164489746, + "num_input_tokens_seen": 10408600, + "step": 95, + "train_runtime": 1332.1107, + "train_tokens_per_second": 7813.615 + }, + { + "epoch": 0.04970178926441352, + "grad_norm": 0.6772254972685579, + "learning_rate": 9.985072812513661e-06, + "loss": 0.4523159980773926, + "num_input_tokens_seen": 11012720, + "step": 100, + "train_runtime": 1400.2315, + "train_tokens_per_second": 7864.928 + }, + { + "epoch": 0.052186878727634195, + "grad_norm": 0.6777537636498243, + "learning_rate": 9.983527789454593e-06, + "loss": 0.46146764755249026, + "num_input_tokens_seen": 11595744, + "step": 105, + "train_runtime": 1465.7185, + "train_tokens_per_second": 7911.304 + }, + { + "epoch": 0.05467196819085487, + "grad_norm": 0.8158227350097417, + "learning_rate": 9.981906828222764e-06, + "loss": 0.4701869010925293, + "num_input_tokens_seen": 12145384, + "step": 110, + "train_runtime": 1532.5514, + "train_tokens_per_second": 7924.944 + }, + { + "epoch": 0.05715705765407555, + "grad_norm": 1.0309855809717265, + "learning_rate": 9.98020995351812e-06, + "loss": 0.45644779205322267, + "num_input_tokens_seen": 12717448, + "step": 115, + "train_runtime": 1605.5364, + "train_tokens_per_second": 7920.996 + }, + { + "epoch": 0.05964214711729622, + "grad_norm": 0.7546116526019456, + "learning_rate": 9.978437191197354e-06, + "loss": 0.4323369026184082, + "num_input_tokens_seen": 13275968, + "step": 120, + "train_runtime": 1678.1785, + "train_tokens_per_second": 7910.939 + }, + { + "epoch": 0.0621272365805169, + "grad_norm": 0.8425049347947811, + "learning_rate": 9.976588568273526e-06, + "loss": 0.44097166061401366, + "num_input_tokens_seen": 13848096, + "step": 125, + "train_runtime": 1733.1806, + "train_tokens_per_second": 7989.99 + }, + { + "epoch": 0.06461232604373758, + "grad_norm": 0.7248966416044177, + "learning_rate": 9.974664112915644e-06, + "loss": 0.44237565994262695, + "num_input_tokens_seen": 14389600, + "step": 130, + "train_runtime": 1805.7036, + "train_tokens_per_second": 7968.971 + }, + { + "epoch": 0.06709741550695825, + "grad_norm": 0.7170274064378713, + "learning_rate": 9.972663854448244e-06, + "loss": 0.4436997413635254, + "num_input_tokens_seen": 14942624, + "step": 135, + "train_runtime": 1870.7248, + "train_tokens_per_second": 7987.612 + }, + { + "epoch": 0.06958250497017893, + "grad_norm": 0.729330446275624, + "learning_rate": 9.970587823350933e-06, + "loss": 0.4589400768280029, + "num_input_tokens_seen": 15508880, + "step": 140, + "train_runtime": 1928.649, + "train_tokens_per_second": 8041.318 + }, + { + "epoch": 0.0720675944333996, + "grad_norm": 0.7217377471042493, + "learning_rate": 9.968436051257927e-06, + "loss": 0.43112521171569823, + "num_input_tokens_seen": 16067408, + "step": 145, + "train_runtime": 2001.9902, + "train_tokens_per_second": 8025.718 + }, + { + "epoch": 0.07455268389662027, + "grad_norm": 0.7417357330469618, + "learning_rate": 9.966208570957576e-06, + "loss": 0.43833422660827637, + "num_input_tokens_seen": 16641360, + "step": 150, + "train_runtime": 2074.374, + "train_tokens_per_second": 8022.353 + }, + { + "epoch": 0.07703777335984095, + "grad_norm": 0.661495628429147, + "learning_rate": 9.963905416391858e-06, + "loss": 0.4669617176055908, + "num_input_tokens_seen": 17205544, + "step": 155, + "train_runtime": 2148.2766, + "train_tokens_per_second": 8008.999 + }, + { + "epoch": 0.07952286282306163, + "grad_norm": 1.0634325439600465, + "learning_rate": 9.96152662265586e-06, + "loss": 0.47125539779663084, + "num_input_tokens_seen": 17787240, + "step": 160, + "train_runtime": 2223.2045, + "train_tokens_per_second": 8000.721 + }, + { + "epoch": 0.08200795228628231, + "grad_norm": 0.755507375336271, + "learning_rate": 9.959072225997246e-06, + "loss": 0.42906894683837893, + "num_input_tokens_seen": 18356552, + "step": 165, + "train_runtime": 2286.4492, + "train_tokens_per_second": 8028.41 + }, + { + "epoch": 0.08449304174950298, + "grad_norm": 0.8311318326618452, + "learning_rate": 9.95654226381571e-06, + "loss": 0.4371389389038086, + "num_input_tokens_seen": 18889552, + "step": 170, + "train_runtime": 2352.8397, + "train_tokens_per_second": 8028.406 + }, + { + "epoch": 0.08697813121272366, + "grad_norm": 0.6897368738648079, + "learning_rate": 9.953936774662396e-06, + "loss": 0.43525238037109376, + "num_input_tokens_seen": 19454128, + "step": 175, + "train_runtime": 2427.6715, + "train_tokens_per_second": 8013.493 + }, + { + "epoch": 0.08946322067594434, + "grad_norm": 0.7993617258900514, + "learning_rate": 9.951255798239317e-06, + "loss": 0.4044791698455811, + "num_input_tokens_seen": 19996088, + "step": 180, + "train_runtime": 2494.0017, + "train_tokens_per_second": 8017.672 + }, + { + "epoch": 0.091948310139165, + "grad_norm": 0.7087103621246671, + "learning_rate": 9.948499375398746e-06, + "loss": 0.44888954162597655, + "num_input_tokens_seen": 20554328, + "step": 185, + "train_runtime": 2564.496, + "train_tokens_per_second": 8014.958 + }, + { + "epoch": 0.09443339960238568, + "grad_norm": 0.6872240626380537, + "learning_rate": 9.945667548142603e-06, + "loss": 0.40648441314697265, + "num_input_tokens_seen": 21111368, + "step": 190, + "train_runtime": 2624.6694, + "train_tokens_per_second": 8043.439 + }, + { + "epoch": 0.09691848906560636, + "grad_norm": 0.7442086897543689, + "learning_rate": 9.942760359621803e-06, + "loss": 0.42204437255859373, + "num_input_tokens_seen": 21664656, + "step": 195, + "train_runtime": 2690.9339, + "train_tokens_per_second": 8050.98 + }, + { + "epoch": 0.09940357852882704, + "grad_norm": 0.7173717455657352, + "learning_rate": 9.939777854135603e-06, + "loss": 0.41335372924804686, + "num_input_tokens_seen": 22254552, + "step": 200, + "train_runtime": 2758.8818, + "train_tokens_per_second": 8066.512 + }, + { + "epoch": 0.10188866799204771, + "grad_norm": 0.6731061622522566, + "learning_rate": 9.936720077130928e-06, + "loss": 0.4478871822357178, + "num_input_tokens_seen": 22804024, + "step": 205, + "train_runtime": 2826.2431, + "train_tokens_per_second": 8068.671 + }, + { + "epoch": 0.10437375745526839, + "grad_norm": 0.6380671912881155, + "learning_rate": 9.933587075201678e-06, + "loss": 0.42775816917419435, + "num_input_tokens_seen": 23356848, + "step": 210, + "train_runtime": 2900.2086, + "train_tokens_per_second": 8053.506 + }, + { + "epoch": 0.10685884691848907, + "grad_norm": 0.6839656053674485, + "learning_rate": 9.930378896088022e-06, + "loss": 0.46932339668273926, + "num_input_tokens_seen": 23923904, + "step": 215, + "train_runtime": 2966.0931, + "train_tokens_per_second": 8065.797 + }, + { + "epoch": 0.10934393638170974, + "grad_norm": 0.6294228660792941, + "learning_rate": 9.92709558867566e-06, + "loss": 0.4336122512817383, + "num_input_tokens_seen": 24486352, + "step": 220, + "train_runtime": 3025.8796, + "train_tokens_per_second": 8092.309 + }, + { + "epoch": 0.11182902584493042, + "grad_norm": 0.661004546035617, + "learning_rate": 9.923737202995089e-06, + "loss": 0.41347699165344237, + "num_input_tokens_seen": 25055736, + "step": 225, + "train_runtime": 3107.3681, + "train_tokens_per_second": 8063.33 + }, + { + "epoch": 0.1143141153081511, + "grad_norm": 0.7423836723730006, + "learning_rate": 9.920303790220835e-06, + "loss": 0.41893601417541504, + "num_input_tokens_seen": 25605264, + "step": 230, + "train_runtime": 3180.8869, + "train_tokens_per_second": 8049.725 + }, + { + "epoch": 0.11679920477137178, + "grad_norm": 0.7794748050178208, + "learning_rate": 9.916795402670674e-06, + "loss": 0.4319512367248535, + "num_input_tokens_seen": 26154472, + "step": 235, + "train_runtime": 3256.7537, + "train_tokens_per_second": 8030.841 + }, + { + "epoch": 0.11928429423459244, + "grad_norm": 0.7139577795218617, + "learning_rate": 9.913212093804835e-06, + "loss": 0.4378546714782715, + "num_input_tokens_seen": 26673248, + "step": 240, + "train_runtime": 3316.079, + "train_tokens_per_second": 8043.611 + }, + { + "epoch": 0.12176938369781312, + "grad_norm": 0.6767799849577883, + "learning_rate": 9.909553918225186e-06, + "loss": 0.42829208374023436, + "num_input_tokens_seen": 27253464, + "step": 245, + "train_runtime": 3391.3447, + "train_tokens_per_second": 8036.182 + }, + { + "epoch": 0.1242544731610338, + "grad_norm": 0.6991762222617712, + "learning_rate": 9.905820931674404e-06, + "loss": 0.42859811782836915, + "num_input_tokens_seen": 27792304, + "step": 250, + "train_runtime": 3464.0294, + "train_tokens_per_second": 8023.114 + }, + { + "epoch": 0.12673956262425448, + "grad_norm": 0.7301292115528666, + "learning_rate": 9.902013191035118e-06, + "loss": 0.4286064147949219, + "num_input_tokens_seen": 28363528, + "step": 255, + "train_runtime": 3540.8198, + "train_tokens_per_second": 8010.441 + }, + { + "epoch": 0.12922465208747516, + "grad_norm": 0.7160088828333275, + "learning_rate": 9.898130754329053e-06, + "loss": 0.41687164306640623, + "num_input_tokens_seen": 28939472, + "step": 260, + "train_runtime": 3599.3546, + "train_tokens_per_second": 8040.184 + }, + { + "epoch": 0.1317097415506958, + "grad_norm": 0.6280424801999446, + "learning_rate": 9.894173680716138e-06, + "loss": 0.418760871887207, + "num_input_tokens_seen": 29510080, + "step": 265, + "train_runtime": 3655.6824, + "train_tokens_per_second": 8072.386 + }, + { + "epoch": 0.1341948310139165, + "grad_norm": 0.6423846031425479, + "learning_rate": 9.890142030493603e-06, + "loss": 0.41259174346923827, + "num_input_tokens_seen": 30092568, + "step": 270, + "train_runtime": 3716.5484, + "train_tokens_per_second": 8096.913 + }, + { + "epoch": 0.13667992047713717, + "grad_norm": 0.6778581431468657, + "learning_rate": 9.886035865095073e-06, + "loss": 0.45127053260803224, + "num_input_tokens_seen": 30677880, + "step": 275, + "train_runtime": 3787.1836, + "train_tokens_per_second": 8100.447 + }, + { + "epoch": 0.13916500994035785, + "grad_norm": 0.7286396272235184, + "learning_rate": 9.881855247089615e-06, + "loss": 0.4469005584716797, + "num_input_tokens_seen": 31242472, + "step": 280, + "train_runtime": 3853.396, + "train_tokens_per_second": 8107.776 + }, + { + "epoch": 0.14165009940357853, + "grad_norm": 0.7206352363765698, + "learning_rate": 9.877600240180793e-06, + "loss": 0.40082669258117676, + "num_input_tokens_seen": 31809864, + "step": 285, + "train_runtime": 3928.0213, + "train_tokens_per_second": 8098.19 + }, + { + "epoch": 0.1441351888667992, + "grad_norm": 0.662700111475498, + "learning_rate": 9.873270909205703e-06, + "loss": 0.40410122871398924, + "num_input_tokens_seen": 32351336, + "step": 290, + "train_runtime": 3990.9788, + "train_tokens_per_second": 8106.116 + }, + { + "epoch": 0.1466202783300199, + "grad_norm": 1.1819170959698453, + "learning_rate": 9.868867320133974e-06, + "loss": 0.4134364128112793, + "num_input_tokens_seen": 32862824, + "step": 295, + "train_runtime": 4061.991, + "train_tokens_per_second": 8090.324 + }, + { + "epoch": 0.14910536779324055, + "grad_norm": 0.6349050134498895, + "learning_rate": 9.864389540066765e-06, + "loss": 0.40877833366394045, + "num_input_tokens_seen": 33455224, + "step": 300, + "train_runtime": 4134.2336, + "train_tokens_per_second": 8092.243 + }, + { + "epoch": 0.15159045725646123, + "grad_norm": 0.7145363527076218, + "learning_rate": 9.859837637235752e-06, + "loss": 0.4419567108154297, + "num_input_tokens_seen": 34011888, + "step": 305, + "train_runtime": 4200.7209, + "train_tokens_per_second": 8096.679 + }, + { + "epoch": 0.1540755467196819, + "grad_norm": 0.6304983059809672, + "learning_rate": 9.855211681002077e-06, + "loss": 0.4136381149291992, + "num_input_tokens_seen": 34563864, + "step": 310, + "train_runtime": 4272.3884, + "train_tokens_per_second": 8090.057 + }, + { + "epoch": 0.15656063618290258, + "grad_norm": 0.7209343441189546, + "learning_rate": 9.850511741855298e-06, + "loss": 0.4224031925201416, + "num_input_tokens_seen": 35098496, + "step": 315, + "train_runtime": 4329.2394, + "train_tokens_per_second": 8107.312 + }, + { + "epoch": 0.15904572564612326, + "grad_norm": 0.7704307506047531, + "learning_rate": 9.845737891412308e-06, + "loss": 0.4627669334411621, + "num_input_tokens_seen": 35636160, + "step": 320, + "train_runtime": 4395.1235, + "train_tokens_per_second": 8108.113 + }, + { + "epoch": 0.16153081510934394, + "grad_norm": 0.6892599049173513, + "learning_rate": 9.840890202416253e-06, + "loss": 0.42969422340393065, + "num_input_tokens_seen": 36196184, + "step": 325, + "train_runtime": 4470.4024, + "train_tokens_per_second": 8096.851 + }, + { + "epoch": 0.16401590457256462, + "grad_norm": 0.7501653604520323, + "learning_rate": 9.835968748735415e-06, + "loss": 0.41391358375549314, + "num_input_tokens_seen": 36764680, + "step": 330, + "train_runtime": 4540.3634, + "train_tokens_per_second": 8097.299 + }, + { + "epoch": 0.16650099403578528, + "grad_norm": 0.705647029353333, + "learning_rate": 9.830973605362093e-06, + "loss": 0.42150249481201174, + "num_input_tokens_seen": 37355168, + "step": 335, + "train_runtime": 4605.5272, + "train_tokens_per_second": 8110.943 + }, + { + "epoch": 0.16898608349900596, + "grad_norm": 0.6797613209132866, + "learning_rate": 9.825904848411457e-06, + "loss": 0.42339191436767576, + "num_input_tokens_seen": 37898784, + "step": 340, + "train_runtime": 4669.2568, + "train_tokens_per_second": 8116.663 + }, + { + "epoch": 0.17147117296222664, + "grad_norm": 0.7340918300956556, + "learning_rate": 9.820762555120386e-06, + "loss": 0.42127156257629395, + "num_input_tokens_seen": 38481464, + "step": 345, + "train_runtime": 4745.8387, + "train_tokens_per_second": 8108.464 + }, + { + "epoch": 0.17395626242544732, + "grad_norm": 0.6804354135392355, + "learning_rate": 9.815546803846297e-06, + "loss": 0.4065572738647461, + "num_input_tokens_seen": 39060736, + "step": 350, + "train_runtime": 4814.6774, + "train_tokens_per_second": 8112.846 + }, + { + "epoch": 0.176441351888668, + "grad_norm": 0.7446666159487537, + "learning_rate": 9.810257674065942e-06, + "loss": 0.39780230522155763, + "num_input_tokens_seen": 39569784, + "step": 355, + "train_runtime": 4877.1833, + "train_tokens_per_second": 8113.245 + }, + { + "epoch": 0.17892644135188868, + "grad_norm": 0.6572454024397968, + "learning_rate": 9.804895246374211e-06, + "loss": 0.4173138618469238, + "num_input_tokens_seen": 40119136, + "step": 360, + "train_runtime": 4944.5934, + "train_tokens_per_second": 8113.738 + }, + { + "epoch": 0.18141153081510936, + "grad_norm": 0.6810638966440239, + "learning_rate": 9.799459602482888e-06, + "loss": 0.4325833797454834, + "num_input_tokens_seen": 40696928, + "step": 365, + "train_runtime": 5012.5173, + "train_tokens_per_second": 8119.06 + }, + { + "epoch": 0.18389662027833, + "grad_norm": 0.7460043268879804, + "learning_rate": 9.793950825219418e-06, + "loss": 0.40879230499267577, + "num_input_tokens_seen": 41261312, + "step": 370, + "train_runtime": 5088.2819, + "train_tokens_per_second": 8109.085 + }, + { + "epoch": 0.1863817097415507, + "grad_norm": 0.6543763054325196, + "learning_rate": 9.788368998525635e-06, + "loss": 0.40566296577453614, + "num_input_tokens_seen": 41817800, + "step": 375, + "train_runtime": 5150.8642, + "train_tokens_per_second": 8118.599 + }, + { + "epoch": 0.18886679920477137, + "grad_norm": 0.7897116477363397, + "learning_rate": 9.782714207456496e-06, + "loss": 0.4157840728759766, + "num_input_tokens_seen": 42346336, + "step": 380, + "train_runtime": 5218.3479, + "train_tokens_per_second": 8114.893 + }, + { + "epoch": 0.19135188866799205, + "grad_norm": 0.5801772953709613, + "learning_rate": 9.776986538178773e-06, + "loss": 0.4044309616088867, + "num_input_tokens_seen": 42916584, + "step": 385, + "train_runtime": 5289.3381, + "train_tokens_per_second": 8113.791 + }, + { + "epoch": 0.19383697813121273, + "grad_norm": 0.6644503896895383, + "learning_rate": 9.771186077969736e-06, + "loss": 0.4201981544494629, + "num_input_tokens_seen": 43524552, + "step": 390, + "train_runtime": 5360.5957, + "train_tokens_per_second": 8119.35 + }, + { + "epoch": 0.1963220675944334, + "grad_norm": 0.6594682011653696, + "learning_rate": 9.76531291521585e-06, + "loss": 0.3854872941970825, + "num_input_tokens_seen": 44099480, + "step": 395, + "train_runtime": 5425.3275, + "train_tokens_per_second": 8128.446 + }, + { + "epoch": 0.1988071570576541, + "grad_norm": 0.6522720740101062, + "learning_rate": 9.759367139411389e-06, + "loss": 0.4173161506652832, + "num_input_tokens_seen": 44659192, + "step": 400, + "train_runtime": 5479.4308, + "train_tokens_per_second": 8150.334 + }, + { + "epoch": 0.20129224652087474, + "grad_norm": 0.6506002420636534, + "learning_rate": 9.753348841157107e-06, + "loss": 0.40453433990478516, + "num_input_tokens_seen": 45225784, + "step": 405, + "train_runtime": 5558.2618, + "train_tokens_per_second": 8136.678 + }, + { + "epoch": 0.20377733598409542, + "grad_norm": 0.7682310982532393, + "learning_rate": 9.747258112158837e-06, + "loss": 0.4161695957183838, + "num_input_tokens_seen": 45769824, + "step": 410, + "train_runtime": 5627.989, + "train_tokens_per_second": 8132.536 + }, + { + "epoch": 0.2062624254473161, + "grad_norm": 0.6993601555199566, + "learning_rate": 9.741095045226098e-06, + "loss": 0.42332983016967773, + "num_input_tokens_seen": 46364216, + "step": 415, + "train_runtime": 5704.4159, + "train_tokens_per_second": 8127.776 + }, + { + "epoch": 0.20874751491053678, + "grad_norm": 0.8818710434381768, + "learning_rate": 9.734859734270692e-06, + "loss": 0.4054297924041748, + "num_input_tokens_seen": 46930928, + "step": 420, + "train_runtime": 5762.8675, + "train_tokens_per_second": 8143.676 + }, + { + "epoch": 0.21123260437375746, + "grad_norm": 0.6952834831712054, + "learning_rate": 9.72855227430525e-06, + "loss": 0.4093897819519043, + "num_input_tokens_seen": 47488416, + "step": 425, + "train_runtime": 5829.918, + "train_tokens_per_second": 8145.641 + }, + { + "epoch": 0.21371769383697814, + "grad_norm": 0.7035993394357021, + "learning_rate": 9.722172761441806e-06, + "loss": 0.40102171897888184, + "num_input_tokens_seen": 48082952, + "step": 430, + "train_runtime": 5898.1334, + "train_tokens_per_second": 8152.232 + }, + { + "epoch": 0.21620278330019882, + "grad_norm": 0.6911870115106311, + "learning_rate": 9.715721292890325e-06, + "loss": 0.40170631408691404, + "num_input_tokens_seen": 48633736, + "step": 435, + "train_runtime": 5971.3436, + "train_tokens_per_second": 8144.522 + }, + { + "epoch": 0.21868787276341947, + "grad_norm": 0.6573411917108836, + "learning_rate": 9.709197966957217e-06, + "loss": 0.4013651371002197, + "num_input_tokens_seen": 49197560, + "step": 440, + "train_runtime": 6041.5015, + "train_tokens_per_second": 8143.267 + }, + { + "epoch": 0.22117296222664015, + "grad_norm": 0.7422260563403945, + "learning_rate": 9.702602883043845e-06, + "loss": 0.4123281478881836, + "num_input_tokens_seen": 49775504, + "step": 445, + "train_runtime": 6108.1405, + "train_tokens_per_second": 8149.044 + }, + { + "epoch": 0.22365805168986083, + "grad_norm": 0.8371929406965495, + "learning_rate": 9.695936141645007e-06, + "loss": 0.39359707832336427, + "num_input_tokens_seen": 50348040, + "step": 450, + "train_runtime": 6173.433, + "train_tokens_per_second": 8155.598 + }, + { + "epoch": 0.2261431411530815, + "grad_norm": 0.7514594863149813, + "learning_rate": 9.689197844347404e-06, + "loss": 0.42499399185180664, + "num_input_tokens_seen": 50859312, + "step": 455, + "train_runtime": 6239.1577, + "train_tokens_per_second": 8151.631 + }, + { + "epoch": 0.2286282306163022, + "grad_norm": 0.7238788450094297, + "learning_rate": 9.6823880938281e-06, + "loss": 0.403135871887207, + "num_input_tokens_seen": 51420904, + "step": 460, + "train_runtime": 6314.7403, + "train_tokens_per_second": 8142.996 + }, + { + "epoch": 0.23111332007952287, + "grad_norm": 0.645737273625517, + "learning_rate": 9.67550699385295e-06, + "loss": 0.3934160709381104, + "num_input_tokens_seen": 52002024, + "step": 465, + "train_runtime": 6379.9154, + "train_tokens_per_second": 8150.896 + }, + { + "epoch": 0.23359840954274355, + "grad_norm": 0.7494232376603538, + "learning_rate": 9.668554649275011e-06, + "loss": 0.41286430358886717, + "num_input_tokens_seen": 52574656, + "step": 470, + "train_runtime": 6454.7081, + "train_tokens_per_second": 8145.164 + }, + { + "epoch": 0.2360834990059642, + "grad_norm": 0.7472569718523798, + "learning_rate": 9.661531166032968e-06, + "loss": 0.3993457078933716, + "num_input_tokens_seen": 53132184, + "step": 475, + "train_runtime": 6523.9363, + "train_tokens_per_second": 8144.191 + }, + { + "epoch": 0.23856858846918488, + "grad_norm": 0.6877935355865147, + "learning_rate": 9.65443665114949e-06, + "loss": 0.3828821897506714, + "num_input_tokens_seen": 53708416, + "step": 480, + "train_runtime": 6593.8237, + "train_tokens_per_second": 8145.261 + }, + { + "epoch": 0.24105367793240556, + "grad_norm": 0.6352472339861172, + "learning_rate": 9.64727121272963e-06, + "loss": 0.3770132541656494, + "num_input_tokens_seen": 54288656, + "step": 485, + "train_runtime": 6664.6736, + "train_tokens_per_second": 8145.734 + }, + { + "epoch": 0.24353876739562624, + "grad_norm": 0.653028852664786, + "learning_rate": 9.64003495995915e-06, + "loss": 0.413008451461792, + "num_input_tokens_seen": 54869496, + "step": 490, + "train_runtime": 6735.0725, + "train_tokens_per_second": 8146.831 + }, + { + "epoch": 0.24602385685884692, + "grad_norm": 0.6121568523371016, + "learning_rate": 9.632728003102873e-06, + "loss": 0.40009346008300783, + "num_input_tokens_seen": 55427408, + "step": 495, + "train_runtime": 6799.8959, + "train_tokens_per_second": 8151.214 + }, + { + "epoch": 0.2485089463220676, + "grad_norm": 0.7120886817536121, + "learning_rate": 9.625350453503001e-06, + "loss": 0.4180534839630127, + "num_input_tokens_seen": 55973256, + "step": 500, + "train_runtime": 6868.4697, + "train_tokens_per_second": 8149.305 + }, + { + "epoch": 0.2509940357852883, + "grad_norm": 0.6788366210649915, + "learning_rate": 9.617902423577418e-06, + "loss": 0.38946239948272704, + "num_input_tokens_seen": 56502248, + "step": 505, + "train_runtime": 6945.3063, + "train_tokens_per_second": 8135.314 + }, + { + "epoch": 0.25347912524850896, + "grad_norm": 0.6723523622436895, + "learning_rate": 9.61038402681797e-06, + "loss": 0.38671092987060546, + "num_input_tokens_seen": 57099144, + "step": 510, + "train_runtime": 7006.3056, + "train_tokens_per_second": 8149.679 + }, + { + "epoch": 0.25596421471172964, + "grad_norm": 0.6883074091669321, + "learning_rate": 9.602795377788745e-06, + "loss": 0.4063846588134766, + "num_input_tokens_seen": 57631048, + "step": 515, + "train_runtime": 7068.9472, + "train_tokens_per_second": 8152.706 + }, + { + "epoch": 0.2584493041749503, + "grad_norm": 0.6993757379347519, + "learning_rate": 9.59513659212432e-06, + "loss": 0.3968632221221924, + "num_input_tokens_seen": 58161800, + "step": 520, + "train_runtime": 7140.0651, + "train_tokens_per_second": 8145.836 + }, + { + "epoch": 0.26093439363817095, + "grad_norm": 0.7105408103332554, + "learning_rate": 9.587407786528012e-06, + "loss": 0.4193892955780029, + "num_input_tokens_seen": 58751864, + "step": 525, + "train_runtime": 7214.3581, + "train_tokens_per_second": 8143.741 + }, + { + "epoch": 0.2634194831013916, + "grad_norm": 0.6615084684963548, + "learning_rate": 9.579609078770076e-06, + "loss": 0.38826470375061034, + "num_input_tokens_seen": 59341128, + "step": 530, + "train_runtime": 7291.9092, + "train_tokens_per_second": 8137.941 + }, + { + "epoch": 0.2659045725646123, + "grad_norm": 0.715899891970891, + "learning_rate": 9.571740587685934e-06, + "loss": 0.40195498466491697, + "num_input_tokens_seen": 59904784, + "step": 535, + "train_runtime": 7366.4075, + "train_tokens_per_second": 8132.157 + }, + { + "epoch": 0.268389662027833, + "grad_norm": 0.6914023809775961, + "learning_rate": 9.563802433174354e-06, + "loss": 0.3986898183822632, + "num_input_tokens_seen": 60457240, + "step": 540, + "train_runtime": 7431.4252, + "train_tokens_per_second": 8135.349 + }, + { + "epoch": 0.27087475149105367, + "grad_norm": 0.6469623320084871, + "learning_rate": 9.555794736195621e-06, + "loss": 0.41936459541320803, + "num_input_tokens_seen": 61018792, + "step": 545, + "train_runtime": 7504.4225, + "train_tokens_per_second": 8131.044 + }, + { + "epoch": 0.27335984095427435, + "grad_norm": 0.6316071072074376, + "learning_rate": 9.547717618769699e-06, + "loss": 0.3822399377822876, + "num_input_tokens_seen": 61587280, + "step": 550, + "train_runtime": 7575.5242, + "train_tokens_per_second": 8129.771 + }, + { + "epoch": 0.275844930417495, + "grad_norm": 0.701021051625865, + "learning_rate": 9.53957120397437e-06, + "loss": 0.40807833671569826, + "num_input_tokens_seen": 62172968, + "step": 555, + "train_runtime": 7652.0272, + "train_tokens_per_second": 8125.032 + }, + { + "epoch": 0.2783300198807157, + "grad_norm": 0.6331565665780977, + "learning_rate": 9.53135561594335e-06, + "loss": 0.3991607427597046, + "num_input_tokens_seen": 62738296, + "step": 560, + "train_runtime": 7720.4402, + "train_tokens_per_second": 8126.259 + }, + { + "epoch": 0.2808151093439364, + "grad_norm": 0.7111143326581448, + "learning_rate": 9.523070979864419e-06, + "loss": 0.3836745500564575, + "num_input_tokens_seen": 63287696, + "step": 565, + "train_runtime": 7785.6342, + "train_tokens_per_second": 8128.778 + }, + { + "epoch": 0.28330019880715707, + "grad_norm": 0.6857878025122368, + "learning_rate": 9.514717421977488e-06, + "loss": 0.40068907737731935, + "num_input_tokens_seen": 63803960, + "step": 570, + "train_runtime": 7855.5786, + "train_tokens_per_second": 8122.121 + }, + { + "epoch": 0.28578528827037775, + "grad_norm": 0.679734463461153, + "learning_rate": 9.506295069572691e-06, + "loss": 0.39016916751861574, + "num_input_tokens_seen": 64387544, + "step": 575, + "train_runtime": 7933.8861, + "train_tokens_per_second": 8115.512 + }, + { + "epoch": 0.2882703777335984, + "grad_norm": 0.6438999905737711, + "learning_rate": 9.497804050988444e-06, + "loss": 0.3936117887496948, + "num_input_tokens_seen": 64959824, + "step": 580, + "train_runtime": 8003.8231, + "train_tokens_per_second": 8116.099 + }, + { + "epoch": 0.2907554671968191, + "grad_norm": 0.7738550684817421, + "learning_rate": 9.489244495609483e-06, + "loss": 0.3829601287841797, + "num_input_tokens_seen": 65505672, + "step": 585, + "train_runtime": 8068.5982, + "train_tokens_per_second": 8118.594 + }, + { + "epoch": 0.2932405566600398, + "grad_norm": 0.5964964558199108, + "learning_rate": 9.4806165338649e-06, + "loss": 0.3989708423614502, + "num_input_tokens_seen": 66085672, + "step": 590, + "train_runtime": 8141.3491, + "train_tokens_per_second": 8117.288 + }, + { + "epoch": 0.2957256461232604, + "grad_norm": 0.6987987841896753, + "learning_rate": 9.47192029722615e-06, + "loss": 0.4029735565185547, + "num_input_tokens_seen": 66666048, + "step": 595, + "train_runtime": 8206.1634, + "train_tokens_per_second": 8123.9 + }, + { + "epoch": 0.2982107355864811, + "grad_norm": 0.6866832077620406, + "learning_rate": 9.463155918205044e-06, + "loss": 0.39454197883605957, + "num_input_tokens_seen": 67219704, + "step": 600, + "train_runtime": 8272.4325, + "train_tokens_per_second": 8125.748 + }, + { + "epoch": 0.30069582504970177, + "grad_norm": 0.6467053653124836, + "learning_rate": 9.45432353035175e-06, + "loss": 0.3835888385772705, + "num_input_tokens_seen": 67804880, + "step": 605, + "train_runtime": 8345.7372, + "train_tokens_per_second": 8124.493 + }, + { + "epoch": 0.30318091451292245, + "grad_norm": 0.8648961152385213, + "learning_rate": 9.445423268252728e-06, + "loss": 0.38858232498168943, + "num_input_tokens_seen": 68411640, + "step": 610, + "train_runtime": 8415.2451, + "train_tokens_per_second": 8129.489 + }, + { + "epoch": 0.30566600397614313, + "grad_norm": 0.6604286061422132, + "learning_rate": 9.436455267528703e-06, + "loss": 0.39548397064208984, + "num_input_tokens_seen": 68947928, + "step": 615, + "train_runtime": 8475.3192, + "train_tokens_per_second": 8135.142 + }, + { + "epoch": 0.3081510934393638, + "grad_norm": 0.6829665779642397, + "learning_rate": 9.427419664832587e-06, + "loss": 0.4033242702484131, + "num_input_tokens_seen": 69509952, + "step": 620, + "train_runtime": 8544.5865, + "train_tokens_per_second": 8134.97 + }, + { + "epoch": 0.3106361829025845, + "grad_norm": 0.7812393604791641, + "learning_rate": 9.418316597847399e-06, + "loss": 0.4224880218505859, + "num_input_tokens_seen": 70059912, + "step": 625, + "train_runtime": 8613.9454, + "train_tokens_per_second": 8133.313 + }, + { + "epoch": 0.31312127236580517, + "grad_norm": 0.6333442228925582, + "learning_rate": 9.409146205284173e-06, + "loss": 0.40122294425964355, + "num_input_tokens_seen": 70619256, + "step": 630, + "train_runtime": 8678.5839, + "train_tokens_per_second": 8137.187 + }, + { + "epoch": 0.31560636182902585, + "grad_norm": 0.7508822861442499, + "learning_rate": 9.399908626879833e-06, + "loss": 0.38468031883239745, + "num_input_tokens_seen": 71129552, + "step": 635, + "train_runtime": 8750.5035, + "train_tokens_per_second": 8128.624 + }, + { + "epoch": 0.31809145129224653, + "grad_norm": 0.6485400299103095, + "learning_rate": 9.39060400339507e-06, + "loss": 0.3871898174285889, + "num_input_tokens_seen": 71696440, + "step": 640, + "train_runtime": 8812.5572, + "train_tokens_per_second": 8135.713 + }, + { + "epoch": 0.3205765407554672, + "grad_norm": 0.6150207837439486, + "learning_rate": 9.3812324766122e-06, + "loss": 0.40517573356628417, + "num_input_tokens_seen": 72283752, + "step": 645, + "train_runtime": 8885.3822, + "train_tokens_per_second": 8135.131 + }, + { + "epoch": 0.3230616302186879, + "grad_norm": 0.6361063463586124, + "learning_rate": 9.371794189333001e-06, + "loss": 0.39126243591308596, + "num_input_tokens_seen": 72878696, + "step": 650, + "train_runtime": 8949.227, + "train_tokens_per_second": 8143.574 + }, + { + "epoch": 0.32554671968190857, + "grad_norm": 0.7063808722971539, + "learning_rate": 9.362289285376535e-06, + "loss": 0.40833339691162107, + "num_input_tokens_seen": 73403960, + "step": 655, + "train_runtime": 9023.6416, + "train_tokens_per_second": 8134.627 + }, + { + "epoch": 0.32803180914512925, + "grad_norm": 0.6549033895959345, + "learning_rate": 9.352717909576953e-06, + "loss": 0.38887796401977537, + "num_input_tokens_seen": 73968104, + "step": 660, + "train_runtime": 9098.1734, + "train_tokens_per_second": 8129.995 + }, + { + "epoch": 0.3305168986083499, + "grad_norm": 0.7050082995725995, + "learning_rate": 9.343080207781303e-06, + "loss": 0.39077062606811525, + "num_input_tokens_seen": 74529736, + "step": 665, + "train_runtime": 9166.8651, + "train_tokens_per_second": 8130.341 + }, + { + "epoch": 0.33300198807157055, + "grad_norm": 0.7206810212457836, + "learning_rate": 9.333376326847287e-06, + "loss": 0.40349912643432617, + "num_input_tokens_seen": 75099920, + "step": 670, + "train_runtime": 9231.9047, + "train_tokens_per_second": 8134.824 + }, + { + "epoch": 0.33548707753479123, + "grad_norm": 1.2766859850539753, + "learning_rate": 9.323606414641042e-06, + "loss": 0.4237051963806152, + "num_input_tokens_seen": 75617560, + "step": 675, + "train_runtime": 9299.6805, + "train_tokens_per_second": 8131.2 + }, + { + "epoch": 0.3379721669980119, + "grad_norm": 0.6546120612034582, + "learning_rate": 9.313770620034875e-06, + "loss": 0.39501242637634276, + "num_input_tokens_seen": 76193520, + "step": 680, + "train_runtime": 9372.6197, + "train_tokens_per_second": 8129.373 + }, + { + "epoch": 0.3404572564612326, + "grad_norm": 0.7077091286308951, + "learning_rate": 9.303869092904997e-06, + "loss": 0.38451781272888186, + "num_input_tokens_seen": 76725112, + "step": 685, + "train_runtime": 9431.7736, + "train_tokens_per_second": 8134.749 + }, + { + "epoch": 0.3429423459244533, + "grad_norm": 0.847094582235997, + "learning_rate": 9.293901984129242e-06, + "loss": 0.42759084701538086, + "num_input_tokens_seen": 77253160, + "step": 690, + "train_runtime": 9507.3404, + "train_tokens_per_second": 8125.633 + }, + { + "epoch": 0.34542743538767395, + "grad_norm": 0.724643275692759, + "learning_rate": 9.283869445584768e-06, + "loss": 0.38852245807647706, + "num_input_tokens_seen": 77834712, + "step": 695, + "train_runtime": 9580.8788, + "train_tokens_per_second": 8123.964 + }, + { + "epoch": 0.34791252485089463, + "grad_norm": 0.6388676139951753, + "learning_rate": 9.273771630145738e-06, + "loss": 0.3962401866912842, + "num_input_tokens_seen": 78415136, + "step": 700, + "train_runtime": 9657.2939, + "train_tokens_per_second": 8119.784 + }, + { + "epoch": 0.3503976143141153, + "grad_norm": 0.7023099375951507, + "learning_rate": 9.263608691680995e-06, + "loss": 0.3903517723083496, + "num_input_tokens_seen": 78979280, + "step": 705, + "train_runtime": 9720.0881, + "train_tokens_per_second": 8125.367 + }, + { + "epoch": 0.352882703777336, + "grad_norm": 0.6944448445664358, + "learning_rate": 9.253380785051713e-06, + "loss": 0.40244059562683104, + "num_input_tokens_seen": 79592584, + "step": 710, + "train_runtime": 9796.9348, + "train_tokens_per_second": 8124.233 + }, + { + "epoch": 0.3553677932405567, + "grad_norm": 0.7101208454465424, + "learning_rate": 9.243088066109046e-06, + "loss": 0.3918424367904663, + "num_input_tokens_seen": 80133568, + "step": 715, + "train_runtime": 9862.7867, + "train_tokens_per_second": 8124.84 + }, + { + "epoch": 0.35785288270377735, + "grad_norm": 0.7046102017305603, + "learning_rate": 9.232730691691744e-06, + "loss": 0.3876338481903076, + "num_input_tokens_seen": 80686376, + "step": 720, + "train_runtime": 9934.4835, + "train_tokens_per_second": 8121.849 + }, + { + "epoch": 0.36033797216699803, + "grad_norm": 0.5849718476461411, + "learning_rate": 9.222308819623764e-06, + "loss": 0.38863554000854494, + "num_input_tokens_seen": 81262208, + "step": 725, + "train_runtime": 10004.5111, + "train_tokens_per_second": 8122.557 + }, + { + "epoch": 0.3628230616302187, + "grad_norm": 0.6364454170804459, + "learning_rate": 9.211822608711871e-06, + "loss": 0.3743600368499756, + "num_input_tokens_seen": 81836168, + "step": 730, + "train_runtime": 10070.199, + "train_tokens_per_second": 8126.569 + }, + { + "epoch": 0.36530815109343934, + "grad_norm": 0.6071928222164013, + "learning_rate": 9.201272218743216e-06, + "loss": 0.3995983600616455, + "num_input_tokens_seen": 82371112, + "step": 735, + "train_runtime": 10135.7095, + "train_tokens_per_second": 8126.822 + }, + { + "epoch": 0.36779324055666, + "grad_norm": 0.7349936527629004, + "learning_rate": 9.190657810482895e-06, + "loss": 0.3718123912811279, + "num_input_tokens_seen": 82944752, + "step": 740, + "train_runtime": 10207.6767, + "train_tokens_per_second": 8125.723 + }, + { + "epoch": 0.3702783300198807, + "grad_norm": 0.5993649435181128, + "learning_rate": 9.179979545671506e-06, + "loss": 0.39644222259521483, + "num_input_tokens_seen": 83524856, + "step": 745, + "train_runtime": 10286.3081, + "train_tokens_per_second": 8120.003 + }, + { + "epoch": 0.3727634194831014, + "grad_norm": 0.7228648169458408, + "learning_rate": 9.169237587022686e-06, + "loss": 0.3852032899856567, + "num_input_tokens_seen": 84076376, + "step": 750, + "train_runtime": 10358.1994, + "train_tokens_per_second": 8116.891 + }, + { + "epoch": 0.37524850894632206, + "grad_norm": 0.6375390856299219, + "learning_rate": 9.158432098220622e-06, + "loss": 0.39706830978393554, + "num_input_tokens_seen": 84659784, + "step": 755, + "train_runtime": 10435.2201, + "train_tokens_per_second": 8112.889 + }, + { + "epoch": 0.37773359840954274, + "grad_norm": 0.6420632496748842, + "learning_rate": 9.14756324391757e-06, + "loss": 0.40331830978393557, + "num_input_tokens_seen": 85229696, + "step": 760, + "train_runtime": 10502.6016, + "train_tokens_per_second": 8115.103 + }, + { + "epoch": 0.3802186878727634, + "grad_norm": 0.6732092960025068, + "learning_rate": 9.136631189731333e-06, + "loss": 0.38048765659332273, + "num_input_tokens_seen": 85791504, + "step": 765, + "train_runtime": 10577.5089, + "train_tokens_per_second": 8110.748 + }, + { + "epoch": 0.3827037773359841, + "grad_norm": 0.7172930233435729, + "learning_rate": 9.125636102242746e-06, + "loss": 0.4042394161224365, + "num_input_tokens_seen": 86339056, + "step": 770, + "train_runtime": 10644.362, + "train_tokens_per_second": 8111.248 + }, + { + "epoch": 0.3851888667992048, + "grad_norm": 0.6733955622585514, + "learning_rate": 9.114578148993137e-06, + "loss": 0.40470151901245116, + "num_input_tokens_seen": 86915040, + "step": 775, + "train_runtime": 10708.2311, + "train_tokens_per_second": 8116.657 + }, + { + "epoch": 0.38767395626242546, + "grad_norm": 0.6506079124141272, + "learning_rate": 9.103457498481767e-06, + "loss": 0.4003398895263672, + "num_input_tokens_seen": 87497840, + "step": 780, + "train_runtime": 10774.5507, + "train_tokens_per_second": 8120.788 + }, + { + "epoch": 0.39015904572564614, + "grad_norm": 0.6197586716604738, + "learning_rate": 9.092274320163272e-06, + "loss": 0.39843001365661623, + "num_input_tokens_seen": 88080960, + "step": 785, + "train_runtime": 10854.3091, + "train_tokens_per_second": 8114.838 + }, + { + "epoch": 0.3926441351888668, + "grad_norm": 0.6214330713076378, + "learning_rate": 9.081028784445074e-06, + "loss": 0.41014394760131834, + "num_input_tokens_seen": 88642200, + "step": 790, + "train_runtime": 10922.0457, + "train_tokens_per_second": 8115.897 + }, + { + "epoch": 0.3951292246520875, + "grad_norm": 0.637025404217726, + "learning_rate": 9.069721062684788e-06, + "loss": 0.39574294090270995, + "num_input_tokens_seen": 89186160, + "step": 795, + "train_runtime": 10996.8642, + "train_tokens_per_second": 8110.145 + }, + { + "epoch": 0.3976143141153082, + "grad_norm": 0.7309686126550612, + "learning_rate": 9.058351327187612e-06, + "loss": 0.4020538330078125, + "num_input_tokens_seen": 89741648, + "step": 800, + "train_runtime": 11065.0273, + "train_tokens_per_second": 8110.386 + }, + { + "epoch": 0.4000994035785288, + "grad_norm": 0.6486776530798172, + "learning_rate": 9.04691975120369e-06, + "loss": 0.38901195526123045, + "num_input_tokens_seen": 90294104, + "step": 805, + "train_runtime": 11138.3975, + "train_tokens_per_second": 8106.561 + }, + { + "epoch": 0.4025844930417495, + "grad_norm": 0.6250833402939229, + "learning_rate": 9.035426508925492e-06, + "loss": 0.41133790016174315, + "num_input_tokens_seen": 90861112, + "step": 810, + "train_runtime": 11201.0576, + "train_tokens_per_second": 8111.833 + }, + { + "epoch": 0.40506958250497016, + "grad_norm": 0.6336674978219595, + "learning_rate": 9.023871775485144e-06, + "loss": 0.3821245670318604, + "num_input_tokens_seen": 91462096, + "step": 815, + "train_runtime": 11277.1005, + "train_tokens_per_second": 8110.427 + }, + { + "epoch": 0.40755467196819084, + "grad_norm": 0.6908474991384238, + "learning_rate": 9.012255726951762e-06, + "loss": 0.37911319732666016, + "num_input_tokens_seen": 91998152, + "step": 820, + "train_runtime": 11345.4991, + "train_tokens_per_second": 8108.78 + }, + { + "epoch": 0.4100397614314115, + "grad_norm": 0.6567025243966448, + "learning_rate": 9.000578540328774e-06, + "loss": 0.3868110179901123, + "num_input_tokens_seen": 92551272, + "step": 825, + "train_runtime": 11397.0237, + "train_tokens_per_second": 8120.653 + }, + { + "epoch": 0.4125248508946322, + "grad_norm": 0.7276807857998491, + "learning_rate": 8.988840393551223e-06, + "loss": 0.4077720642089844, + "num_input_tokens_seen": 93098784, + "step": 830, + "train_runtime": 11472.7767, + "train_tokens_per_second": 8114.756 + }, + { + "epoch": 0.4150099403578529, + "grad_norm": 0.6040724723816785, + "learning_rate": 8.977041465483047e-06, + "loss": 0.3834868907928467, + "num_input_tokens_seen": 93624944, + "step": 835, + "train_runtime": 11539.9357, + "train_tokens_per_second": 8113.125 + }, + { + "epoch": 0.41749502982107356, + "grad_norm": 0.6567273831952852, + "learning_rate": 8.965181935914362e-06, + "loss": 0.39317729473114016, + "num_input_tokens_seen": 94168944, + "step": 840, + "train_runtime": 11610.9127, + "train_tokens_per_second": 8110.383 + }, + { + "epoch": 0.41998011928429424, + "grad_norm": 0.6221685627736298, + "learning_rate": 8.95326198555872e-06, + "loss": 0.373657751083374, + "num_input_tokens_seen": 94717336, + "step": 845, + "train_runtime": 11674.633, + "train_tokens_per_second": 8113.089 + }, + { + "epoch": 0.4224652087475149, + "grad_norm": 0.7165153564917169, + "learning_rate": 8.941281796050353e-06, + "loss": 0.3971364498138428, + "num_input_tokens_seen": 95241328, + "step": 850, + "train_runtime": 11744.4826, + "train_tokens_per_second": 8109.453 + }, + { + "epoch": 0.4249502982107356, + "grad_norm": 0.691704234623846, + "learning_rate": 8.92924154994141e-06, + "loss": 0.3897407054901123, + "num_input_tokens_seen": 95766208, + "step": 855, + "train_runtime": 11813.3453, + "train_tokens_per_second": 8106.612 + }, + { + "epoch": 0.4274353876739563, + "grad_norm": 0.6885009290097122, + "learning_rate": 8.917141430699168e-06, + "loss": 0.3914926528930664, + "num_input_tokens_seen": 96331064, + "step": 860, + "train_runtime": 11886.9855, + "train_tokens_per_second": 8103.91 + }, + { + "epoch": 0.42992047713717696, + "grad_norm": 0.6685732541434032, + "learning_rate": 8.904981622703249e-06, + "loss": 0.41732587814331057, + "num_input_tokens_seen": 96907896, + "step": 865, + "train_runtime": 11959.3909, + "train_tokens_per_second": 8103.08 + }, + { + "epoch": 0.43240556660039764, + "grad_norm": 0.5918522728076266, + "learning_rate": 8.892762311242794e-06, + "loss": 0.4009408473968506, + "num_input_tokens_seen": 97474824, + "step": 870, + "train_runtime": 12030.6848, + "train_tokens_per_second": 8102.184 + }, + { + "epoch": 0.43489065606361826, + "grad_norm": 0.8623254383775136, + "learning_rate": 8.88048368251365e-06, + "loss": 0.38512749671936036, + "num_input_tokens_seen": 98021000, + "step": 875, + "train_runtime": 12100.1611, + "train_tokens_per_second": 8100.801 + }, + { + "epoch": 0.43737574552683894, + "grad_norm": 0.6698686591004054, + "learning_rate": 8.868145923615536e-06, + "loss": 0.4004067897796631, + "num_input_tokens_seen": 98606472, + "step": 880, + "train_runtime": 12178.5079, + "train_tokens_per_second": 8096.761 + }, + { + "epoch": 0.4398608349900596, + "grad_norm": 0.6474326780847197, + "learning_rate": 8.855749222549182e-06, + "loss": 0.3953530311584473, + "num_input_tokens_seen": 99163032, + "step": 885, + "train_runtime": 12244.6471, + "train_tokens_per_second": 8098.48 + }, + { + "epoch": 0.4423459244532803, + "grad_norm": 0.6224874146345041, + "learning_rate": 8.84329376821347e-06, + "loss": 0.39556407928466797, + "num_input_tokens_seen": 99734536, + "step": 890, + "train_runtime": 12320.0818, + "train_tokens_per_second": 8095.282 + }, + { + "epoch": 0.444831013916501, + "grad_norm": 1.167987554467554, + "learning_rate": 8.830779750402552e-06, + "loss": 0.38228645324707033, + "num_input_tokens_seen": 100336096, + "step": 895, + "train_runtime": 12394.7286, + "train_tokens_per_second": 8095.062 + }, + { + "epoch": 0.44731610337972166, + "grad_norm": 0.6553225039451944, + "learning_rate": 8.818207359802964e-06, + "loss": 0.37019195556640627, + "num_input_tokens_seen": 100900888, + "step": 900, + "train_runtime": 12473.426, + "train_tokens_per_second": 8089.268 + }, + { + "epoch": 0.44980119284294234, + "grad_norm": 0.6618406196252455, + "learning_rate": 8.805576787990717e-06, + "loss": 0.4208672523498535, + "num_input_tokens_seen": 101473232, + "step": 905, + "train_runtime": 12535.3608, + "train_tokens_per_second": 8094.959 + }, + { + "epoch": 0.452286282306163, + "grad_norm": 0.6588885148577126, + "learning_rate": 8.792888227428377e-06, + "loss": 0.3735737562179565, + "num_input_tokens_seen": 102025888, + "step": 910, + "train_runtime": 12608.3576, + "train_tokens_per_second": 8091.925 + }, + { + "epoch": 0.4547713717693837, + "grad_norm": 0.692978307316363, + "learning_rate": 8.780141871462131e-06, + "loss": 0.40465331077575684, + "num_input_tokens_seen": 102578928, + "step": 915, + "train_runtime": 12678.4908, + "train_tokens_per_second": 8090.784 + }, + { + "epoch": 0.4572564612326044, + "grad_norm": 0.6525049342881567, + "learning_rate": 8.767337914318847e-06, + "loss": 0.3807732820510864, + "num_input_tokens_seen": 103154336, + "step": 920, + "train_runtime": 12742.4883, + "train_tokens_per_second": 8095.306 + }, + { + "epoch": 0.45974155069582506, + "grad_norm": 0.7309713627006861, + "learning_rate": 8.754476551103103e-06, + "loss": 0.393044638633728, + "num_input_tokens_seen": 103696528, + "step": 925, + "train_runtime": 12820.0603, + "train_tokens_per_second": 8088.615 + }, + { + "epoch": 0.46222664015904574, + "grad_norm": 0.6611357704173997, + "learning_rate": 8.741557977794231e-06, + "loss": 0.37778701782226565, + "num_input_tokens_seen": 104259320, + "step": 930, + "train_runtime": 12888.7384, + "train_tokens_per_second": 8089.18 + }, + { + "epoch": 0.4647117296222664, + "grad_norm": 0.6691733672826592, + "learning_rate": 8.728582391243314e-06, + "loss": 0.3868418216705322, + "num_input_tokens_seen": 104836024, + "step": 935, + "train_runtime": 12955.4548, + "train_tokens_per_second": 8092.037 + }, + { + "epoch": 0.4671968190854871, + "grad_norm": 0.610418650928282, + "learning_rate": 8.715549989170194e-06, + "loss": 0.3844210147857666, + "num_input_tokens_seen": 105379672, + "step": 940, + "train_runtime": 13029.738, + "train_tokens_per_second": 8087.628 + }, + { + "epoch": 0.4696819085487077, + "grad_norm": 0.6617871653045767, + "learning_rate": 8.702460970160461e-06, + "loss": 0.3901623249053955, + "num_input_tokens_seen": 105977136, + "step": 945, + "train_runtime": 13102.4668, + "train_tokens_per_second": 8088.335 + }, + { + "epoch": 0.4721669980119284, + "grad_norm": 0.6659417096656759, + "learning_rate": 8.689315533662422e-06, + "loss": 0.3892323970794678, + "num_input_tokens_seen": 106512832, + "step": 950, + "train_runtime": 13175.831, + "train_tokens_per_second": 8083.956 + }, + { + "epoch": 0.4746520874751491, + "grad_norm": 0.6515881473306753, + "learning_rate": 8.676113879984066e-06, + "loss": 0.3954622745513916, + "num_input_tokens_seen": 107088296, + "step": 955, + "train_runtime": 13245.5794, + "train_tokens_per_second": 8084.833 + }, + { + "epoch": 0.47713717693836977, + "grad_norm": 0.683441432314069, + "learning_rate": 8.662856210290011e-06, + "loss": 0.38210899829864503, + "num_input_tokens_seen": 107613528, + "step": 960, + "train_runtime": 13313.226, + "train_tokens_per_second": 8083.204 + }, + { + "epoch": 0.47962226640159045, + "grad_norm": 0.6453497287907337, + "learning_rate": 8.649542726598437e-06, + "loss": 0.40354013442993164, + "num_input_tokens_seen": 108180144, + "step": 965, + "train_runtime": 13385.4655, + "train_tokens_per_second": 8081.911 + }, + { + "epoch": 0.4821073558648111, + "grad_norm": 4.06598238495918, + "learning_rate": 8.636173631778004e-06, + "loss": 0.3946369171142578, + "num_input_tokens_seen": 108762072, + "step": 970, + "train_runtime": 13454.2167, + "train_tokens_per_second": 8083.865 + }, + { + "epoch": 0.4845924453280318, + "grad_norm": 0.7014921502795326, + "learning_rate": 8.622749129544774e-06, + "loss": 0.38191866874694824, + "num_input_tokens_seen": 109342496, + "step": 975, + "train_runtime": 13516.7536, + "train_tokens_per_second": 8089.405 + }, + { + "epoch": 0.4870775347912525, + "grad_norm": 0.6018900891142692, + "learning_rate": 8.60926942445909e-06, + "loss": 0.39454472064971924, + "num_input_tokens_seen": 109892888, + "step": 980, + "train_runtime": 13585.9169, + "train_tokens_per_second": 8088.735 + }, + { + "epoch": 0.48956262425447317, + "grad_norm": 0.8443712614972618, + "learning_rate": 8.595734721922472e-06, + "loss": 0.3783087730407715, + "num_input_tokens_seen": 110488784, + "step": 985, + "train_runtime": 13658.5619, + "train_tokens_per_second": 8089.342 + }, + { + "epoch": 0.49204771371769385, + "grad_norm": 0.6465720342647037, + "learning_rate": 8.58214522817448e-06, + "loss": 0.40636625289916994, + "num_input_tokens_seen": 111051672, + "step": 990, + "train_runtime": 13737.7614, + "train_tokens_per_second": 8083.68 + }, + { + "epoch": 0.4945328031809145, + "grad_norm": 0.6282542532285146, + "learning_rate": 8.568501150289571e-06, + "loss": 0.38943657875061033, + "num_input_tokens_seen": 111614256, + "step": 995, + "train_runtime": 13802.9585, + "train_tokens_per_second": 8086.256 + }, + { + "epoch": 0.4970178926441352, + "grad_norm": 0.6402509235410074, + "learning_rate": 8.554802696173952e-06, + "loss": 0.3624922037124634, + "num_input_tokens_seen": 112180632, + "step": 1000, + "train_runtime": 13877.7512, + "train_tokens_per_second": 8083.488 + }, + { + "epoch": 0.4995029821073559, + "grad_norm": 0.6358140885715353, + "learning_rate": 8.5410500745624e-06, + "loss": 0.41681599617004395, + "num_input_tokens_seen": 112738584, + "step": 1005, + "train_runtime": 13937.7359, + "train_tokens_per_second": 8088.73 + }, + { + "epoch": 0.5019880715705766, + "grad_norm": 0.655191044508995, + "learning_rate": 8.527243495015093e-06, + "loss": 0.40607786178588867, + "num_input_tokens_seen": 113274456, + "step": 1010, + "train_runtime": 14007.6145, + "train_tokens_per_second": 8086.634 + }, + { + "epoch": 0.5044731610337972, + "grad_norm": 0.6922933445516329, + "learning_rate": 8.513383167914406e-06, + "loss": 0.3852816581726074, + "num_input_tokens_seen": 113844560, + "step": 1015, + "train_runtime": 14085.0289, + "train_tokens_per_second": 8082.664 + }, + { + "epoch": 0.5069582504970179, + "grad_norm": 0.7045126061031313, + "learning_rate": 8.499469304461711e-06, + "loss": 0.3989840507507324, + "num_input_tokens_seen": 114410088, + "step": 1020, + "train_runtime": 14161.1215, + "train_tokens_per_second": 8079.169 + }, + { + "epoch": 0.5094433399602386, + "grad_norm": 0.7241386038488862, + "learning_rate": 8.485502116674164e-06, + "loss": 0.38511319160461427, + "num_input_tokens_seen": 114958920, + "step": 1025, + "train_runtime": 14225.1432, + "train_tokens_per_second": 8081.389 + }, + { + "epoch": 0.5119284294234593, + "grad_norm": 0.707901759928824, + "learning_rate": 8.471481817381461e-06, + "loss": 0.3943337440490723, + "num_input_tokens_seen": 115531768, + "step": 1030, + "train_runtime": 14291.4805, + "train_tokens_per_second": 8083.961 + }, + { + "epoch": 0.51441351888668, + "grad_norm": 0.6407962508436487, + "learning_rate": 8.457408620222606e-06, + "loss": 0.3882716655731201, + "num_input_tokens_seen": 116140696, + "step": 1035, + "train_runtime": 14357.8181, + "train_tokens_per_second": 8089.021 + }, + { + "epoch": 0.5168986083499006, + "grad_norm": 0.6039192410277909, + "learning_rate": 8.44328273964265e-06, + "loss": 0.3741127014160156, + "num_input_tokens_seen": 116766688, + "step": 1040, + "train_runtime": 14437.1437, + "train_tokens_per_second": 8087.936 + }, + { + "epoch": 0.5193836978131213, + "grad_norm": 0.6778414376038925, + "learning_rate": 8.429104390889428e-06, + "loss": 0.36663236618041994, + "num_input_tokens_seen": 117352872, + "step": 1045, + "train_runtime": 14507.687, + "train_tokens_per_second": 8089.013 + }, + { + "epoch": 0.5218687872763419, + "grad_norm": 0.7873398498367088, + "learning_rate": 8.414873790010273e-06, + "loss": 0.38754978179931643, + "num_input_tokens_seen": 117928536, + "step": 1050, + "train_runtime": 14583.9954, + "train_tokens_per_second": 8086.161 + }, + { + "epoch": 0.5243538767395626, + "grad_norm": 0.7688694726129269, + "learning_rate": 8.400591153848731e-06, + "loss": 0.3773825645446777, + "num_input_tokens_seen": 118499352, + "step": 1055, + "train_runtime": 14657.9337, + "train_tokens_per_second": 8084.315 + }, + { + "epoch": 0.5268389662027833, + "grad_norm": 0.6825595915690184, + "learning_rate": 8.386256700041253e-06, + "loss": 0.38932156562805176, + "num_input_tokens_seen": 119071752, + "step": 1060, + "train_runtime": 14729.5361, + "train_tokens_per_second": 8083.877 + }, + { + "epoch": 0.5293240556660039, + "grad_norm": 0.6865901984473853, + "learning_rate": 8.37187064701387e-06, + "loss": 0.37840564250946046, + "num_input_tokens_seen": 119646592, + "step": 1065, + "train_runtime": 14795.4073, + "train_tokens_per_second": 8086.739 + }, + { + "epoch": 0.5318091451292246, + "grad_norm": 0.6560972250410257, + "learning_rate": 8.357433213978889e-06, + "loss": 0.405233097076416, + "num_input_tokens_seen": 120206464, + "step": 1070, + "train_runtime": 14856.4765, + "train_tokens_per_second": 8091.183 + }, + { + "epoch": 0.5342942345924453, + "grad_norm": 0.636106085370861, + "learning_rate": 8.342944620931522e-06, + "loss": 0.38074421882629395, + "num_input_tokens_seen": 120750128, + "step": 1075, + "train_runtime": 14921.9978, + "train_tokens_per_second": 8092.089 + }, + { + "epoch": 0.536779324055666, + "grad_norm": 0.6074845338444187, + "learning_rate": 8.328405088646557e-06, + "loss": 0.39549171924591064, + "num_input_tokens_seen": 121351360, + "step": 1080, + "train_runtime": 14997.2426, + "train_tokens_per_second": 8091.578 + }, + { + "epoch": 0.5392644135188867, + "grad_norm": 0.5939211493585828, + "learning_rate": 8.313814838674983e-06, + "loss": 0.3823257923126221, + "num_input_tokens_seen": 121920816, + "step": 1085, + "train_runtime": 15067.8712, + "train_tokens_per_second": 8091.443 + }, + { + "epoch": 0.5417495029821073, + "grad_norm": 0.6481772623574006, + "learning_rate": 8.299174093340619e-06, + "loss": 0.4083685874938965, + "num_input_tokens_seen": 122487496, + "step": 1090, + "train_runtime": 15129.4565, + "train_tokens_per_second": 8095.961 + }, + { + "epoch": 0.544234592445328, + "grad_norm": 0.81158799662579, + "learning_rate": 8.284483075736723e-06, + "loss": 0.3880143165588379, + "num_input_tokens_seen": 123066456, + "step": 1095, + "train_runtime": 15205.0918, + "train_tokens_per_second": 8093.766 + }, + { + "epoch": 0.5467196819085487, + "grad_norm": 0.6375832184541546, + "learning_rate": 8.26974200972259e-06, + "loss": 0.39032974243164065, + "num_input_tokens_seen": 123650328, + "step": 1100, + "train_runtime": 15279.4272, + "train_tokens_per_second": 8092.602 + }, + { + "epoch": 0.5492047713717694, + "grad_norm": 0.6584422685262109, + "learning_rate": 8.254951119920149e-06, + "loss": 0.3937986850738525, + "num_input_tokens_seen": 124179544, + "step": 1105, + "train_runtime": 15352.2892, + "train_tokens_per_second": 8088.666 + }, + { + "epoch": 0.55168986083499, + "grad_norm": 0.6207226466409983, + "learning_rate": 8.240110631710536e-06, + "loss": 0.3630195379257202, + "num_input_tokens_seen": 124736368, + "step": 1110, + "train_runtime": 15419.5775, + "train_tokens_per_second": 8089.48 + }, + { + "epoch": 0.5541749502982107, + "grad_norm": 0.6329597405670121, + "learning_rate": 8.225220771230658e-06, + "loss": 0.3958831787109375, + "num_input_tokens_seen": 125316664, + "step": 1115, + "train_runtime": 15487.3306, + "train_tokens_per_second": 8091.56 + }, + { + "epoch": 0.5566600397614314, + "grad_norm": 0.5526793955696551, + "learning_rate": 8.210281765369743e-06, + "loss": 0.3745615482330322, + "num_input_tokens_seen": 125915520, + "step": 1120, + "train_runtime": 15556.9074, + "train_tokens_per_second": 8093.866 + }, + { + "epoch": 0.5591451292246521, + "grad_norm": 0.5859273875364904, + "learning_rate": 8.195293841765897e-06, + "loss": 0.3736714839935303, + "num_input_tokens_seen": 126476976, + "step": 1125, + "train_runtime": 15628.6615, + "train_tokens_per_second": 8092.63 + }, + { + "epoch": 0.5616302186878728, + "grad_norm": 0.659199210263915, + "learning_rate": 8.180257228802621e-06, + "loss": 0.36822996139526365, + "num_input_tokens_seen": 127043000, + "step": 1130, + "train_runtime": 15698.9748, + "train_tokens_per_second": 8092.439 + }, + { + "epoch": 0.5641153081510935, + "grad_norm": 0.6976627605454907, + "learning_rate": 8.165172155605339e-06, + "loss": 0.3859762907028198, + "num_input_tokens_seen": 127599192, + "step": 1135, + "train_runtime": 15764.5745, + "train_tokens_per_second": 8094.046 + }, + { + "epoch": 0.5666003976143141, + "grad_norm": 0.6025077772690013, + "learning_rate": 8.150038852037903e-06, + "loss": 0.3925613880157471, + "num_input_tokens_seen": 128169864, + "step": 1140, + "train_runtime": 15833.0499, + "train_tokens_per_second": 8095.084 + }, + { + "epoch": 0.5690854870775348, + "grad_norm": 0.6814747502877871, + "learning_rate": 8.134857548699091e-06, + "loss": 0.3835327625274658, + "num_input_tokens_seen": 128697136, + "step": 1145, + "train_runtime": 15895.9708, + "train_tokens_per_second": 8096.211 + }, + { + "epoch": 0.5715705765407555, + "grad_norm": 0.6003090687300624, + "learning_rate": 8.119628476919095e-06, + "loss": 0.3627269983291626, + "num_input_tokens_seen": 129232632, + "step": 1150, + "train_runtime": 15966.8686, + "train_tokens_per_second": 8093.799 + }, + { + "epoch": 0.5740556660039762, + "grad_norm": 0.6628084793416656, + "learning_rate": 8.104351868755992e-06, + "loss": 0.37092711925506594, + "num_input_tokens_seen": 129759592, + "step": 1155, + "train_runtime": 16035.7968, + "train_tokens_per_second": 8091.871 + }, + { + "epoch": 0.5765407554671969, + "grad_norm": 0.7189877059058183, + "learning_rate": 8.089027956992214e-06, + "loss": 0.4017021179199219, + "num_input_tokens_seen": 130294208, + "step": 1160, + "train_runtime": 16096.8303, + "train_tokens_per_second": 8094.402 + }, + { + "epoch": 0.5790258449304175, + "grad_norm": 0.7119757741808245, + "learning_rate": 8.073656975130995e-06, + "loss": 0.3734596729278564, + "num_input_tokens_seen": 130870896, + "step": 1165, + "train_runtime": 16173.6617, + "train_tokens_per_second": 8091.606 + }, + { + "epoch": 0.5815109343936382, + "grad_norm": 0.6243270508977441, + "learning_rate": 8.058239157392815e-06, + "loss": 0.3912842035293579, + "num_input_tokens_seen": 131440608, + "step": 1170, + "train_runtime": 16240.9535, + "train_tokens_per_second": 8093.158 + }, + { + "epoch": 0.5839960238568589, + "grad_norm": 0.6415228748591238, + "learning_rate": 8.042774738711832e-06, + "loss": 0.3868302345275879, + "num_input_tokens_seen": 132026040, + "step": 1175, + "train_runtime": 16302.608, + "train_tokens_per_second": 8098.461 + }, + { + "epoch": 0.5864811133200796, + "grad_norm": 0.660675692185134, + "learning_rate": 8.027263954732305e-06, + "loss": 0.3828722953796387, + "num_input_tokens_seen": 132608688, + "step": 1180, + "train_runtime": 16363.41, + "train_tokens_per_second": 8103.976 + }, + { + "epoch": 0.5889662027833003, + "grad_norm": 0.6457629518981075, + "learning_rate": 8.011707041804993e-06, + "loss": 0.3846240520477295, + "num_input_tokens_seen": 133188232, + "step": 1185, + "train_runtime": 16433.3802, + "train_tokens_per_second": 8104.737 + }, + { + "epoch": 0.5914512922465208, + "grad_norm": 0.6806781120578339, + "learning_rate": 7.996104236983566e-06, + "loss": 0.3660452842712402, + "num_input_tokens_seen": 133723488, + "step": 1190, + "train_runtime": 16497.8253, + "train_tokens_per_second": 8105.522 + }, + { + "epoch": 0.5939363817097415, + "grad_norm": 0.6042226790950421, + "learning_rate": 7.980455778020986e-06, + "loss": 0.3905783653259277, + "num_input_tokens_seen": 134307272, + "step": 1195, + "train_runtime": 16570.4397, + "train_tokens_per_second": 8105.233 + }, + { + "epoch": 0.5964214711729622, + "grad_norm": 0.562342001632669, + "learning_rate": 7.96476190336588e-06, + "loss": 0.36987967491149903, + "num_input_tokens_seen": 134905480, + "step": 1200, + "train_runtime": 16639.755, + "train_tokens_per_second": 8107.42 + }, + { + "epoch": 0.5989065606361829, + "grad_norm": 0.742157777215612, + "learning_rate": 7.949022852158921e-06, + "loss": 0.37041635513305665, + "num_input_tokens_seen": 135476824, + "step": 1205, + "train_runtime": 16708.7869, + "train_tokens_per_second": 8108.118 + }, + { + "epoch": 0.6013916500994035, + "grad_norm": 0.7383638768557731, + "learning_rate": 7.933238864229166e-06, + "loss": 0.3739286422729492, + "num_input_tokens_seen": 136021520, + "step": 1210, + "train_runtime": 16774.9487, + "train_tokens_per_second": 8108.61 + }, + { + "epoch": 0.6038767395626242, + "grad_norm": 0.6657530106331175, + "learning_rate": 7.917410180090416e-06, + "loss": 0.3720057964324951, + "num_input_tokens_seen": 136589416, + "step": 1215, + "train_runtime": 16845.9544, + "train_tokens_per_second": 8108.144 + }, + { + "epoch": 0.6063618290258449, + "grad_norm": 0.6102353754325571, + "learning_rate": 7.901537040937543e-06, + "loss": 0.36920747756958006, + "num_input_tokens_seen": 137190088, + "step": 1220, + "train_runtime": 16919.7712, + "train_tokens_per_second": 8108.271 + }, + { + "epoch": 0.6088469184890656, + "grad_norm": 0.5734944612414656, + "learning_rate": 7.88561968864282e-06, + "loss": 0.38166260719299316, + "num_input_tokens_seen": 137778720, + "step": 1225, + "train_runtime": 16988.1987, + "train_tokens_per_second": 8110.261 + }, + { + "epoch": 0.6113320079522863, + "grad_norm": 0.7324677714151097, + "learning_rate": 7.869658365752229e-06, + "loss": 0.3543116092681885, + "num_input_tokens_seen": 138343216, + "step": 1230, + "train_runtime": 17064.2491, + "train_tokens_per_second": 8107.196 + }, + { + "epoch": 0.6138170974155069, + "grad_norm": 0.6333421052732945, + "learning_rate": 7.853653315481769e-06, + "loss": 0.382243275642395, + "num_input_tokens_seen": 138944400, + "step": 1235, + "train_runtime": 17142.5464, + "train_tokens_per_second": 8105.237 + }, + { + "epoch": 0.6163021868787276, + "grad_norm": 0.6712985934854361, + "learning_rate": 7.83760478171375e-06, + "loss": 0.39281835556030276, + "num_input_tokens_seen": 139495352, + "step": 1240, + "train_runtime": 17214.7266, + "train_tokens_per_second": 8103.257 + }, + { + "epoch": 0.6187872763419483, + "grad_norm": 0.6377938809983201, + "learning_rate": 7.82151300899308e-06, + "loss": 0.3440825939178467, + "num_input_tokens_seen": 140064928, + "step": 1245, + "train_runtime": 17271.6076, + "train_tokens_per_second": 8109.548 + }, + { + "epoch": 0.621272365805169, + "grad_norm": 0.6186734384142927, + "learning_rate": 7.805378242523529e-06, + "loss": 0.37058846950531005, + "num_input_tokens_seen": 140601768, + "step": 1250, + "train_runtime": 17334.764, + "train_tokens_per_second": 8110.971 + }, + { + "epoch": 0.6237574552683897, + "grad_norm": 0.640286957080056, + "learning_rate": 7.789200728164004e-06, + "loss": 0.36435441970825194, + "num_input_tokens_seen": 141182728, + "step": 1255, + "train_runtime": 17414.2994, + "train_tokens_per_second": 8107.287 + }, + { + "epoch": 0.6262425447316103, + "grad_norm": 0.6497575652941571, + "learning_rate": 7.772980712424794e-06, + "loss": 0.38532819747924807, + "num_input_tokens_seen": 141758672, + "step": 1260, + "train_runtime": 17489.3553, + "train_tokens_per_second": 8105.426 + }, + { + "epoch": 0.628727634194831, + "grad_norm": 0.5673102834438465, + "learning_rate": 7.756718442463817e-06, + "loss": 0.3860148906707764, + "num_input_tokens_seen": 142337456, + "step": 1265, + "train_runtime": 17558.2407, + "train_tokens_per_second": 8106.59 + }, + { + "epoch": 0.6312127236580517, + "grad_norm": 0.5753452534076092, + "learning_rate": 7.740414166082857e-06, + "loss": 0.37022929191589354, + "num_input_tokens_seen": 142914400, + "step": 1270, + "train_runtime": 17626.4592, + "train_tokens_per_second": 8107.947 + }, + { + "epoch": 0.6336978131212724, + "grad_norm": 0.6306375674628032, + "learning_rate": 7.724068131723781e-06, + "loss": 0.37794060707092286, + "num_input_tokens_seen": 143501152, + "step": 1275, + "train_runtime": 17694.4825, + "train_tokens_per_second": 8109.938 + }, + { + "epoch": 0.6361829025844931, + "grad_norm": 0.6727820101242223, + "learning_rate": 7.707680588464759e-06, + "loss": 0.37308588027954104, + "num_input_tokens_seen": 144099960, + "step": 1280, + "train_runtime": 17762.404, + "train_tokens_per_second": 8112.638 + }, + { + "epoch": 0.6386679920477137, + "grad_norm": 0.6220365845071045, + "learning_rate": 7.691251786016466e-06, + "loss": 0.3680124759674072, + "num_input_tokens_seen": 144669264, + "step": 1285, + "train_runtime": 17829.2163, + "train_tokens_per_second": 8114.168 + }, + { + "epoch": 0.6411530815109344, + "grad_norm": 0.6431513529985349, + "learning_rate": 7.67478197471828e-06, + "loss": 0.3770967960357666, + "num_input_tokens_seen": 145212928, + "step": 1290, + "train_runtime": 17888.1679, + "train_tokens_per_second": 8117.82 + }, + { + "epoch": 0.6436381709741551, + "grad_norm": 0.6280779164316928, + "learning_rate": 7.65827140553446e-06, + "loss": 0.3764733552932739, + "num_input_tokens_seen": 145797632, + "step": 1295, + "train_runtime": 17965.0868, + "train_tokens_per_second": 8115.61 + }, + { + "epoch": 0.6461232604373758, + "grad_norm": 0.7692882747612924, + "learning_rate": 7.641720330050335e-06, + "loss": 0.3946826457977295, + "num_input_tokens_seen": 146395432, + "step": 1300, + "train_runtime": 18036.9172, + "train_tokens_per_second": 8116.433 + }, + { + "epoch": 0.6486083499005965, + "grad_norm": 0.5791676866606946, + "learning_rate": 7.625129000468455e-06, + "loss": 0.37627956867218015, + "num_input_tokens_seen": 146985480, + "step": 1305, + "train_runtime": 18097.5271, + "train_tokens_per_second": 8121.854 + }, + { + "epoch": 0.6510934393638171, + "grad_norm": 0.7578857552729431, + "learning_rate": 7.608497669604758e-06, + "loss": 0.39343085289001467, + "num_input_tokens_seen": 147574704, + "step": 1310, + "train_runtime": 18168.814, + "train_tokens_per_second": 8122.418 + }, + { + "epoch": 0.6535785288270378, + "grad_norm": 0.5694109222114526, + "learning_rate": 7.591826590884714e-06, + "loss": 0.37366428375244143, + "num_input_tokens_seen": 148159576, + "step": 1315, + "train_runtime": 18236.4333, + "train_tokens_per_second": 8124.372 + }, + { + "epoch": 0.6560636182902585, + "grad_norm": 0.626832671444105, + "learning_rate": 7.575116018339468e-06, + "loss": 0.3778612375259399, + "num_input_tokens_seen": 148698336, + "step": 1320, + "train_runtime": 18310.7689, + "train_tokens_per_second": 8120.813 + }, + { + "epoch": 0.6585487077534792, + "grad_norm": 0.5934886755150511, + "learning_rate": 7.558366206601959e-06, + "loss": 0.392897891998291, + "num_input_tokens_seen": 149247272, + "step": 1325, + "train_runtime": 18383.9895, + "train_tokens_per_second": 8118.329 + }, + { + "epoch": 0.6610337972166997, + "grad_norm": 0.5807546911283734, + "learning_rate": 7.541577410903051e-06, + "loss": 0.37039999961853026, + "num_input_tokens_seen": 149837744, + "step": 1330, + "train_runtime": 18453.4609, + "train_tokens_per_second": 8119.764 + }, + { + "epoch": 0.6635188866799204, + "grad_norm": 0.6295430870892492, + "learning_rate": 7.5247498870676385e-06, + "loss": 0.3697408676147461, + "num_input_tokens_seen": 150405216, + "step": 1335, + "train_runtime": 18525.8576, + "train_tokens_per_second": 8118.664 + }, + { + "epoch": 0.6660039761431411, + "grad_norm": 0.6121907368436152, + "learning_rate": 7.507883891510748e-06, + "loss": 0.39031453132629396, + "num_input_tokens_seen": 150959584, + "step": 1340, + "train_runtime": 18597.6357, + "train_tokens_per_second": 8117.138 + }, + { + "epoch": 0.6684890656063618, + "grad_norm": 0.5883566650638485, + "learning_rate": 7.490979681233632e-06, + "loss": 0.3696321725845337, + "num_input_tokens_seen": 151531504, + "step": 1345, + "train_runtime": 18662.0136, + "train_tokens_per_second": 8119.783 + }, + { + "epoch": 0.6709741550695825, + "grad_norm": 0.6230296069159287, + "learning_rate": 7.474037513819854e-06, + "loss": 0.36713156700134275, + "num_input_tokens_seen": 152118072, + "step": 1350, + "train_runtime": 18729.4197, + "train_tokens_per_second": 8121.879 + }, + { + "epoch": 0.6734592445328031, + "grad_norm": 0.6503566554892475, + "learning_rate": 7.457057647431359e-06, + "loss": 0.37275185585021975, + "num_input_tokens_seen": 152689736, + "step": 1355, + "train_runtime": 18795.5153, + "train_tokens_per_second": 8123.732 + }, + { + "epoch": 0.6759443339960238, + "grad_norm": 0.6780715854584781, + "learning_rate": 7.440040340804544e-06, + "loss": 0.38737239837646487, + "num_input_tokens_seen": 153265432, + "step": 1360, + "train_runtime": 18855.2674, + "train_tokens_per_second": 8128.521 + }, + { + "epoch": 0.6784294234592445, + "grad_norm": 0.6011465963199712, + "learning_rate": 7.4229858532463165e-06, + "loss": 0.377882719039917, + "num_input_tokens_seen": 153827688, + "step": 1365, + "train_runtime": 18916.1623, + "train_tokens_per_second": 8132.077 + }, + { + "epoch": 0.6809145129224652, + "grad_norm": 0.6859801824488166, + "learning_rate": 7.405894444630139e-06, + "loss": 0.39091172218322756, + "num_input_tokens_seen": 154383160, + "step": 1370, + "train_runtime": 18981.7307, + "train_tokens_per_second": 8133.25 + }, + { + "epoch": 0.6833996023856859, + "grad_norm": 0.7401084359843174, + "learning_rate": 7.388766375392069e-06, + "loss": 0.3995502948760986, + "num_input_tokens_seen": 154956304, + "step": 1375, + "train_runtime": 19056.5453, + "train_tokens_per_second": 8131.395 + }, + { + "epoch": 0.6858846918489065, + "grad_norm": 0.6278629560645018, + "learning_rate": 7.371601906526798e-06, + "loss": 0.3924835681915283, + "num_input_tokens_seen": 155539704, + "step": 1380, + "train_runtime": 19129.2416, + "train_tokens_per_second": 8130.992 + }, + { + "epoch": 0.6883697813121272, + "grad_norm": 0.632656964383803, + "learning_rate": 7.354401299583663e-06, + "loss": 0.3653599500656128, + "num_input_tokens_seen": 156109744, + "step": 1385, + "train_runtime": 19196.2196, + "train_tokens_per_second": 8132.317 + }, + { + "epoch": 0.6908548707753479, + "grad_norm": 0.6798835858095437, + "learning_rate": 7.3371648166626695e-06, + "loss": 0.37296769618988035, + "num_input_tokens_seen": 156671784, + "step": 1390, + "train_runtime": 19259.1299, + "train_tokens_per_second": 8134.936 + }, + { + "epoch": 0.6933399602385686, + "grad_norm": 0.6193944959064848, + "learning_rate": 7.3198927204104974e-06, + "loss": 0.38158082962036133, + "num_input_tokens_seen": 157274000, + "step": 1395, + "train_runtime": 19334.7946, + "train_tokens_per_second": 8134.247 + }, + { + "epoch": 0.6958250497017893, + "grad_norm": 0.6562734928047045, + "learning_rate": 7.302585274016492e-06, + "loss": 0.36525468826293944, + "num_input_tokens_seen": 157860416, + "step": 1400, + "train_runtime": 19405.0749, + "train_tokens_per_second": 8135.007 + }, + { + "epoch": 0.69831013916501, + "grad_norm": 0.6467875664763257, + "learning_rate": 7.285242741208662e-06, + "loss": 0.3774902582168579, + "num_input_tokens_seen": 158412552, + "step": 1405, + "train_runtime": 19476.9711, + "train_tokens_per_second": 8133.326 + }, + { + "epoch": 0.7007952286282306, + "grad_norm": 0.6562795038922349, + "learning_rate": 7.267865386249657e-06, + "loss": 0.3876355648040771, + "num_input_tokens_seen": 158952128, + "step": 1410, + "train_runtime": 19548.8662, + "train_tokens_per_second": 8131.015 + }, + { + "epoch": 0.7032803180914513, + "grad_norm": 0.6242657809964064, + "learning_rate": 7.250453473932737e-06, + "loss": 0.37313523292541506, + "num_input_tokens_seen": 159537712, + "step": 1415, + "train_runtime": 19623.8101, + "train_tokens_per_second": 8129.803 + }, + { + "epoch": 0.705765407554672, + "grad_norm": 0.6007436211080148, + "learning_rate": 7.233007269577748e-06, + "loss": 0.3646975517272949, + "num_input_tokens_seen": 160088072, + "step": 1420, + "train_runtime": 19685.2067, + "train_tokens_per_second": 8132.405 + }, + { + "epoch": 0.7082504970178927, + "grad_norm": 0.6857638946160585, + "learning_rate": 7.215527039027063e-06, + "loss": 0.3755317211151123, + "num_input_tokens_seen": 160647592, + "step": 1425, + "train_runtime": 19759.3561, + "train_tokens_per_second": 8130.204 + }, + { + "epoch": 0.7107355864811133, + "grad_norm": 0.6740440365502317, + "learning_rate": 7.19801304864155e-06, + "loss": 0.3877701759338379, + "num_input_tokens_seen": 161226352, + "step": 1430, + "train_runtime": 19830.2038, + "train_tokens_per_second": 8130.343 + }, + { + "epoch": 0.713220675944334, + "grad_norm": 0.670158587798895, + "learning_rate": 7.180465565296499e-06, + "loss": 0.38379640579223634, + "num_input_tokens_seen": 161781648, + "step": 1435, + "train_runtime": 19896.867, + "train_tokens_per_second": 8131.011 + }, + { + "epoch": 0.7157057654075547, + "grad_norm": 0.6496913670152225, + "learning_rate": 7.162884856377561e-06, + "loss": 0.38007800579071044, + "num_input_tokens_seen": 162308816, + "step": 1440, + "train_runtime": 19968.7987, + "train_tokens_per_second": 8128.121 + }, + { + "epoch": 0.7181908548707754, + "grad_norm": 0.6458110268551169, + "learning_rate": 7.145271189776671e-06, + "loss": 0.37517600059509276, + "num_input_tokens_seen": 162860320, + "step": 1445, + "train_runtime": 20036.3648, + "train_tokens_per_second": 8128.237 + }, + { + "epoch": 0.7206759443339961, + "grad_norm": 0.6259957110185805, + "learning_rate": 7.127624833887972e-06, + "loss": 0.381079363822937, + "num_input_tokens_seen": 163423192, + "step": 1450, + "train_runtime": 20097.4601, + "train_tokens_per_second": 8131.535 + }, + { + "epoch": 0.7231610337972167, + "grad_norm": 0.6883099889629782, + "learning_rate": 7.109946057603715e-06, + "loss": 0.354444694519043, + "num_input_tokens_seen": 164002768, + "step": 1455, + "train_runtime": 20169.7942, + "train_tokens_per_second": 8131.108 + }, + { + "epoch": 0.7256461232604374, + "grad_norm": 0.7105202074547773, + "learning_rate": 7.092235130310173e-06, + "loss": 0.3573470115661621, + "num_input_tokens_seen": 164567208, + "step": 1460, + "train_runtime": 20238.6169, + "train_tokens_per_second": 8131.347 + }, + { + "epoch": 0.7281312127236581, + "grad_norm": 0.6103797428673515, + "learning_rate": 7.074492321883529e-06, + "loss": 0.3674468994140625, + "num_input_tokens_seen": 165152808, + "step": 1465, + "train_runtime": 20291.2575, + "train_tokens_per_second": 8139.112 + }, + { + "epoch": 0.7306163021868787, + "grad_norm": 0.6312950432397176, + "learning_rate": 7.056717902685766e-06, + "loss": 0.36470980644226075, + "num_input_tokens_seen": 165672592, + "step": 1470, + "train_runtime": 20370.2056, + "train_tokens_per_second": 8133.084 + }, + { + "epoch": 0.7331013916500994, + "grad_norm": 0.6324132099709199, + "learning_rate": 7.038912143560544e-06, + "loss": 0.36964056491851804, + "num_input_tokens_seen": 166220392, + "step": 1475, + "train_runtime": 20442.4995, + "train_tokens_per_second": 8131.119 + }, + { + "epoch": 0.73558648111332, + "grad_norm": 0.6532083359524653, + "learning_rate": 7.0210753158290775e-06, + "loss": 0.3760281801223755, + "num_input_tokens_seen": 166799584, + "step": 1480, + "train_runtime": 20510.1687, + "train_tokens_per_second": 8132.531 + }, + { + "epoch": 0.7380715705765407, + "grad_norm": 0.6707624646370762, + "learning_rate": 7.003207691286002e-06, + "loss": 0.34601802825927735, + "num_input_tokens_seen": 167374976, + "step": 1485, + "train_runtime": 20588.8271, + "train_tokens_per_second": 8129.408 + }, + { + "epoch": 0.7405566600397614, + "grad_norm": 0.6264520062030375, + "learning_rate": 6.985309542195223e-06, + "loss": 0.36390366554260256, + "num_input_tokens_seen": 167968576, + "step": 1490, + "train_runtime": 20654.3184, + "train_tokens_per_second": 8132.371 + }, + { + "epoch": 0.7430417495029821, + "grad_norm": 0.6884500100471885, + "learning_rate": 6.967381141285782e-06, + "loss": 0.38413434028625487, + "num_input_tokens_seen": 168512152, + "step": 1495, + "train_runtime": 20731.0823, + "train_tokens_per_second": 8128.478 + }, + { + "epoch": 0.7455268389662028, + "grad_norm": 0.6084887950391518, + "learning_rate": 6.949422761747686e-06, + "loss": 0.3661206722259521, + "num_input_tokens_seen": 169117760, + "step": 1500, + "train_runtime": 20806.7346, + "train_tokens_per_second": 8128.03 + }, + { + "epoch": 0.7480119284294234, + "grad_norm": 0.6120245127276173, + "learning_rate": 6.931434677227757e-06, + "loss": 0.3716851472854614, + "num_input_tokens_seen": 169679680, + "step": 1505, + "train_runtime": 20874.8979, + "train_tokens_per_second": 8128.408 + }, + { + "epoch": 0.7504970178926441, + "grad_norm": 0.6421899460992642, + "learning_rate": 6.913417161825449e-06, + "loss": 0.3560441017150879, + "num_input_tokens_seen": 170245984, + "step": 1510, + "train_runtime": 20936.7667, + "train_tokens_per_second": 8131.436 + }, + { + "epoch": 0.7529821073558648, + "grad_norm": 0.6092828908557618, + "learning_rate": 6.895370490088689e-06, + "loss": 0.3637753963470459, + "num_input_tokens_seen": 170835960, + "step": 1515, + "train_runtime": 21012.0106, + "train_tokens_per_second": 8130.396 + }, + { + "epoch": 0.7554671968190855, + "grad_norm": 0.5934542091805799, + "learning_rate": 6.877294937009674e-06, + "loss": 0.36241917610168456, + "num_input_tokens_seen": 171414384, + "step": 1520, + "train_runtime": 21078.3625, + "train_tokens_per_second": 8132.244 + }, + { + "epoch": 0.7579522862823062, + "grad_norm": 0.6024497397273366, + "learning_rate": 6.859190778020699e-06, + "loss": 0.36909372806549073, + "num_input_tokens_seen": 172000600, + "step": 1525, + "train_runtime": 21151.1544, + "train_tokens_per_second": 8131.972 + }, + { + "epoch": 0.7604373757455268, + "grad_norm": 0.6730872247525814, + "learning_rate": 6.84105828898994e-06, + "loss": 0.35951976776123046, + "num_input_tokens_seen": 172542928, + "step": 1530, + "train_runtime": 21225.773, + "train_tokens_per_second": 8128.935 + }, + { + "epoch": 0.7629224652087475, + "grad_norm": 0.6018986454782195, + "learning_rate": 6.822897746217273e-06, + "loss": 0.38000712394714353, + "num_input_tokens_seen": 173125952, + "step": 1535, + "train_runtime": 21302.4516, + "train_tokens_per_second": 8127.044 + }, + { + "epoch": 0.7654075546719682, + "grad_norm": 0.6779080007730667, + "learning_rate": 6.804709426430044e-06, + "loss": 0.38850367069244385, + "num_input_tokens_seen": 173685288, + "step": 1540, + "train_runtime": 21374.849, + "train_tokens_per_second": 8125.685 + }, + { + "epoch": 0.7678926441351889, + "grad_norm": 1.1716673949458725, + "learning_rate": 6.7864936067788655e-06, + "loss": 0.37443814277648924, + "num_input_tokens_seen": 174228168, + "step": 1545, + "train_runtime": 21433.4408, + "train_tokens_per_second": 8128.801 + }, + { + "epoch": 0.7703777335984096, + "grad_norm": 0.6363044002977479, + "learning_rate": 6.768250564833386e-06, + "loss": 0.3762331962585449, + "num_input_tokens_seen": 174785480, + "step": 1550, + "train_runtime": 21499.8435, + "train_tokens_per_second": 8129.616 + }, + { + "epoch": 0.7728628230616302, + "grad_norm": 0.6554813903842375, + "learning_rate": 6.749980578578065e-06, + "loss": 0.3846780776977539, + "num_input_tokens_seen": 175319664, + "step": 1555, + "train_runtime": 21558.7725, + "train_tokens_per_second": 8132.173 + }, + { + "epoch": 0.7753479125248509, + "grad_norm": 0.5823578528588909, + "learning_rate": 6.73168392640793e-06, + "loss": 0.3836991786956787, + "num_input_tokens_seen": 175902832, + "step": 1560, + "train_runtime": 21619.053, + "train_tokens_per_second": 8136.473 + }, + { + "epoch": 0.7778330019880716, + "grad_norm": 0.7173080164331932, + "learning_rate": 6.713360887124345e-06, + "loss": 0.38359010219573975, + "num_input_tokens_seen": 176470592, + "step": 1565, + "train_runtime": 21689.6316, + "train_tokens_per_second": 8136.173 + }, + { + "epoch": 0.7803180914512923, + "grad_norm": 0.7091939720182803, + "learning_rate": 6.695011739930753e-06, + "loss": 0.3817564010620117, + "num_input_tokens_seen": 177051456, + "step": 1570, + "train_runtime": 21760.9902, + "train_tokens_per_second": 8136.186 + }, + { + "epoch": 0.782803180914513, + "grad_norm": 0.6788158743530545, + "learning_rate": 6.6766367644284256e-06, + "loss": 0.36972239017486574, + "num_input_tokens_seen": 177585568, + "step": 1575, + "train_runtime": 21827.5285, + "train_tokens_per_second": 8135.853 + }, + { + "epoch": 0.7852882703777336, + "grad_norm": 0.6541589775723099, + "learning_rate": 6.658236240612206e-06, + "loss": 0.4031380653381348, + "num_input_tokens_seen": 178160320, + "step": 1580, + "train_runtime": 21902.24, + "train_tokens_per_second": 8134.342 + }, + { + "epoch": 0.7877733598409543, + "grad_norm": 0.5604530784928683, + "learning_rate": 6.639810448866229e-06, + "loss": 0.3602180242538452, + "num_input_tokens_seen": 178733080, + "step": 1585, + "train_runtime": 21973.5423, + "train_tokens_per_second": 8134.013 + }, + { + "epoch": 0.790258449304175, + "grad_norm": 0.6141158580733346, + "learning_rate": 6.62135966995967e-06, + "loss": 0.3882112979888916, + "num_input_tokens_seen": 179302480, + "step": 1590, + "train_runtime": 22043.3679, + "train_tokens_per_second": 8134.078 + }, + { + "epoch": 0.7927435387673957, + "grad_norm": 0.6858196188615735, + "learning_rate": 6.602884185042443e-06, + "loss": 0.3694867372512817, + "num_input_tokens_seen": 179849296, + "step": 1595, + "train_runtime": 22114.4242, + "train_tokens_per_second": 8132.669 + }, + { + "epoch": 0.7952286282306164, + "grad_norm": 0.6800424046137808, + "learning_rate": 6.584384275640936e-06, + "loss": 0.38325796127319334, + "num_input_tokens_seen": 180390272, + "step": 1600, + "train_runtime": 22186.7809, + "train_tokens_per_second": 8130.529 + }, + { + "epoch": 0.797713717693837, + "grad_norm": 0.6055839865866774, + "learning_rate": 6.565860223653711e-06, + "loss": 0.380050802230835, + "num_input_tokens_seen": 180962120, + "step": 1605, + "train_runtime": 22252.135, + "train_tokens_per_second": 8132.349 + }, + { + "epoch": 0.8001988071570576, + "grad_norm": 0.6534675953398222, + "learning_rate": 6.547312311347209e-06, + "loss": 0.35499546527862547, + "num_input_tokens_seen": 181501104, + "step": 1610, + "train_runtime": 22310.4882, + "train_tokens_per_second": 8135.237 + }, + { + "epoch": 0.8026838966202783, + "grad_norm": 0.6100901808662069, + "learning_rate": 6.528740821351454e-06, + "loss": 0.35903818607330323, + "num_input_tokens_seen": 182029648, + "step": 1615, + "train_runtime": 22380.1867, + "train_tokens_per_second": 8133.518 + }, + { + "epoch": 0.805168986083499, + "grad_norm": 0.7216005820941789, + "learning_rate": 6.51014603665574e-06, + "loss": 0.3705106019973755, + "num_input_tokens_seen": 182580304, + "step": 1620, + "train_runtime": 22452.8035, + "train_tokens_per_second": 8131.737 + }, + { + "epoch": 0.8076540755467196, + "grad_norm": 0.724774261056537, + "learning_rate": 6.491528240604321e-06, + "loss": 0.35998854637145994, + "num_input_tokens_seen": 183176344, + "step": 1625, + "train_runtime": 22526.9861, + "train_tokens_per_second": 8131.418 + }, + { + "epoch": 0.8101391650099403, + "grad_norm": 0.5743129566941019, + "learning_rate": 6.472887716892099e-06, + "loss": 0.3614665985107422, + "num_input_tokens_seen": 183746040, + "step": 1630, + "train_runtime": 22600.0157, + "train_tokens_per_second": 8130.35 + }, + { + "epoch": 0.812624254473161, + "grad_norm": 0.6191864717572078, + "learning_rate": 6.454224749560292e-06, + "loss": 0.37230873107910156, + "num_input_tokens_seen": 184318032, + "step": 1635, + "train_runtime": 22668.5456, + "train_tokens_per_second": 8131.004 + }, + { + "epoch": 0.8151093439363817, + "grad_norm": 0.6389574654498897, + "learning_rate": 6.435539622992111e-06, + "loss": 0.3798834800720215, + "num_input_tokens_seen": 184873432, + "step": 1640, + "train_runtime": 22735.9801, + "train_tokens_per_second": 8131.316 + }, + { + "epoch": 0.8175944333996024, + "grad_norm": 0.7066198331408383, + "learning_rate": 6.416832621908427e-06, + "loss": 0.3715982437133789, + "num_input_tokens_seen": 185462104, + "step": 1645, + "train_runtime": 22808.9383, + "train_tokens_per_second": 8131.115 + }, + { + "epoch": 0.820079522862823, + "grad_norm": 0.8915397102983567, + "learning_rate": 6.398104031363431e-06, + "loss": 0.3546753406524658, + "num_input_tokens_seen": 186010952, + "step": 1650, + "train_runtime": 22874.6983, + "train_tokens_per_second": 8131.734 + }, + { + "epoch": 0.8225646123260437, + "grad_norm": 0.5916235464663082, + "learning_rate": 6.379354136740289e-06, + "loss": 0.35317416191101075, + "num_input_tokens_seen": 186574808, + "step": 1655, + "train_runtime": 22946.8516, + "train_tokens_per_second": 8130.737 + }, + { + "epoch": 0.8250497017892644, + "grad_norm": 0.6415938798199239, + "learning_rate": 6.360583223746797e-06, + "loss": 0.38242692947387696, + "num_input_tokens_seen": 187175160, + "step": 1660, + "train_runtime": 23017.774, + "train_tokens_per_second": 8131.766 + }, + { + "epoch": 0.8275347912524851, + "grad_norm": 0.6334738073655161, + "learning_rate": 6.3417915784110225e-06, + "loss": 0.37494263648986814, + "num_input_tokens_seen": 187705168, + "step": 1665, + "train_runtime": 23091.7158, + "train_tokens_per_second": 8128.68 + }, + { + "epoch": 0.8300198807157058, + "grad_norm": 0.6878866420362602, + "learning_rate": 6.3229794870769556e-06, + "loss": 0.37105731964111327, + "num_input_tokens_seen": 188248088, + "step": 1670, + "train_runtime": 23158.3978, + "train_tokens_per_second": 8128.718 + }, + { + "epoch": 0.8325049701789264, + "grad_norm": 0.591740735189097, + "learning_rate": 6.304147236400132e-06, + "loss": 0.3681910037994385, + "num_input_tokens_seen": 188824240, + "step": 1675, + "train_runtime": 23232.3056, + "train_tokens_per_second": 8127.658 + }, + { + "epoch": 0.8349900596421471, + "grad_norm": 0.6539791777058457, + "learning_rate": 6.285295113343275e-06, + "loss": 0.3783636331558228, + "num_input_tokens_seen": 189404880, + "step": 1680, + "train_runtime": 23311.8398, + "train_tokens_per_second": 8124.836 + }, + { + "epoch": 0.8374751491053678, + "grad_norm": 0.6379984466848608, + "learning_rate": 6.266423405171919e-06, + "loss": 0.37495999336242675, + "num_input_tokens_seen": 189960232, + "step": 1685, + "train_runtime": 23380.7956, + "train_tokens_per_second": 8124.627 + }, + { + "epoch": 0.8399602385685885, + "grad_norm": 0.6440128255736416, + "learning_rate": 6.247532399450034e-06, + "loss": 0.37276477813720704, + "num_input_tokens_seen": 190501944, + "step": 1690, + "train_runtime": 23433.8731, + "train_tokens_per_second": 8129.341 + }, + { + "epoch": 0.8424453280318092, + "grad_norm": 0.6357598871877349, + "learning_rate": 6.228622384035644e-06, + "loss": 0.36922607421875, + "num_input_tokens_seen": 191047552, + "step": 1695, + "train_runtime": 23510.0285, + "train_tokens_per_second": 8126.215 + }, + { + "epoch": 0.8449304174950298, + "grad_norm": 0.6785386642899698, + "learning_rate": 6.209693647076438e-06, + "loss": 0.3879129886627197, + "num_input_tokens_seen": 191606544, + "step": 1700, + "train_runtime": 23575.4821, + "train_tokens_per_second": 8127.365 + }, + { + "epoch": 0.8474155069582505, + "grad_norm": 0.5816351130772778, + "learning_rate": 6.190746477005383e-06, + "loss": 0.3573521614074707, + "num_input_tokens_seen": 192167832, + "step": 1705, + "train_runtime": 23646.4167, + "train_tokens_per_second": 8126.721 + }, + { + "epoch": 0.8499005964214712, + "grad_norm": 0.611254277728532, + "learning_rate": 6.171781162536321e-06, + "loss": 0.3799440383911133, + "num_input_tokens_seen": 192737368, + "step": 1710, + "train_runtime": 23710.5416, + "train_tokens_per_second": 8128.763 + }, + { + "epoch": 0.8523856858846919, + "grad_norm": 0.6578262189767939, + "learning_rate": 6.1527979926595804e-06, + "loss": 0.3529336452484131, + "num_input_tokens_seen": 193310376, + "step": 1715, + "train_runtime": 23777.0396, + "train_tokens_per_second": 8130.128 + }, + { + "epoch": 0.8548707753479126, + "grad_norm": 0.6971330894757081, + "learning_rate": 6.133797256637569e-06, + "loss": 0.40113162994384766, + "num_input_tokens_seen": 193889080, + "step": 1720, + "train_runtime": 23848.2218, + "train_tokens_per_second": 8130.127 + }, + { + "epoch": 0.8573558648111332, + "grad_norm": 0.6088194817802597, + "learning_rate": 6.114779244000358e-06, + "loss": 0.361390209197998, + "num_input_tokens_seen": 194470016, + "step": 1725, + "train_runtime": 23918.3977, + "train_tokens_per_second": 8130.562 + }, + { + "epoch": 0.8598409542743539, + "grad_norm": 0.6832990839129883, + "learning_rate": 6.095744244541284e-06, + "loss": 0.3713616609573364, + "num_input_tokens_seen": 195028640, + "step": 1730, + "train_runtime": 23992.1158, + "train_tokens_per_second": 8128.864 + }, + { + "epoch": 0.8623260437375746, + "grad_norm": 0.6317118407149473, + "learning_rate": 6.076692548312519e-06, + "loss": 0.35507752895355227, + "num_input_tokens_seen": 195575456, + "step": 1735, + "train_runtime": 24062.9782, + "train_tokens_per_second": 8127.65 + }, + { + "epoch": 0.8648111332007953, + "grad_norm": 0.7146221369152111, + "learning_rate": 6.057624445620665e-06, + "loss": 0.36352009773254396, + "num_input_tokens_seen": 196157544, + "step": 1740, + "train_runtime": 24131.5968, + "train_tokens_per_second": 8128.66 + }, + { + "epoch": 0.867296222664016, + "grad_norm": 0.7088936706828921, + "learning_rate": 6.038540227022316e-06, + "loss": 0.384311580657959, + "num_input_tokens_seen": 196750768, + "step": 1745, + "train_runtime": 24205.335, + "train_tokens_per_second": 8128.405 + }, + { + "epoch": 0.8697813121272365, + "grad_norm": 0.5972050885888135, + "learning_rate": 6.0194401833196415e-06, + "loss": 0.3668705940246582, + "num_input_tokens_seen": 197320664, + "step": 1750, + "train_runtime": 24279.6655, + "train_tokens_per_second": 8126.993 + }, + { + "epoch": 0.8722664015904572, + "grad_norm": 0.7021082461120339, + "learning_rate": 6.00032460555595e-06, + "loss": 0.3702070236206055, + "num_input_tokens_seen": 197878640, + "step": 1755, + "train_runtime": 24353.6434, + "train_tokens_per_second": 8125.217 + }, + { + "epoch": 0.8747514910536779, + "grad_norm": 0.6056290116544734, + "learning_rate": 5.981193785011259e-06, + "loss": 0.37683701515197754, + "num_input_tokens_seen": 198436200, + "step": 1760, + "train_runtime": 24414.8279, + "train_tokens_per_second": 8127.692 + }, + { + "epoch": 0.8772365805168986, + "grad_norm": 0.6479478521739579, + "learning_rate": 5.962048013197847e-06, + "loss": 0.37996611595153806, + "num_input_tokens_seen": 199006328, + "step": 1765, + "train_runtime": 24475.8906, + "train_tokens_per_second": 8130.708 + }, + { + "epoch": 0.8797216699801192, + "grad_norm": 0.6581445368624954, + "learning_rate": 5.942887581855822e-06, + "loss": 0.376295804977417, + "num_input_tokens_seen": 199495608, + "step": 1770, + "train_runtime": 24552.12, + "train_tokens_per_second": 8125.392 + }, + { + "epoch": 0.8822067594433399, + "grad_norm": 0.6730435073799281, + "learning_rate": 5.923712782948667e-06, + "loss": 0.38019070625305174, + "num_input_tokens_seen": 200051688, + "step": 1775, + "train_runtime": 24620.8508, + "train_tokens_per_second": 8125.295 + }, + { + "epoch": 0.8846918489065606, + "grad_norm": 0.6783482204025797, + "learning_rate": 5.904523908658803e-06, + "loss": 0.34175782203674315, + "num_input_tokens_seen": 200613432, + "step": 1780, + "train_runtime": 24692.092, + "train_tokens_per_second": 8124.602 + }, + { + "epoch": 0.8871769383697813, + "grad_norm": 0.6255358527217343, + "learning_rate": 5.88532125138312e-06, + "loss": 0.39377293586730955, + "num_input_tokens_seen": 201181112, + "step": 1785, + "train_runtime": 24770.3682, + "train_tokens_per_second": 8121.846 + }, + { + "epoch": 0.889662027833002, + "grad_norm": 0.6511607731402379, + "learning_rate": 5.866105103728539e-06, + "loss": 0.3675422668457031, + "num_input_tokens_seen": 201721712, + "step": 1790, + "train_runtime": 24846.4881, + "train_tokens_per_second": 8118.721 + }, + { + "epoch": 0.8921471172962226, + "grad_norm": 0.6097925440982134, + "learning_rate": 5.846875758507542e-06, + "loss": 0.35813026428222655, + "num_input_tokens_seen": 202289352, + "step": 1795, + "train_runtime": 24919.0734, + "train_tokens_per_second": 8117.852 + }, + { + "epoch": 0.8946322067594433, + "grad_norm": 0.6669576909090674, + "learning_rate": 5.82763350873371e-06, + "loss": 0.3726003408432007, + "num_input_tokens_seen": 202832392, + "step": 1800, + "train_runtime": 24984.6576, + "train_tokens_per_second": 8118.278 + }, + { + "epoch": 0.897117296222664, + "grad_norm": 0.6371272532557793, + "learning_rate": 5.808378647617268e-06, + "loss": 0.3683046817779541, + "num_input_tokens_seen": 203388544, + "step": 1805, + "train_runtime": 25057.8811, + "train_tokens_per_second": 8116.75 + }, + { + "epoch": 0.8996023856858847, + "grad_norm": 0.5971440620896147, + "learning_rate": 5.789111468560605e-06, + "loss": 0.354137659072876, + "num_input_tokens_seen": 203963352, + "step": 1810, + "train_runtime": 25110.4549, + "train_tokens_per_second": 8122.647 + }, + { + "epoch": 0.9020874751491054, + "grad_norm": 0.616124922832976, + "learning_rate": 5.769832265153812e-06, + "loss": 0.36770176887512207, + "num_input_tokens_seen": 204522408, + "step": 1815, + "train_runtime": 25182.1083, + "train_tokens_per_second": 8121.735 + }, + { + "epoch": 0.904572564612326, + "grad_norm": 0.7280646962260139, + "learning_rate": 5.750541331170203e-06, + "loss": 0.3491917610168457, + "num_input_tokens_seen": 205048168, + "step": 1820, + "train_runtime": 25257.0811, + "train_tokens_per_second": 8118.443 + }, + { + "epoch": 0.9070576540755467, + "grad_norm": 0.7016016822435476, + "learning_rate": 5.731238960561845e-06, + "loss": 0.38056321144104005, + "num_input_tokens_seen": 205560424, + "step": 1825, + "train_runtime": 25326.0118, + "train_tokens_per_second": 8116.573 + }, + { + "epoch": 0.9095427435387674, + "grad_norm": 0.6523741364307268, + "learning_rate": 5.711925447455066e-06, + "loss": 0.35477895736694337, + "num_input_tokens_seen": 206107504, + "step": 1830, + "train_runtime": 25399.3723, + "train_tokens_per_second": 8114.669 + }, + { + "epoch": 0.9120278330019881, + "grad_norm": 0.5876235650343216, + "learning_rate": 5.692601086145988e-06, + "loss": 0.37775001525878904, + "num_input_tokens_seen": 206655872, + "step": 1835, + "train_runtime": 25473.2908, + "train_tokens_per_second": 8112.649 + }, + { + "epoch": 0.9145129224652088, + "grad_norm": 2.4172905516795384, + "learning_rate": 5.6732661710960355e-06, + "loss": 0.36883745193481443, + "num_input_tokens_seen": 207202568, + "step": 1840, + "train_runtime": 25539.5779, + "train_tokens_per_second": 8112.999 + }, + { + "epoch": 0.9169980119284294, + "grad_norm": 0.6055102040624764, + "learning_rate": 5.653920996927449e-06, + "loss": 0.36173162460327146, + "num_input_tokens_seen": 207777648, + "step": 1845, + "train_runtime": 25616.9455, + "train_tokens_per_second": 8110.945 + }, + { + "epoch": 0.9194831013916501, + "grad_norm": 0.6490933435224526, + "learning_rate": 5.634565858418793e-06, + "loss": 0.34039714336395266, + "num_input_tokens_seen": 208319072, + "step": 1850, + "train_runtime": 25683.4467, + "train_tokens_per_second": 8111.025 + }, + { + "epoch": 0.9219681908548708, + "grad_norm": 0.6396971123670695, + "learning_rate": 5.615201050500473e-06, + "loss": 0.3785147428512573, + "num_input_tokens_seen": 208864528, + "step": 1855, + "train_runtime": 25762.3124, + "train_tokens_per_second": 8107.367 + }, + { + "epoch": 0.9244532803180915, + "grad_norm": 0.6030033026681769, + "learning_rate": 5.595826868250227e-06, + "loss": 0.3572536468505859, + "num_input_tokens_seen": 209417888, + "step": 1860, + "train_runtime": 25836.9927, + "train_tokens_per_second": 8105.351 + }, + { + "epoch": 0.9269383697813122, + "grad_norm": 1.5543801436852138, + "learning_rate": 5.576443606888642e-06, + "loss": 0.36751155853271483, + "num_input_tokens_seen": 209964888, + "step": 1865, + "train_runtime": 25913.7805, + "train_tokens_per_second": 8102.441 + }, + { + "epoch": 0.9294234592445328, + "grad_norm": 0.6169500594832261, + "learning_rate": 5.557051561774654e-06, + "loss": 0.3864218950271606, + "num_input_tokens_seen": 210502752, + "step": 1870, + "train_runtime": 25984.8225, + "train_tokens_per_second": 8100.989 + }, + { + "epoch": 0.9319085487077535, + "grad_norm": 0.8278711622199485, + "learning_rate": 5.537651028401037e-06, + "loss": 0.3558746337890625, + "num_input_tokens_seen": 211057776, + "step": 1875, + "train_runtime": 26055.4876, + "train_tokens_per_second": 8100.32 + }, + { + "epoch": 0.9343936381709742, + "grad_norm": 0.5907160213079181, + "learning_rate": 5.518242302389916e-06, + "loss": 0.359229588508606, + "num_input_tokens_seen": 211637880, + "step": 1880, + "train_runtime": 26127.8154, + "train_tokens_per_second": 8100.099 + }, + { + "epoch": 0.9368787276341949, + "grad_norm": 0.6755321695474246, + "learning_rate": 5.498825679488248e-06, + "loss": 0.37452447414398193, + "num_input_tokens_seen": 212215848, + "step": 1885, + "train_runtime": 26203.1682, + "train_tokens_per_second": 8098.862 + }, + { + "epoch": 0.9393638170974155, + "grad_norm": 0.6456676100993128, + "learning_rate": 5.479401455563325e-06, + "loss": 0.3868560791015625, + "num_input_tokens_seen": 212790280, + "step": 1890, + "train_runtime": 26285.2163, + "train_tokens_per_second": 8095.436 + }, + { + "epoch": 0.9418489065606361, + "grad_norm": 0.5803041422067553, + "learning_rate": 5.459969926598258e-06, + "loss": 0.36788253784179686, + "num_input_tokens_seen": 213387504, + "step": 1895, + "train_runtime": 26354.212, + "train_tokens_per_second": 8096.903 + }, + { + "epoch": 0.9443339960238568, + "grad_norm": 0.6154300943059695, + "learning_rate": 5.440531388687476e-06, + "loss": 0.3790719509124756, + "num_input_tokens_seen": 213953472, + "step": 1900, + "train_runtime": 26420.0944, + "train_tokens_per_second": 8098.134 + }, + { + "epoch": 0.9468190854870775, + "grad_norm": 0.5910251247683235, + "learning_rate": 5.421086138032207e-06, + "loss": 0.3631014347076416, + "num_input_tokens_seen": 214516584, + "step": 1905, + "train_runtime": 26489.5147, + "train_tokens_per_second": 8098.17 + }, + { + "epoch": 0.9493041749502982, + "grad_norm": 0.6672670920810775, + "learning_rate": 5.401634470935964e-06, + "loss": 0.3704709768295288, + "num_input_tokens_seen": 215066736, + "step": 1910, + "train_runtime": 26570.8482, + "train_tokens_per_second": 8094.086 + }, + { + "epoch": 0.9517892644135189, + "grad_norm": 0.6737920942821918, + "learning_rate": 5.382176683800039e-06, + "loss": 0.35127840042114256, + "num_input_tokens_seen": 215632776, + "step": 1915, + "train_runtime": 26633.6264, + "train_tokens_per_second": 8096.26 + }, + { + "epoch": 0.9542743538767395, + "grad_norm": 0.6307301076945065, + "learning_rate": 5.3627130731189715e-06, + "loss": 0.3761229753494263, + "num_input_tokens_seen": 216180176, + "step": 1920, + "train_runtime": 26705.7076, + "train_tokens_per_second": 8094.905 + }, + { + "epoch": 0.9567594433399602, + "grad_norm": 0.6481023910750464, + "learning_rate": 5.343243935476047e-06, + "loss": 0.35740041732788086, + "num_input_tokens_seen": 216737480, + "step": 1925, + "train_runtime": 26764.3695, + "train_tokens_per_second": 8097.986 + }, + { + "epoch": 0.9592445328031809, + "grad_norm": 0.905370333286984, + "learning_rate": 5.323769567538767e-06, + "loss": 0.3636675596237183, + "num_input_tokens_seen": 217313168, + "step": 1930, + "train_runtime": 26833.2208, + "train_tokens_per_second": 8098.661 + }, + { + "epoch": 0.9617296222664016, + "grad_norm": 0.6967039165703843, + "learning_rate": 5.304290266054331e-06, + "loss": 0.37630290985107423, + "num_input_tokens_seen": 217865872, + "step": 1935, + "train_runtime": 26891.5964, + "train_tokens_per_second": 8101.634 + }, + { + "epoch": 0.9642147117296223, + "grad_norm": 0.8154514723115697, + "learning_rate": 5.284806327845115e-06, + "loss": 0.37077414989471436, + "num_input_tokens_seen": 218402664, + "step": 1940, + "train_runtime": 26968.7914, + "train_tokens_per_second": 8098.348 + }, + { + "epoch": 0.9666998011928429, + "grad_norm": 0.6291794387386056, + "learning_rate": 5.265318049804149e-06, + "loss": 0.3803543567657471, + "num_input_tokens_seen": 218962152, + "step": 1945, + "train_runtime": 27042.5358, + "train_tokens_per_second": 8096.953 + }, + { + "epoch": 0.9691848906560636, + "grad_norm": 0.6705078379912762, + "learning_rate": 5.245825728890593e-06, + "loss": 0.3520726442337036, + "num_input_tokens_seen": 219502664, + "step": 1950, + "train_runtime": 27114.1339, + "train_tokens_per_second": 8095.507 + }, + { + "epoch": 0.9716699801192843, + "grad_norm": 0.7091745812039305, + "learning_rate": 5.2263296621252114e-06, + "loss": 0.3566716194152832, + "num_input_tokens_seen": 220059856, + "step": 1955, + "train_runtime": 27185.932, + "train_tokens_per_second": 8094.622 + }, + { + "epoch": 0.974155069582505, + "grad_norm": 0.6347854456155624, + "learning_rate": 5.2068301465858475e-06, + "loss": 0.37149648666381835, + "num_input_tokens_seen": 220611704, + "step": 1960, + "train_runtime": 27251.5972, + "train_tokens_per_second": 8095.368 + }, + { + "epoch": 0.9766401590457257, + "grad_norm": 0.6321021787921602, + "learning_rate": 5.187327479402895e-06, + "loss": 0.379772424697876, + "num_input_tokens_seen": 221165456, + "step": 1965, + "train_runtime": 27322.076, + "train_tokens_per_second": 8094.753 + }, + { + "epoch": 0.9791252485089463, + "grad_norm": 0.6851251954721802, + "learning_rate": 5.167821957754774e-06, + "loss": 0.3727846622467041, + "num_input_tokens_seen": 221738272, + "step": 1970, + "train_runtime": 27389.6635, + "train_tokens_per_second": 8095.692 + }, + { + "epoch": 0.981610337972167, + "grad_norm": 0.6383190626147188, + "learning_rate": 5.148313878863399e-06, + "loss": 0.3708191871643066, + "num_input_tokens_seen": 222281008, + "step": 1975, + "train_runtime": 27463.5519, + "train_tokens_per_second": 8093.673 + }, + { + "epoch": 0.9840954274353877, + "grad_norm": 0.6768802371987076, + "learning_rate": 5.128803539989652e-06, + "loss": 0.360191798210144, + "num_input_tokens_seen": 222844704, + "step": 1980, + "train_runtime": 27532.9692, + "train_tokens_per_second": 8093.74 + }, + { + "epoch": 0.9865805168986084, + "grad_norm": 0.649124040652731, + "learning_rate": 5.109291238428852e-06, + "loss": 0.3667256832122803, + "num_input_tokens_seen": 223387592, + "step": 1985, + "train_runtime": 27606.8018, + "train_tokens_per_second": 8091.759 + }, + { + "epoch": 0.989065606361829, + "grad_norm": 0.6326655768343108, + "learning_rate": 5.0897772715062235e-06, + "loss": 0.36615288257598877, + "num_input_tokens_seen": 223941728, + "step": 1990, + "train_runtime": 27675.9319, + "train_tokens_per_second": 8091.57 + }, + { + "epoch": 0.9915506958250497, + "grad_norm": 0.6170508955770921, + "learning_rate": 5.070261936572372e-06, + "loss": 0.35810768604278564, + "num_input_tokens_seen": 224492544, + "step": 1995, + "train_runtime": 27740.6607, + "train_tokens_per_second": 8092.545 + }, + { + "epoch": 0.9940357852882704, + "grad_norm": 0.602289801473698, + "learning_rate": 5.050745530998744e-06, + "loss": 0.3417465925216675, + "num_input_tokens_seen": 225033760, + "step": 2000, + "train_runtime": 27817.0549, + "train_tokens_per_second": 8089.777 + }, + { + "epoch": 0.9965208747514911, + "grad_norm": 0.7304680712034809, + "learning_rate": 5.031228352173102e-06, + "loss": 0.36499967575073244, + "num_input_tokens_seen": 225618616, + "step": 2005, + "train_runtime": 27882.8699, + "train_tokens_per_second": 8091.657 + }, + { + "epoch": 0.9990059642147118, + "grad_norm": 0.7100089813196016, + "learning_rate": 5.011710697494992e-06, + "loss": 0.35567431449890136, + "num_input_tokens_seen": 226160112, + "step": 2010, + "train_runtime": 27954.2297, + "train_tokens_per_second": 8090.372 + }, + { + "epoch": 1.0014910536779325, + "grad_norm": 0.7220342711104758, + "learning_rate": 4.992192864371208e-06, + "loss": 0.3508726119995117, + "num_input_tokens_seen": 226737064, + "step": 2015, + "train_runtime": 28022.8018, + "train_tokens_per_second": 8091.163 + }, + { + "epoch": 1.0039761431411531, + "grad_norm": 0.6316258903569337, + "learning_rate": 4.972675150211267e-06, + "loss": 0.31991915702819823, + "num_input_tokens_seen": 227316536, + "step": 2020, + "train_runtime": 28094.4464, + "train_tokens_per_second": 8091.156 + }, + { + "epoch": 1.0064612326043738, + "grad_norm": 0.6211320216915867, + "learning_rate": 4.953157852422869e-06, + "loss": 0.33379387855529785, + "num_input_tokens_seen": 227859184, + "step": 2025, + "train_runtime": 28160.738, + "train_tokens_per_second": 8091.378 + }, + { + "epoch": 1.0089463220675945, + "grad_norm": 0.767588613787823, + "learning_rate": 4.933641268407374e-06, + "loss": 0.3425900459289551, + "num_input_tokens_seen": 228400328, + "step": 2030, + "train_runtime": 28227.6141, + "train_tokens_per_second": 8091.379 + }, + { + "epoch": 1.0114314115308152, + "grad_norm": 0.6152211007519458, + "learning_rate": 4.914125695555264e-06, + "loss": 0.3301522731781006, + "num_input_tokens_seen": 228981912, + "step": 2035, + "train_runtime": 28294.637, + "train_tokens_per_second": 8092.767 + }, + { + "epoch": 1.0139165009940359, + "grad_norm": 0.6248840645533891, + "learning_rate": 4.894611431241612e-06, + "loss": 0.3237950801849365, + "num_input_tokens_seen": 229537504, + "step": 2040, + "train_runtime": 28369.1905, + "train_tokens_per_second": 8091.084 + }, + { + "epoch": 1.0164015904572565, + "grad_norm": 0.5479980565152459, + "learning_rate": 4.875098772821551e-06, + "loss": 0.3146378517150879, + "num_input_tokens_seen": 230138128, + "step": 2045, + "train_runtime": 28449.9416, + "train_tokens_per_second": 8089.23 + }, + { + "epoch": 1.0188866799204772, + "grad_norm": 0.646207788278426, + "learning_rate": 4.8555880176257445e-06, + "loss": 0.33398947715759275, + "num_input_tokens_seen": 230678136, + "step": 2050, + "train_runtime": 28509.5788, + "train_tokens_per_second": 8091.25 + }, + { + "epoch": 1.021371769383698, + "grad_norm": 0.572908153342372, + "learning_rate": 4.836079462955857e-06, + "loss": 0.3242894172668457, + "num_input_tokens_seen": 231204040, + "step": 2055, + "train_runtime": 28584.1306, + "train_tokens_per_second": 8088.545 + }, + { + "epoch": 1.0238568588469186, + "grad_norm": 0.6594371617737297, + "learning_rate": 4.81657340608002e-06, + "loss": 0.33183932304382324, + "num_input_tokens_seen": 231785888, + "step": 2060, + "train_runtime": 28657.3946, + "train_tokens_per_second": 8088.17 + }, + { + "epoch": 1.0263419483101393, + "grad_norm": 0.6141889733406628, + "learning_rate": 4.7970701442283045e-06, + "loss": 0.31254689693450927, + "num_input_tokens_seen": 232313496, + "step": 2065, + "train_runtime": 28723.097, + "train_tokens_per_second": 8088.038 + }, + { + "epoch": 1.02882703777336, + "grad_norm": 0.9103089897475846, + "learning_rate": 4.7775699745881885e-06, + "loss": 0.31564185619354246, + "num_input_tokens_seen": 232844408, + "step": 2070, + "train_runtime": 28788.5389, + "train_tokens_per_second": 8088.094 + }, + { + "epoch": 1.0313121272365806, + "grad_norm": 0.7483983432390716, + "learning_rate": 4.758073194300036e-06, + "loss": 0.32052605152130126, + "num_input_tokens_seen": 233389336, + "step": 2075, + "train_runtime": 28854.9254, + "train_tokens_per_second": 8088.371 + }, + { + "epoch": 1.0337972166998013, + "grad_norm": 0.6386677878485151, + "learning_rate": 4.738580100452562e-06, + "loss": 0.3223442554473877, + "num_input_tokens_seen": 233976072, + "step": 2080, + "train_runtime": 28925.8427, + "train_tokens_per_second": 8088.825 + }, + { + "epoch": 1.036282306163022, + "grad_norm": 0.6599612147692426, + "learning_rate": 4.719090990078309e-06, + "loss": 0.3202527046203613, + "num_input_tokens_seen": 234543496, + "step": 2085, + "train_runtime": 28990.7436, + "train_tokens_per_second": 8090.289 + }, + { + "epoch": 1.0387673956262427, + "grad_norm": 0.6789824733475135, + "learning_rate": 4.699606160149119e-06, + "loss": 0.3402568340301514, + "num_input_tokens_seen": 235092984, + "step": 2090, + "train_runtime": 29060.2989, + "train_tokens_per_second": 8089.834 + }, + { + "epoch": 1.0412524850894633, + "grad_norm": 0.6211248431195834, + "learning_rate": 4.680125907571611e-06, + "loss": 0.3191435098648071, + "num_input_tokens_seen": 235624312, + "step": 2095, + "train_runtime": 29125.8068, + "train_tokens_per_second": 8089.881 + }, + { + "epoch": 1.0437375745526838, + "grad_norm": 0.6531595574371873, + "learning_rate": 4.660650529182653e-06, + "loss": 0.3352505683898926, + "num_input_tokens_seen": 236188872, + "step": 2100, + "train_runtime": 29200.0952, + "train_tokens_per_second": 8088.634 + }, + { + "epoch": 1.0462226640159045, + "grad_norm": 0.636947726877122, + "learning_rate": 4.6411803217448425e-06, + "loss": 0.3199373722076416, + "num_input_tokens_seen": 236756720, + "step": 2105, + "train_runtime": 29269.1579, + "train_tokens_per_second": 8088.949 + }, + { + "epoch": 1.0487077534791251, + "grad_norm": 0.6442717467568068, + "learning_rate": 4.621715581941981e-06, + "loss": 0.32409305572509767, + "num_input_tokens_seen": 237330608, + "step": 2110, + "train_runtime": 29346.7346, + "train_tokens_per_second": 8087.121 + }, + { + "epoch": 1.0511928429423458, + "grad_norm": 0.6853790472708381, + "learning_rate": 4.602256606374558e-06, + "loss": 0.3354381561279297, + "num_input_tokens_seen": 237886696, + "step": 2115, + "train_runtime": 29416.4256, + "train_tokens_per_second": 8086.866 + }, + { + "epoch": 1.0536779324055665, + "grad_norm": 0.6126900427173739, + "learning_rate": 4.582803691555226e-06, + "loss": 0.3016113042831421, + "num_input_tokens_seen": 238463936, + "step": 2120, + "train_runtime": 29481.7313, + "train_tokens_per_second": 8088.532 + }, + { + "epoch": 1.0561630218687872, + "grad_norm": 0.6399196449668154, + "learning_rate": 4.563357133904288e-06, + "loss": 0.31397280693054197, + "num_input_tokens_seen": 239038720, + "step": 2125, + "train_runtime": 29540.4228, + "train_tokens_per_second": 8091.919 + }, + { + "epoch": 1.0586481113320079, + "grad_norm": 0.6469459363473768, + "learning_rate": 4.543917229745172e-06, + "loss": 0.3278268098831177, + "num_input_tokens_seen": 239595744, + "step": 2130, + "train_runtime": 29612.1449, + "train_tokens_per_second": 8091.131 + }, + { + "epoch": 1.0611332007952285, + "grad_norm": 0.5846807950952326, + "learning_rate": 4.524484275299925e-06, + "loss": 0.3242122888565063, + "num_input_tokens_seen": 240176144, + "step": 2135, + "train_runtime": 29680.0843, + "train_tokens_per_second": 8092.165 + }, + { + "epoch": 1.0636182902584492, + "grad_norm": 0.6187000739579367, + "learning_rate": 4.5050585666846975e-06, + "loss": 0.319428825378418, + "num_input_tokens_seen": 240776312, + "step": 2140, + "train_runtime": 29754.3293, + "train_tokens_per_second": 8092.144 + }, + { + "epoch": 1.06610337972167, + "grad_norm": 0.6142268871255283, + "learning_rate": 4.485640399905224e-06, + "loss": 0.32778201103210447, + "num_input_tokens_seen": 241324080, + "step": 2145, + "train_runtime": 29826.6771, + "train_tokens_per_second": 8090.881 + }, + { + "epoch": 1.0685884691848906, + "grad_norm": 0.6405587203034582, + "learning_rate": 4.466230070852322e-06, + "loss": 0.3494364023208618, + "num_input_tokens_seen": 241902248, + "step": 2150, + "train_runtime": 29888.6183, + "train_tokens_per_second": 8093.457 + }, + { + "epoch": 1.0710735586481113, + "grad_norm": 0.8151302431756986, + "learning_rate": 4.446827875297376e-06, + "loss": 0.3194560527801514, + "num_input_tokens_seen": 242478536, + "step": 2155, + "train_runtime": 29962.5325, + "train_tokens_per_second": 8092.725 + }, + { + "epoch": 1.073558648111332, + "grad_norm": 0.6115821306941586, + "learning_rate": 4.427434108887835e-06, + "loss": 0.3152233839035034, + "num_input_tokens_seen": 243054488, + "step": 2160, + "train_runtime": 30034.0332, + "train_tokens_per_second": 8092.636 + }, + { + "epoch": 1.0760437375745526, + "grad_norm": 0.61408241635913, + "learning_rate": 4.4080490671427075e-06, + "loss": 0.2981253147125244, + "num_input_tokens_seen": 243606416, + "step": 2165, + "train_runtime": 30106.6566, + "train_tokens_per_second": 8091.447 + }, + { + "epoch": 1.0785288270377733, + "grad_norm": 0.6666482668627652, + "learning_rate": 4.3886730454480546e-06, + "loss": 0.33526015281677246, + "num_input_tokens_seen": 244146392, + "step": 2170, + "train_runtime": 30182.4373, + "train_tokens_per_second": 8089.022 + }, + { + "epoch": 1.081013916500994, + "grad_norm": 0.6430409556136693, + "learning_rate": 4.3693063390524915e-06, + "loss": 0.3174752235412598, + "num_input_tokens_seen": 244692544, + "step": 2175, + "train_runtime": 30254.6422, + "train_tokens_per_second": 8087.769 + }, + { + "epoch": 1.0834990059642147, + "grad_norm": 0.652326856959876, + "learning_rate": 4.349949243062692e-06, + "loss": 0.3108466625213623, + "num_input_tokens_seen": 245252688, + "step": 2180, + "train_runtime": 30329.765, + "train_tokens_per_second": 8086.205 + }, + { + "epoch": 1.0859840954274353, + "grad_norm": 0.630003418633941, + "learning_rate": 4.330602052438882e-06, + "loss": 0.3311713933944702, + "num_input_tokens_seen": 245832416, + "step": 2185, + "train_runtime": 30403.6988, + "train_tokens_per_second": 8085.609 + }, + { + "epoch": 1.088469184890656, + "grad_norm": 0.6366408624029698, + "learning_rate": 4.311265061990358e-06, + "loss": 0.3274925708770752, + "num_input_tokens_seen": 246416552, + "step": 2190, + "train_runtime": 30466.8745, + "train_tokens_per_second": 8088.015 + }, + { + "epoch": 1.0909542743538767, + "grad_norm": 0.5900715901891197, + "learning_rate": 4.291938566370981e-06, + "loss": 0.3136040210723877, + "num_input_tokens_seen": 246969200, + "step": 2195, + "train_runtime": 30535.7817, + "train_tokens_per_second": 8087.862 + }, + { + "epoch": 1.0934393638170974, + "grad_norm": 0.6012730805946693, + "learning_rate": 4.272622860074698e-06, + "loss": 0.3395073413848877, + "num_input_tokens_seen": 247478144, + "step": 2200, + "train_runtime": 30611.4595, + "train_tokens_per_second": 8084.493 + }, + { + "epoch": 1.095924453280318, + "grad_norm": 0.6560838523717596, + "learning_rate": 4.2533182374310495e-06, + "loss": 0.29720206260681153, + "num_input_tokens_seen": 248051264, + "step": 2205, + "train_runtime": 30684.982, + "train_tokens_per_second": 8083.8 + }, + { + "epoch": 1.0984095427435387, + "grad_norm": 0.6825555931511044, + "learning_rate": 4.234024992600687e-06, + "loss": 0.3173992156982422, + "num_input_tokens_seen": 248590704, + "step": 2210, + "train_runtime": 30758.5544, + "train_tokens_per_second": 8082.002 + }, + { + "epoch": 1.1008946322067594, + "grad_norm": 0.6098839490760868, + "learning_rate": 4.2147434195708804e-06, + "loss": 0.31537957191467286, + "num_input_tokens_seen": 249109680, + "step": 2215, + "train_runtime": 30825.2679, + "train_tokens_per_second": 8081.347 + }, + { + "epoch": 1.10337972166998, + "grad_norm": 0.6386612958474456, + "learning_rate": 4.195473812151058e-06, + "loss": 0.3233582258224487, + "num_input_tokens_seen": 249704216, + "step": 2220, + "train_runtime": 30892.3844, + "train_tokens_per_second": 8083.035 + }, + { + "epoch": 1.1058648111332008, + "grad_norm": 0.636697916752228, + "learning_rate": 4.1762164639683095e-06, + "loss": 0.31053619384765624, + "num_input_tokens_seen": 250252552, + "step": 2225, + "train_runtime": 30964.4613, + "train_tokens_per_second": 8081.928 + }, + { + "epoch": 1.1083499005964215, + "grad_norm": 0.6317469311704326, + "learning_rate": 4.156971668462925e-06, + "loss": 0.3271548509597778, + "num_input_tokens_seen": 250822296, + "step": 2230, + "train_runtime": 31041.7921, + "train_tokens_per_second": 8080.149 + }, + { + "epoch": 1.1108349900596421, + "grad_norm": 0.6389020439902426, + "learning_rate": 4.137739718883917e-06, + "loss": 0.32387576103210447, + "num_input_tokens_seen": 251378216, + "step": 2235, + "train_runtime": 31108.9605, + "train_tokens_per_second": 8080.573 + }, + { + "epoch": 1.1133200795228628, + "grad_norm": 0.6261621892785184, + "learning_rate": 4.118520908284554e-06, + "loss": 0.32201275825500486, + "num_input_tokens_seen": 251947928, + "step": 2240, + "train_runtime": 31177.9588, + "train_tokens_per_second": 8080.963 + }, + { + "epoch": 1.1158051689860835, + "grad_norm": 0.6341116492609907, + "learning_rate": 4.099315529517896e-06, + "loss": 0.3337748050689697, + "num_input_tokens_seen": 252521112, + "step": 2245, + "train_runtime": 31249.2317, + "train_tokens_per_second": 8080.874 + }, + { + "epoch": 1.1182902584493042, + "grad_norm": 0.5755586278792829, + "learning_rate": 4.080123875232332e-06, + "loss": 0.3051743984222412, + "num_input_tokens_seen": 253104688, + "step": 2250, + "train_runtime": 31325.9284, + "train_tokens_per_second": 8079.719 + }, + { + "epoch": 1.1207753479125249, + "grad_norm": 0.5744938963501268, + "learning_rate": 4.060946237867115e-06, + "loss": 0.3200753450393677, + "num_input_tokens_seen": 253677416, + "step": 2255, + "train_runtime": 31395.7294, + "train_tokens_per_second": 8079.998 + }, + { + "epoch": 1.1232604373757455, + "grad_norm": 0.6578306620624749, + "learning_rate": 4.041782909647915e-06, + "loss": 0.3317903518676758, + "num_input_tokens_seen": 254282824, + "step": 2260, + "train_runtime": 31471.9345, + "train_tokens_per_second": 8079.669 + }, + { + "epoch": 1.1257455268389662, + "grad_norm": 0.6531894627559202, + "learning_rate": 4.022634182582359e-06, + "loss": 0.3205811262130737, + "num_input_tokens_seen": 254855648, + "step": 2265, + "train_runtime": 31537.6213, + "train_tokens_per_second": 8081.004 + }, + { + "epoch": 1.128230616302187, + "grad_norm": 0.5834929249713257, + "learning_rate": 4.003500348455586e-06, + "loss": 0.3306148529052734, + "num_input_tokens_seen": 255427096, + "step": 2270, + "train_runtime": 31609.822, + "train_tokens_per_second": 8080.624 + }, + { + "epoch": 1.1307157057654076, + "grad_norm": 0.6847755111833836, + "learning_rate": 3.984381698825797e-06, + "loss": 0.3347146987915039, + "num_input_tokens_seen": 255990328, + "step": 2275, + "train_runtime": 31679.2751, + "train_tokens_per_second": 8080.688 + }, + { + "epoch": 1.1332007952286283, + "grad_norm": 0.6833813157942104, + "learning_rate": 3.965278525019818e-06, + "loss": 0.3315311908721924, + "num_input_tokens_seen": 256536512, + "step": 2280, + "train_runtime": 31747.5821, + "train_tokens_per_second": 8080.506 + }, + { + "epoch": 1.135685884691849, + "grad_norm": 0.6990559836523956, + "learning_rate": 3.94619111812865e-06, + "loss": 0.34252352714538575, + "num_input_tokens_seen": 257097856, + "step": 2285, + "train_runtime": 31821.5405, + "train_tokens_per_second": 8079.365 + }, + { + "epoch": 1.1381709741550696, + "grad_norm": 0.7771992785218264, + "learning_rate": 3.927119769003045e-06, + "loss": 0.3083313226699829, + "num_input_tokens_seen": 257661360, + "step": 2290, + "train_runtime": 31897.236, + "train_tokens_per_second": 8077.859 + }, + { + "epoch": 1.1406560636182903, + "grad_norm": 0.6231858999571254, + "learning_rate": 3.908064768249072e-06, + "loss": 0.3196516990661621, + "num_input_tokens_seen": 258243568, + "step": 2295, + "train_runtime": 31972.5262, + "train_tokens_per_second": 8077.046 + }, + { + "epoch": 1.143141153081511, + "grad_norm": 0.5968720642356495, + "learning_rate": 3.889026406223679e-06, + "loss": 0.3293436050415039, + "num_input_tokens_seen": 258850704, + "step": 2300, + "train_runtime": 32027.1608, + "train_tokens_per_second": 8082.225 + }, + { + "epoch": 1.1456262425447317, + "grad_norm": 0.7239474168849465, + "learning_rate": 3.870004973030283e-06, + "loss": 0.32830061912536623, + "num_input_tokens_seen": 259420696, + "step": 2305, + "train_runtime": 32093.8142, + "train_tokens_per_second": 8083.199 + }, + { + "epoch": 1.1481113320079523, + "grad_norm": 0.595752681335604, + "learning_rate": 3.8510007585143375e-06, + "loss": 0.3143561124801636, + "num_input_tokens_seen": 260014448, + "step": 2310, + "train_runtime": 32161.1403, + "train_tokens_per_second": 8084.74 + }, + { + "epoch": 1.150596421471173, + "grad_norm": 0.6162562980204516, + "learning_rate": 3.8320140522589215e-06, + "loss": 0.33693206310272217, + "num_input_tokens_seen": 260587896, + "step": 2315, + "train_runtime": 32226.074, + "train_tokens_per_second": 8086.244 + }, + { + "epoch": 1.1530815109343937, + "grad_norm": 0.6143498128120426, + "learning_rate": 3.8130451435803282e-06, + "loss": 0.33428802490234377, + "num_input_tokens_seen": 261141896, + "step": 2320, + "train_runtime": 32293.1201, + "train_tokens_per_second": 8086.611 + }, + { + "epoch": 1.1555666003976144, + "grad_norm": 0.6929894402647042, + "learning_rate": 3.794094321523652e-06, + "loss": 0.3260087013244629, + "num_input_tokens_seen": 261702968, + "step": 2325, + "train_runtime": 32371.474, + "train_tokens_per_second": 8084.37 + }, + { + "epoch": 1.158051689860835, + "grad_norm": 0.6497606792439974, + "learning_rate": 3.7751618748583873e-06, + "loss": 0.345662522315979, + "num_input_tokens_seen": 262272216, + "step": 2330, + "train_runtime": 32440.2344, + "train_tokens_per_second": 8084.782 + }, + { + "epoch": 1.1605367793240557, + "grad_norm": 0.6613596197721965, + "learning_rate": 3.756248092074028e-06, + "loss": 0.3290838718414307, + "num_input_tokens_seen": 262827032, + "step": 2335, + "train_runtime": 32518.1051, + "train_tokens_per_second": 8082.483 + }, + { + "epoch": 1.1630218687872764, + "grad_norm": 0.7094004518614818, + "learning_rate": 3.737353261375668e-06, + "loss": 0.3275127649307251, + "num_input_tokens_seen": 263329496, + "step": 2340, + "train_runtime": 32588.7633, + "train_tokens_per_second": 8080.377 + }, + { + "epoch": 1.165506958250497, + "grad_norm": 0.7002145901744623, + "learning_rate": 3.718477670679616e-06, + "loss": 0.3150747299194336, + "num_input_tokens_seen": 263937296, + "step": 2345, + "train_runtime": 32655.9487, + "train_tokens_per_second": 8082.365 + }, + { + "epoch": 1.1679920477137178, + "grad_norm": 0.6635493599794974, + "learning_rate": 3.699621607609003e-06, + "loss": 0.3206862688064575, + "num_input_tokens_seen": 264506672, + "step": 2350, + "train_runtime": 32728.8067, + "train_tokens_per_second": 8081.77 + }, + { + "epoch": 1.1704771371769385, + "grad_norm": 0.5871404302500475, + "learning_rate": 3.6807853594894007e-06, + "loss": 0.3272248268127441, + "num_input_tokens_seen": 265098816, + "step": 2355, + "train_runtime": 32799.8193, + "train_tokens_per_second": 8082.326 + }, + { + "epoch": 1.1729622266401591, + "grad_norm": 0.6703183065801952, + "learning_rate": 3.661969213344446e-06, + "loss": 0.32170169353485106, + "num_input_tokens_seen": 265640160, + "step": 2360, + "train_runtime": 32866.9165, + "train_tokens_per_second": 8082.296 + }, + { + "epoch": 1.1754473161033796, + "grad_norm": 0.602513563183077, + "learning_rate": 3.6431734558914633e-06, + "loss": 0.3153726577758789, + "num_input_tokens_seen": 266192824, + "step": 2365, + "train_runtime": 32930.0122, + "train_tokens_per_second": 8083.593 + }, + { + "epoch": 1.1779324055666005, + "grad_norm": 0.5856326412574623, + "learning_rate": 3.6243983735370967e-06, + "loss": 0.3210511445999146, + "num_input_tokens_seen": 266753208, + "step": 2370, + "train_runtime": 32998.7329, + "train_tokens_per_second": 8083.741 + }, + { + "epoch": 1.180417495029821, + "grad_norm": 0.6463277530426302, + "learning_rate": 3.6056442523729497e-06, + "loss": 0.33164658546447756, + "num_input_tokens_seen": 267338872, + "step": 2375, + "train_runtime": 33058.3541, + "train_tokens_per_second": 8086.878 + }, + { + "epoch": 1.1829025844930419, + "grad_norm": 0.5771305337108241, + "learning_rate": 3.5869113781712196e-06, + "loss": 0.3182060956954956, + "num_input_tokens_seen": 267915232, + "step": 2380, + "train_runtime": 33130.367, + "train_tokens_per_second": 8086.697 + }, + { + "epoch": 1.1853876739562623, + "grad_norm": 0.628003990194654, + "learning_rate": 3.568200036380348e-06, + "loss": 0.3286311149597168, + "num_input_tokens_seen": 268464608, + "step": 2385, + "train_runtime": 33201.9465, + "train_tokens_per_second": 8085.809 + }, + { + "epoch": 1.1878727634194832, + "grad_norm": 0.7480220207953597, + "learning_rate": 3.549510512120668e-06, + "loss": 0.33164024353027344, + "num_input_tokens_seen": 269014224, + "step": 2390, + "train_runtime": 33276.8319, + "train_tokens_per_second": 8084.13 + }, + { + "epoch": 1.1903578528827037, + "grad_norm": 0.730153837622979, + "learning_rate": 3.5308430901800627e-06, + "loss": 0.31567792892456054, + "num_input_tokens_seen": 269564560, + "step": 2395, + "train_runtime": 33352.8577, + "train_tokens_per_second": 8082.203 + }, + { + "epoch": 1.1928429423459244, + "grad_norm": 0.5843494247611316, + "learning_rate": 3.512198055009619e-06, + "loss": 0.3049311637878418, + "num_input_tokens_seen": 270139240, + "step": 2400, + "train_runtime": 33420.0861, + "train_tokens_per_second": 8083.14 + }, + { + "epoch": 1.195328031809145, + "grad_norm": 0.5793905581100733, + "learning_rate": 3.493575690719304e-06, + "loss": 0.32205421924591066, + "num_input_tokens_seen": 270703424, + "step": 2405, + "train_runtime": 33482.6524, + "train_tokens_per_second": 8084.886 + }, + { + "epoch": 1.1978131212723657, + "grad_norm": 0.6018827521593433, + "learning_rate": 3.4749762810736254e-06, + "loss": 0.31123852729797363, + "num_input_tokens_seen": 271286064, + "step": 2410, + "train_runtime": 33552.1596, + "train_tokens_per_second": 8085.502 + }, + { + "epoch": 1.2002982107355864, + "grad_norm": 0.6808928666697777, + "learning_rate": 3.4564001094873132e-06, + "loss": 0.32583189010620117, + "num_input_tokens_seen": 271807856, + "step": 2415, + "train_runtime": 33620.7165, + "train_tokens_per_second": 8084.535 + }, + { + "epoch": 1.202783300198807, + "grad_norm": 0.6180584776366593, + "learning_rate": 3.4378474590210016e-06, + "loss": 0.32810773849487307, + "num_input_tokens_seen": 272360400, + "step": 2420, + "train_runtime": 33690.7748, + "train_tokens_per_second": 8084.124 + }, + { + "epoch": 1.2052683896620278, + "grad_norm": 0.6351757893695958, + "learning_rate": 3.4193186123769083e-06, + "loss": 0.34054317474365237, + "num_input_tokens_seen": 272917656, + "step": 2425, + "train_runtime": 33753.6709, + "train_tokens_per_second": 8085.57 + }, + { + "epoch": 1.2077534791252484, + "grad_norm": 0.6652341868779129, + "learning_rate": 3.400813851894539e-06, + "loss": 0.32586188316345216, + "num_input_tokens_seen": 273469104, + "step": 2430, + "train_runtime": 33826.367, + "train_tokens_per_second": 8084.495 + }, + { + "epoch": 1.2102385685884691, + "grad_norm": 0.6218837895255407, + "learning_rate": 3.382333459546374e-06, + "loss": 0.3159704923629761, + "num_input_tokens_seen": 274021080, + "step": 2435, + "train_runtime": 33885.609, + "train_tokens_per_second": 8086.651 + }, + { + "epoch": 1.2127236580516898, + "grad_norm": 0.637896223267975, + "learning_rate": 3.363877716933578e-06, + "loss": 0.33318562507629396, + "num_input_tokens_seen": 274611008, + "step": 2440, + "train_runtime": 33958.6769, + "train_tokens_per_second": 8086.623 + }, + { + "epoch": 1.2152087475149105, + "grad_norm": 0.6035773864104889, + "learning_rate": 3.345446905281706e-06, + "loss": 0.3152783393859863, + "num_input_tokens_seen": 275218032, + "step": 2445, + "train_runtime": 34019.1894, + "train_tokens_per_second": 8090.082 + }, + { + "epoch": 1.2176938369781312, + "grad_norm": 0.6548051481387002, + "learning_rate": 3.3270413054364213e-06, + "loss": 0.3222240447998047, + "num_input_tokens_seen": 275800712, + "step": 2450, + "train_runtime": 34089.1439, + "train_tokens_per_second": 8090.573 + }, + { + "epoch": 1.2201789264413518, + "grad_norm": 0.6336896903624539, + "learning_rate": 3.3086611978592094e-06, + "loss": 0.3394973039627075, + "num_input_tokens_seen": 276361504, + "step": 2455, + "train_runtime": 34146.9155, + "train_tokens_per_second": 8093.308 + }, + { + "epoch": 1.2226640159045725, + "grad_norm": 0.668188510482118, + "learning_rate": 3.2903068626231128e-06, + "loss": 0.32337279319763185, + "num_input_tokens_seen": 276961304, + "step": 2460, + "train_runtime": 34216.8666, + "train_tokens_per_second": 8094.292 + }, + { + "epoch": 1.2251491053677932, + "grad_norm": 0.6989543846259106, + "learning_rate": 3.2719785794084593e-06, + "loss": 0.3227228164672852, + "num_input_tokens_seen": 277507944, + "step": 2465, + "train_runtime": 34288.6046, + "train_tokens_per_second": 8093.299 + }, + { + "epoch": 1.2276341948310139, + "grad_norm": 0.5758534154829603, + "learning_rate": 3.253676627498597e-06, + "loss": 0.32222661972045896, + "num_input_tokens_seen": 278107072, + "step": 2470, + "train_runtime": 34360.6406, + "train_tokens_per_second": 8093.769 + }, + { + "epoch": 1.2301192842942346, + "grad_norm": 0.6351372561755276, + "learning_rate": 3.2354012857756457e-06, + "loss": 0.31956522464752196, + "num_input_tokens_seen": 278666560, + "step": 2475, + "train_runtime": 34436.457, + "train_tokens_per_second": 8092.196 + }, + { + "epoch": 1.2326043737574552, + "grad_norm": 0.595795778235017, + "learning_rate": 3.217152832716243e-06, + "loss": 0.32536227703094484, + "num_input_tokens_seen": 279215880, + "step": 2480, + "train_runtime": 34507.4318, + "train_tokens_per_second": 8091.471 + }, + { + "epoch": 1.235089463220676, + "grad_norm": 0.6324007197929544, + "learning_rate": 3.198931546387297e-06, + "loss": 0.3594836235046387, + "num_input_tokens_seen": 279792376, + "step": 2485, + "train_runtime": 34571.6799, + "train_tokens_per_second": 8093.109 + }, + { + "epoch": 1.2375745526838966, + "grad_norm": 0.6465568636310962, + "learning_rate": 3.180737704441762e-06, + "loss": 0.32520527839660646, + "num_input_tokens_seen": 280344912, + "step": 2490, + "train_runtime": 34635.6455, + "train_tokens_per_second": 8094.115 + }, + { + "epoch": 1.2400596421471173, + "grad_norm": 0.6772821809022134, + "learning_rate": 3.162571584114389e-06, + "loss": 0.34560179710388184, + "num_input_tokens_seen": 280915880, + "step": 2495, + "train_runtime": 34713.8145, + "train_tokens_per_second": 8092.337 + }, + { + "epoch": 1.242544731610338, + "grad_norm": 0.6488056406494763, + "learning_rate": 3.1444334622175197e-06, + "loss": 0.3284440994262695, + "num_input_tokens_seen": 281467400, + "step": 2500, + "train_runtime": 34784.8153, + "train_tokens_per_second": 8091.674 + }, + { + "epoch": 1.2450298210735586, + "grad_norm": 0.5418898831725447, + "learning_rate": 3.126323615136857e-06, + "loss": 0.3169837951660156, + "num_input_tokens_seen": 282051512, + "step": 2505, + "train_runtime": 34854.7249, + "train_tokens_per_second": 8092.203 + }, + { + "epoch": 1.2475149105367793, + "grad_norm": 0.6087455796213493, + "learning_rate": 3.1082423188272583e-06, + "loss": 0.33939690589904786, + "num_input_tokens_seen": 282629440, + "step": 2510, + "train_runtime": 34920.7332, + "train_tokens_per_second": 8093.457 + }, + { + "epoch": 1.25, + "grad_norm": 0.6124826130804227, + "learning_rate": 3.090189848808527e-06, + "loss": 0.30511353015899656, + "num_input_tokens_seen": 283204528, + "step": 2515, + "train_runtime": 34981.7285, + "train_tokens_per_second": 8095.784 + }, + { + "epoch": 1.2524850894632207, + "grad_norm": 0.5567122299549848, + "learning_rate": 3.0721664801612195e-06, + "loss": 0.32858710289001464, + "num_input_tokens_seen": 283801304, + "step": 2520, + "train_runtime": 35052.067, + "train_tokens_per_second": 8096.564 + }, + { + "epoch": 1.2549701789264414, + "grad_norm": 0.6343492277139788, + "learning_rate": 3.0541724875224467e-06, + "loss": 0.3222800731658936, + "num_input_tokens_seen": 284358344, + "step": 2525, + "train_runtime": 35123.8774, + "train_tokens_per_second": 8095.87 + }, + { + "epoch": 1.257455268389662, + "grad_norm": 0.6361025578248685, + "learning_rate": 3.036208145081694e-06, + "loss": 0.3197443962097168, + "num_input_tokens_seen": 284941192, + "step": 2530, + "train_runtime": 35205.8222, + "train_tokens_per_second": 8093.582 + }, + { + "epoch": 1.2599403578528827, + "grad_norm": 0.6320358293731507, + "learning_rate": 3.018273726576642e-06, + "loss": 0.33939423561096194, + "num_input_tokens_seen": 285494600, + "step": 2535, + "train_runtime": 35279.0001, + "train_tokens_per_second": 8092.48 + }, + { + "epoch": 1.2624254473161034, + "grad_norm": 0.6146055270104064, + "learning_rate": 3.000369505288997e-06, + "loss": 0.32036757469177246, + "num_input_tokens_seen": 286077960, + "step": 2540, + "train_runtime": 35349.9295, + "train_tokens_per_second": 8092.745 + }, + { + "epoch": 1.264910536779324, + "grad_norm": 0.6392388893453828, + "learning_rate": 2.9824957540403237e-06, + "loss": 0.3130642414093018, + "num_input_tokens_seen": 286629640, + "step": 2545, + "train_runtime": 35416.6842, + "train_tokens_per_second": 8093.068 + }, + { + "epoch": 1.2673956262425448, + "grad_norm": 0.669561774097629, + "learning_rate": 2.9646527451878894e-06, + "loss": 0.3235614538192749, + "num_input_tokens_seen": 287192416, + "step": 2550, + "train_runtime": 35479.2193, + "train_tokens_per_second": 8094.666 + }, + { + "epoch": 1.2698807157057654, + "grad_norm": 0.6700874935201345, + "learning_rate": 2.9468407506205123e-06, + "loss": 0.33912458419799807, + "num_input_tokens_seen": 287752528, + "step": 2555, + "train_runtime": 35546.1782, + "train_tokens_per_second": 8095.175 + }, + { + "epoch": 1.2723658051689861, + "grad_norm": 0.6277721550332024, + "learning_rate": 2.9290600417544224e-06, + "loss": 0.3259101867675781, + "num_input_tokens_seen": 288329024, + "step": 2560, + "train_runtime": 35608.4572, + "train_tokens_per_second": 8097.206 + }, + { + "epoch": 1.2748508946322068, + "grad_norm": 0.6053800812688522, + "learning_rate": 2.911310889529124e-06, + "loss": 0.319730281829834, + "num_input_tokens_seen": 288906296, + "step": 2565, + "train_runtime": 35667.4207, + "train_tokens_per_second": 8100.005 + }, + { + "epoch": 1.2773359840954275, + "grad_norm": 0.625190928142748, + "learning_rate": 2.8935935644032627e-06, + "loss": 0.32789390087127684, + "num_input_tokens_seen": 289479392, + "step": 2570, + "train_runtime": 35738.4492, + "train_tokens_per_second": 8099.943 + }, + { + "epoch": 1.2798210735586482, + "grad_norm": 0.6318225132472286, + "learning_rate": 2.875908336350511e-06, + "loss": 0.3394835710525513, + "num_input_tokens_seen": 290056952, + "step": 2575, + "train_runtime": 35809.8943, + "train_tokens_per_second": 8099.911 + }, + { + "epoch": 1.2823061630218688, + "grad_norm": 0.6215402620707905, + "learning_rate": 2.8582554748554515e-06, + "loss": 0.30529584884643557, + "num_input_tokens_seen": 290589240, + "step": 2580, + "train_runtime": 35885.7677, + "train_tokens_per_second": 8097.618 + }, + { + "epoch": 1.2847912524850895, + "grad_norm": 0.6317983547445536, + "learning_rate": 2.8406352489094698e-06, + "loss": 0.3314957618713379, + "num_input_tokens_seen": 291131192, + "step": 2585, + "train_runtime": 35954.3698, + "train_tokens_per_second": 8097.241 + }, + { + "epoch": 1.2872763419483102, + "grad_norm": 0.6268849807663828, + "learning_rate": 2.823047927006656e-06, + "loss": 0.3251540422439575, + "num_input_tokens_seen": 291686600, + "step": 2590, + "train_runtime": 36015.8122, + "train_tokens_per_second": 8098.848 + }, + { + "epoch": 1.2897614314115309, + "grad_norm": 0.6424720638305138, + "learning_rate": 2.805493777139715e-06, + "loss": 0.3126816749572754, + "num_input_tokens_seen": 292243120, + "step": 2595, + "train_runtime": 36078.3501, + "train_tokens_per_second": 8100.235 + }, + { + "epoch": 1.2922465208747516, + "grad_norm": 0.5707651828667896, + "learning_rate": 2.7879730667958816e-06, + "loss": 0.318155574798584, + "num_input_tokens_seen": 292821864, + "step": 2600, + "train_runtime": 36142.6453, + "train_tokens_per_second": 8101.838 + }, + { + "epoch": 1.2947316103379722, + "grad_norm": 0.5918765452260516, + "learning_rate": 2.770486062952842e-06, + "loss": 0.32808628082275393, + "num_input_tokens_seen": 293385056, + "step": 2605, + "train_runtime": 36212.7231, + "train_tokens_per_second": 8101.712 + }, + { + "epoch": 1.297216699801193, + "grad_norm": 0.6188612607564651, + "learning_rate": 2.7530330320746724e-06, + "loss": 0.33396780490875244, + "num_input_tokens_seen": 293970056, + "step": 2610, + "train_runtime": 36279.688, + "train_tokens_per_second": 8102.883 + }, + { + "epoch": 1.2997017892644136, + "grad_norm": 0.604409390562338, + "learning_rate": 2.7356142401077723e-06, + "loss": 0.2983008146286011, + "num_input_tokens_seen": 294551352, + "step": 2615, + "train_runtime": 36354.9121, + "train_tokens_per_second": 8102.106 + }, + { + "epoch": 1.302186878727634, + "grad_norm": 0.6308342727707753, + "learning_rate": 2.718229952476816e-06, + "loss": 0.3142648220062256, + "num_input_tokens_seen": 295122520, + "step": 2620, + "train_runtime": 36423.1788, + "train_tokens_per_second": 8102.602 + }, + { + "epoch": 1.304671968190855, + "grad_norm": 0.6294211926843355, + "learning_rate": 2.7008804340807017e-06, + "loss": 0.33000755310058594, + "num_input_tokens_seen": 295672928, + "step": 2625, + "train_runtime": 36495.322, + "train_tokens_per_second": 8101.667 + }, + { + "epoch": 1.3071570576540754, + "grad_norm": 0.5584334023404497, + "learning_rate": 2.683565949288524e-06, + "loss": 0.3415862798690796, + "num_input_tokens_seen": 296236992, + "step": 2630, + "train_runtime": 36571.9987, + "train_tokens_per_second": 8100.104 + }, + { + "epoch": 1.3096421471172963, + "grad_norm": 0.74724042191735, + "learning_rate": 2.6662867619355415e-06, + "loss": 0.32191967964172363, + "num_input_tokens_seen": 296779688, + "step": 2635, + "train_runtime": 36630.3938, + "train_tokens_per_second": 8102.006 + }, + { + "epoch": 1.3121272365805168, + "grad_norm": 0.591378818689317, + "learning_rate": 2.6490431353191547e-06, + "loss": 0.3372189521789551, + "num_input_tokens_seen": 297332984, + "step": 2640, + "train_runtime": 36701.3486, + "train_tokens_per_second": 8101.418 + }, + { + "epoch": 1.3146123260437377, + "grad_norm": 0.6342795071942039, + "learning_rate": 2.6318353321948966e-06, + "loss": 0.3189488649368286, + "num_input_tokens_seen": 297894552, + "step": 2645, + "train_runtime": 36770.1677, + "train_tokens_per_second": 8101.528 + }, + { + "epoch": 1.3170974155069581, + "grad_norm": 0.6098561735055383, + "learning_rate": 2.614663614772426e-06, + "loss": 0.30476808547973633, + "num_input_tokens_seen": 298479264, + "step": 2650, + "train_runtime": 36845.1921, + "train_tokens_per_second": 8100.901 + }, + { + "epoch": 1.319582504970179, + "grad_norm": 0.6234581715995534, + "learning_rate": 2.5975282447115353e-06, + "loss": 0.3280471324920654, + "num_input_tokens_seen": 299022280, + "step": 2655, + "train_runtime": 36908.6859, + "train_tokens_per_second": 8101.678 + }, + { + "epoch": 1.3220675944333995, + "grad_norm": 0.655004841687211, + "learning_rate": 2.580429483118161e-06, + "loss": 0.3062290191650391, + "num_input_tokens_seen": 299560984, + "step": 2660, + "train_runtime": 36978.04, + "train_tokens_per_second": 8101.051 + }, + { + "epoch": 1.3245526838966204, + "grad_norm": 0.6393687328102885, + "learning_rate": 2.5633675905404064e-06, + "loss": 0.3150776386260986, + "num_input_tokens_seen": 300152320, + "step": 2665, + "train_runtime": 37047.0582, + "train_tokens_per_second": 8101.921 + }, + { + "epoch": 1.3270377733598409, + "grad_norm": 0.6908188208617738, + "learning_rate": 2.54634282696457e-06, + "loss": 0.3288137435913086, + "num_input_tokens_seen": 300719784, + "step": 2670, + "train_runtime": 37118.2652, + "train_tokens_per_second": 8101.666 + }, + { + "epoch": 1.3295228628230618, + "grad_norm": 0.7164670265290806, + "learning_rate": 2.5293554518111862e-06, + "loss": 0.3155046463012695, + "num_input_tokens_seen": 301255880, + "step": 2675, + "train_runtime": 37176.1967, + "train_tokens_per_second": 8103.462 + }, + { + "epoch": 1.3320079522862822, + "grad_norm": 0.6319171746318435, + "learning_rate": 2.5124057239310705e-06, + "loss": 0.30266408920288085, + "num_input_tokens_seen": 301854936, + "step": 2680, + "train_runtime": 37246.969, + "train_tokens_per_second": 8104.148 + }, + { + "epoch": 1.3344930417495031, + "grad_norm": 0.6315527652604963, + "learning_rate": 2.4954939016013725e-06, + "loss": 0.3076614856719971, + "num_input_tokens_seen": 302417560, + "step": 2685, + "train_runtime": 37312.9776, + "train_tokens_per_second": 8104.889 + }, + { + "epoch": 1.3369781312127236, + "grad_norm": 0.5998438469712011, + "learning_rate": 2.478620242521646e-06, + "loss": 0.3289927005767822, + "num_input_tokens_seen": 302977032, + "step": 2690, + "train_runtime": 37388.1948, + "train_tokens_per_second": 8103.548 + }, + { + "epoch": 1.3394632206759443, + "grad_norm": 0.6543357588939832, + "learning_rate": 2.46178500380992e-06, + "loss": 0.3116851806640625, + "num_input_tokens_seen": 303551736, + "step": 2695, + "train_runtime": 37457.6066, + "train_tokens_per_second": 8103.874 + }, + { + "epoch": 1.341948310139165, + "grad_norm": 0.6427896034534485, + "learning_rate": 2.4449884419987795e-06, + "loss": 0.3331099271774292, + "num_input_tokens_seen": 304119136, + "step": 2700, + "train_runtime": 37531.3612, + "train_tokens_per_second": 8103.067 + }, + { + "epoch": 1.3444333996023856, + "grad_norm": 0.7718855194513118, + "learning_rate": 2.4282308130314564e-06, + "loss": 0.3319978713989258, + "num_input_tokens_seen": 304663536, + "step": 2705, + "train_runtime": 37605.0785, + "train_tokens_per_second": 8101.659 + }, + { + "epoch": 1.3469184890656063, + "grad_norm": 0.6607281365182268, + "learning_rate": 2.411512372257932e-06, + "loss": 0.3363766670227051, + "num_input_tokens_seen": 305232432, + "step": 2710, + "train_runtime": 37678.571, + "train_tokens_per_second": 8100.956 + }, + { + "epoch": 1.349403578528827, + "grad_norm": 0.6871894775171712, + "learning_rate": 2.3948333744310433e-06, + "loss": 0.3225851535797119, + "num_input_tokens_seen": 305778064, + "step": 2715, + "train_runtime": 37750.264, + "train_tokens_per_second": 8100.025 + }, + { + "epoch": 1.3518886679920477, + "grad_norm": 0.6264127228474534, + "learning_rate": 2.3781940737026015e-06, + "loss": 0.3287069320678711, + "num_input_tokens_seen": 306322552, + "step": 2720, + "train_runtime": 37828.0515, + "train_tokens_per_second": 8097.762 + }, + { + "epoch": 1.3543737574552683, + "grad_norm": 0.7153335955526043, + "learning_rate": 2.361594723619522e-06, + "loss": 0.30616672039031984, + "num_input_tokens_seen": 306886408, + "step": 2725, + "train_runtime": 37906.9066, + "train_tokens_per_second": 8095.791 + }, + { + "epoch": 1.356858846918489, + "grad_norm": 0.664957383957733, + "learning_rate": 2.3450355771199574e-06, + "loss": 0.3269395589828491, + "num_input_tokens_seen": 307470920, + "step": 2730, + "train_runtime": 37979.5851, + "train_tokens_per_second": 8095.689 + }, + { + "epoch": 1.3593439363817097, + "grad_norm": 0.6008217681232805, + "learning_rate": 2.3285168865294467e-06, + "loss": 0.327655029296875, + "num_input_tokens_seen": 307989416, + "step": 2735, + "train_runtime": 38043.1653, + "train_tokens_per_second": 8095.788 + }, + { + "epoch": 1.3618290258449304, + "grad_norm": 0.6682266175114023, + "learning_rate": 2.3120389035570638e-06, + "loss": 0.33198771476745603, + "num_input_tokens_seen": 308525672, + "step": 2740, + "train_runtime": 38095.6433, + "train_tokens_per_second": 8098.713 + }, + { + "epoch": 1.364314115308151, + "grad_norm": 0.7038032050296804, + "learning_rate": 2.295601879291591e-06, + "loss": 0.30571699142456055, + "num_input_tokens_seen": 309062512, + "step": 2745, + "train_runtime": 38164.7981, + "train_tokens_per_second": 8098.104 + }, + { + "epoch": 1.3667992047713717, + "grad_norm": 0.62970568538427, + "learning_rate": 2.279206064197687e-06, + "loss": 0.32982263565063474, + "num_input_tokens_seen": 309640392, + "step": 2750, + "train_runtime": 38239.1669, + "train_tokens_per_second": 8097.467 + }, + { + "epoch": 1.3692842942345924, + "grad_norm": 0.7051661660636465, + "learning_rate": 2.2628517081120727e-06, + "loss": 0.31272001266479493, + "num_input_tokens_seen": 310195472, + "step": 2755, + "train_runtime": 38313.1232, + "train_tokens_per_second": 8096.324 + }, + { + "epoch": 1.371769383697813, + "grad_norm": 0.6631271056464644, + "learning_rate": 2.246539060239723e-06, + "loss": 0.3255968809127808, + "num_input_tokens_seen": 310748712, + "step": 2760, + "train_runtime": 38382.5698, + "train_tokens_per_second": 8096.089 + }, + { + "epoch": 1.3742544731610338, + "grad_norm": 0.6028604146021433, + "learning_rate": 2.2302683691500697e-06, + "loss": 0.3258115291595459, + "num_input_tokens_seen": 311289008, + "step": 2765, + "train_runtime": 38454.6479, + "train_tokens_per_second": 8094.964 + }, + { + "epoch": 1.3767395626242545, + "grad_norm": 0.6041675714561272, + "learning_rate": 2.214039882773217e-06, + "loss": 0.3096216440200806, + "num_input_tokens_seen": 311869152, + "step": 2770, + "train_runtime": 38530.892, + "train_tokens_per_second": 8094.003 + }, + { + "epoch": 1.3792246520874751, + "grad_norm": 0.5966704809746424, + "learning_rate": 2.197853848396153e-06, + "loss": 0.33348238468170166, + "num_input_tokens_seen": 312462248, + "step": 2775, + "train_runtime": 38605.4601, + "train_tokens_per_second": 8093.732 + }, + { + "epoch": 1.3817097415506958, + "grad_norm": 0.718133326033276, + "learning_rate": 2.1817105126589965e-06, + "loss": 0.32853498458862307, + "num_input_tokens_seen": 313022608, + "step": 2780, + "train_runtime": 38667.4308, + "train_tokens_per_second": 8095.252 + }, + { + "epoch": 1.3841948310139165, + "grad_norm": 0.6715408976856361, + "learning_rate": 2.1656101215512302e-06, + "loss": 0.324610161781311, + "num_input_tokens_seen": 313565440, + "step": 2785, + "train_runtime": 38738.3196, + "train_tokens_per_second": 8094.451 + }, + { + "epoch": 1.3866799204771372, + "grad_norm": 0.6806927452705147, + "learning_rate": 2.149552920407952e-06, + "loss": 0.3303879976272583, + "num_input_tokens_seen": 314095456, + "step": 2790, + "train_runtime": 38804.6849, + "train_tokens_per_second": 8094.266 + }, + { + "epoch": 1.3891650099403579, + "grad_norm": 0.6726696199133415, + "learning_rate": 2.1335391539061385e-06, + "loss": 0.3162594556808472, + "num_input_tokens_seen": 314670712, + "step": 2795, + "train_runtime": 38871.8138, + "train_tokens_per_second": 8095.087 + }, + { + "epoch": 1.3916500994035785, + "grad_norm": 1.0932365465508813, + "learning_rate": 2.117569066060917e-06, + "loss": 0.3130491256713867, + "num_input_tokens_seen": 315243624, + "step": 2800, + "train_runtime": 38945.6506, + "train_tokens_per_second": 8094.45 + }, + { + "epoch": 1.3941351888667992, + "grad_norm": 0.6400767813779699, + "learning_rate": 2.1016429002218486e-06, + "loss": 0.31258111000061034, + "num_input_tokens_seen": 315828656, + "step": 2805, + "train_runtime": 39019.0222, + "train_tokens_per_second": 8094.223 + }, + { + "epoch": 1.39662027833002, + "grad_norm": 0.6278428108473181, + "learning_rate": 2.0857608990692125e-06, + "loss": 0.3067617416381836, + "num_input_tokens_seen": 316388800, + "step": 2810, + "train_runtime": 39089.4439, + "train_tokens_per_second": 8093.97 + }, + { + "epoch": 1.3991053677932406, + "grad_norm": 0.6486188945063442, + "learning_rate": 2.0699233046103183e-06, + "loss": 0.3267924070358276, + "num_input_tokens_seen": 316971752, + "step": 2815, + "train_runtime": 39158.5322, + "train_tokens_per_second": 8094.577 + }, + { + "epoch": 1.4015904572564613, + "grad_norm": 0.6791996820229442, + "learning_rate": 2.0541303581758137e-06, + "loss": 0.3356635570526123, + "num_input_tokens_seen": 317533040, + "step": 2820, + "train_runtime": 39223.782, + "train_tokens_per_second": 8095.421 + }, + { + "epoch": 1.404075546719682, + "grad_norm": 0.5989036795284862, + "learning_rate": 2.0383823004160054e-06, + "loss": 0.3247509479522705, + "num_input_tokens_seen": 318084288, + "step": 2825, + "train_runtime": 39287.696, + "train_tokens_per_second": 8096.283 + }, + { + "epoch": 1.4065606361829026, + "grad_norm": 0.6970233193088624, + "learning_rate": 2.022679371297195e-06, + "loss": 0.3100812196731567, + "num_input_tokens_seen": 318628168, + "step": 2830, + "train_runtime": 39359.9802, + "train_tokens_per_second": 8095.232 + }, + { + "epoch": 1.4090457256461233, + "grad_norm": 0.6543192934579225, + "learning_rate": 2.0070218100980205e-06, + "loss": 0.32055277824401857, + "num_input_tokens_seen": 319215384, + "step": 2835, + "train_runtime": 39431.9598, + "train_tokens_per_second": 8095.347 + }, + { + "epoch": 1.411530815109344, + "grad_norm": 0.6452251650887227, + "learning_rate": 1.991409855405812e-06, + "loss": 0.31725139617919923, + "num_input_tokens_seen": 319784152, + "step": 2840, + "train_runtime": 39498.5922, + "train_tokens_per_second": 8096.09 + }, + { + "epoch": 1.4140159045725647, + "grad_norm": 0.757194486452002, + "learning_rate": 1.9758437451129548e-06, + "loss": 0.33520400524139404, + "num_input_tokens_seen": 320322216, + "step": 2845, + "train_runtime": 39565.9129, + "train_tokens_per_second": 8095.914 + }, + { + "epoch": 1.4165009940357853, + "grad_norm": 0.6751423258115048, + "learning_rate": 1.9603237164132637e-06, + "loss": 0.3213017225265503, + "num_input_tokens_seen": 320866760, + "step": 2850, + "train_runtime": 39633.0139, + "train_tokens_per_second": 8095.946 + }, + { + "epoch": 1.418986083499006, + "grad_norm": 0.6203989744125594, + "learning_rate": 1.944850005798372e-06, + "loss": 0.3139894962310791, + "num_input_tokens_seen": 321438408, + "step": 2855, + "train_runtime": 39703.683, + "train_tokens_per_second": 8095.934 + }, + { + "epoch": 1.4214711729622267, + "grad_norm": 0.6509657060223445, + "learning_rate": 1.929422849054126e-06, + "loss": 0.31051268577575686, + "num_input_tokens_seen": 321974064, + "step": 2860, + "train_runtime": 39776.2073, + "train_tokens_per_second": 8094.64 + }, + { + "epoch": 1.4239562624254474, + "grad_norm": 0.6149524542703355, + "learning_rate": 1.9140424812569865e-06, + "loss": 0.31319990158081057, + "num_input_tokens_seen": 322536080, + "step": 2865, + "train_runtime": 39844.3429, + "train_tokens_per_second": 8094.903 + }, + { + "epoch": 1.426441351888668, + "grad_norm": 0.6192685584235147, + "learning_rate": 1.8987091367704574e-06, + "loss": 0.32322757244110106, + "num_input_tokens_seen": 323089904, + "step": 2870, + "train_runtime": 39901.8007, + "train_tokens_per_second": 8097.126 + }, + { + "epoch": 1.4289264413518887, + "grad_norm": 0.682898207054021, + "learning_rate": 1.8834230492415078e-06, + "loss": 0.33103797435760496, + "num_input_tokens_seen": 323691944, + "step": 2875, + "train_runtime": 39976.5752, + "train_tokens_per_second": 8097.04 + }, + { + "epoch": 1.4314115308151094, + "grad_norm": 0.652736686762331, + "learning_rate": 1.8681844515970138e-06, + "loss": 0.30420145988464353, + "num_input_tokens_seen": 324270928, + "step": 2880, + "train_runtime": 40042.2876, + "train_tokens_per_second": 8098.212 + }, + { + "epoch": 1.43389662027833, + "grad_norm": 0.5980591342134998, + "learning_rate": 1.8529935760402062e-06, + "loss": 0.31876139640808104, + "num_input_tokens_seen": 324833432, + "step": 2885, + "train_runtime": 40122.833, + "train_tokens_per_second": 8095.974 + }, + { + "epoch": 1.4363817097415508, + "grad_norm": 0.6518475592488924, + "learning_rate": 1.8378506540471363e-06, + "loss": 0.33255369663238527, + "num_input_tokens_seen": 325409592, + "step": 2890, + "train_runtime": 40197.1839, + "train_tokens_per_second": 8095.333 + }, + { + "epoch": 1.4388667992047715, + "grad_norm": 0.6908940064701478, + "learning_rate": 1.8227559163631475e-06, + "loss": 0.3353426933288574, + "num_input_tokens_seen": 325958872, + "step": 2895, + "train_runtime": 40270.9491, + "train_tokens_per_second": 8094.144 + }, + { + "epoch": 1.4413518886679921, + "grad_norm": 0.5991321078036727, + "learning_rate": 1.8077095929993565e-06, + "loss": 0.34195680618286134, + "num_input_tokens_seen": 326534080, + "step": 2900, + "train_runtime": 40348.0139, + "train_tokens_per_second": 8092.941 + }, + { + "epoch": 1.4438369781312126, + "grad_norm": 0.6645254352037387, + "learning_rate": 1.7927119132291526e-06, + "loss": 0.3130086660385132, + "num_input_tokens_seen": 327083344, + "step": 2905, + "train_runtime": 40419.3163, + "train_tokens_per_second": 8092.253 + }, + { + "epoch": 1.4463220675944335, + "grad_norm": 0.6175759887273283, + "learning_rate": 1.7777631055847012e-06, + "loss": 0.30218257904052737, + "num_input_tokens_seen": 327657304, + "step": 2910, + "train_runtime": 40485.9716, + "train_tokens_per_second": 8093.107 + }, + { + "epoch": 1.448807157057654, + "grad_norm": 0.7813084350324443, + "learning_rate": 1.7628633978534621e-06, + "loss": 0.3164646863937378, + "num_input_tokens_seen": 328195488, + "step": 2915, + "train_runtime": 40547.3586, + "train_tokens_per_second": 8094.127 + }, + { + "epoch": 1.4512922465208749, + "grad_norm": 0.6544006075480672, + "learning_rate": 1.74801301707472e-06, + "loss": 0.32097840309143066, + "num_input_tokens_seen": 328772600, + "step": 2920, + "train_runtime": 40623.1936, + "train_tokens_per_second": 8093.224 + }, + { + "epoch": 1.4537773359840953, + "grad_norm": 0.6083901049343592, + "learning_rate": 1.7332121895361209e-06, + "loss": 0.31046261787414553, + "num_input_tokens_seen": 329340472, + "step": 2925, + "train_runtime": 40690.3113, + "train_tokens_per_second": 8093.83 + }, + { + "epoch": 1.4562624254473162, + "grad_norm": 0.6575847535819591, + "learning_rate": 1.7184611407702289e-06, + "loss": 0.316282844543457, + "num_input_tokens_seen": 329913176, + "step": 2930, + "train_runtime": 40766.168, + "train_tokens_per_second": 8092.818 + }, + { + "epoch": 1.4587475149105367, + "grad_norm": 0.6888187821970063, + "learning_rate": 1.703760095551088e-06, + "loss": 0.31734318733215333, + "num_input_tokens_seen": 330482216, + "step": 2935, + "train_runtime": 40832.292, + "train_tokens_per_second": 8093.648 + }, + { + "epoch": 1.4612326043737576, + "grad_norm": 0.7072648332515346, + "learning_rate": 1.6891092778907942e-06, + "loss": 0.3230691194534302, + "num_input_tokens_seen": 331016600, + "step": 2940, + "train_runtime": 40895.4504, + "train_tokens_per_second": 8094.216 + }, + { + "epoch": 1.463717693836978, + "grad_norm": 0.6127877570120994, + "learning_rate": 1.6745089110360862e-06, + "loss": 0.3366373538970947, + "num_input_tokens_seen": 331587248, + "step": 2945, + "train_runtime": 40962.3175, + "train_tokens_per_second": 8094.934 + }, + { + "epoch": 1.466202783300199, + "grad_norm": 0.6887352858196922, + "learning_rate": 1.659959217464941e-06, + "loss": 0.3315255880355835, + "num_input_tokens_seen": 332178608, + "step": 2950, + "train_runtime": 41039.505, + "train_tokens_per_second": 8094.118 + }, + { + "epoch": 1.4686878727634194, + "grad_norm": 0.6190277506247961, + "learning_rate": 1.645460418883184e-06, + "loss": 0.32088706493377683, + "num_input_tokens_seen": 332752984, + "step": 2955, + "train_runtime": 41113.8835, + "train_tokens_per_second": 8093.446 + }, + { + "epoch": 1.4711729622266403, + "grad_norm": 0.9255515530741937, + "learning_rate": 1.631012736221112e-06, + "loss": 0.31818628311157227, + "num_input_tokens_seen": 333310568, + "step": 2960, + "train_runtime": 41178.7993, + "train_tokens_per_second": 8094.227 + }, + { + "epoch": 1.4736580516898607, + "grad_norm": 0.6300217617458029, + "learning_rate": 1.6166163896301251e-06, + "loss": 0.30632126331329346, + "num_input_tokens_seen": 333851096, + "step": 2965, + "train_runtime": 41247.1822, + "train_tokens_per_second": 8093.913 + }, + { + "epoch": 1.4761431411530814, + "grad_norm": 0.6779953695507507, + "learning_rate": 1.6022715984793736e-06, + "loss": 0.32839550971984866, + "num_input_tokens_seen": 334405512, + "step": 2970, + "train_runtime": 41312.1086, + "train_tokens_per_second": 8094.613 + }, + { + "epoch": 1.478628230616302, + "grad_norm": 0.5609991972065584, + "learning_rate": 1.587978581352415e-06, + "loss": 0.31574239730834963, + "num_input_tokens_seen": 334981064, + "step": 2975, + "train_runtime": 41381.2778, + "train_tokens_per_second": 8094.991 + }, + { + "epoch": 1.4811133200795228, + "grad_norm": 0.5921412924705993, + "learning_rate": 1.5737375560438794e-06, + "loss": 0.3063444852828979, + "num_input_tokens_seen": 335532376, + "step": 2980, + "train_runtime": 41449.2757, + "train_tokens_per_second": 8095.012 + }, + { + "epoch": 1.4835984095427435, + "grad_norm": 0.6543399543812114, + "learning_rate": 1.5595487395561575e-06, + "loss": 0.32524838447570803, + "num_input_tokens_seen": 336051664, + "step": 2985, + "train_runtime": 41507.85, + "train_tokens_per_second": 8096.099 + }, + { + "epoch": 1.4860834990059641, + "grad_norm": 0.7793034374101312, + "learning_rate": 1.5454123480960903e-06, + "loss": 0.30900025367736816, + "num_input_tokens_seen": 336592464, + "step": 2990, + "train_runtime": 41580.0213, + "train_tokens_per_second": 8095.053 + }, + { + "epoch": 1.4885685884691848, + "grad_norm": 0.585215729709737, + "learning_rate": 1.531328597071674e-06, + "loss": 0.3024424076080322, + "num_input_tokens_seen": 337150392, + "step": 2995, + "train_runtime": 41651.5755, + "train_tokens_per_second": 8094.541 + }, + { + "epoch": 1.4910536779324055, + "grad_norm": 0.6698192386537378, + "learning_rate": 1.517297701088779e-06, + "loss": 0.3347722053527832, + "num_input_tokens_seen": 337716448, + "step": 3000, + "train_runtime": 41724.4605, + "train_tokens_per_second": 8093.968 + }, + { + "epoch": 1.4935387673956262, + "grad_norm": 0.5924415445415384, + "learning_rate": 1.5033198739478793e-06, + "loss": 0.3223055124282837, + "num_input_tokens_seen": 338285816, + "step": 3005, + "train_runtime": 41793.7955, + "train_tokens_per_second": 8094.164 + }, + { + "epoch": 1.4960238568588469, + "grad_norm": 0.6522488694464076, + "learning_rate": 1.4893953286407947e-06, + "loss": 0.32070398330688477, + "num_input_tokens_seen": 338839456, + "step": 3010, + "train_runtime": 41855.7502, + "train_tokens_per_second": 8095.41 + }, + { + "epoch": 1.4985089463220675, + "grad_norm": 0.6008849067333178, + "learning_rate": 1.4755242773474448e-06, + "loss": 0.34209237098693845, + "num_input_tokens_seen": 339423080, + "step": 3015, + "train_runtime": 41933.7026, + "train_tokens_per_second": 8094.279 + }, + { + "epoch": 1.5009940357852882, + "grad_norm": 0.6382130773115663, + "learning_rate": 1.4617069314326172e-06, + "loss": 0.31463284492492677, + "num_input_tokens_seen": 340009296, + "step": 3020, + "train_runtime": 42001.6842, + "train_tokens_per_second": 8095.135 + }, + { + "epoch": 1.503479125248509, + "grad_norm": 0.6187763336985728, + "learning_rate": 1.447943501442744e-06, + "loss": 0.31114366054534914, + "num_input_tokens_seen": 340586864, + "step": 3025, + "train_runtime": 42072.6709, + "train_tokens_per_second": 8095.204 + }, + { + "epoch": 1.5059642147117296, + "grad_norm": 0.6647223277686793, + "learning_rate": 1.4342341971026975e-06, + "loss": 0.3133885860443115, + "num_input_tokens_seen": 341156544, + "step": 3030, + "train_runtime": 42145.9895, + "train_tokens_per_second": 8094.638 + }, + { + "epoch": 1.5084493041749503, + "grad_norm": 0.6132246241878623, + "learning_rate": 1.4205792273125917e-06, + "loss": 0.31753480434417725, + "num_input_tokens_seen": 341723976, + "step": 3035, + "train_runtime": 42211.0493, + "train_tokens_per_second": 8095.605 + }, + { + "epoch": 1.510934393638171, + "grad_norm": 0.8984737673406371, + "learning_rate": 1.4069788001445978e-06, + "loss": 0.3069683313369751, + "num_input_tokens_seen": 342256416, + "step": 3040, + "train_runtime": 42266.9618, + "train_tokens_per_second": 8097.493 + }, + { + "epoch": 1.5134194831013916, + "grad_norm": 0.5917049715483013, + "learning_rate": 1.3934331228397774e-06, + "loss": 0.31336321830749514, + "num_input_tokens_seen": 342817928, + "step": 3045, + "train_runtime": 42344.4457, + "train_tokens_per_second": 8095.936 + }, + { + "epoch": 1.5159045725646123, + "grad_norm": 0.6324853808854811, + "learning_rate": 1.3799424018049245e-06, + "loss": 0.336625862121582, + "num_input_tokens_seen": 343389200, + "step": 3050, + "train_runtime": 42416.5441, + "train_tokens_per_second": 8095.643 + }, + { + "epoch": 1.518389662027833, + "grad_norm": 0.6156596717252506, + "learning_rate": 1.366506842609413e-06, + "loss": 0.30888302326202394, + "num_input_tokens_seen": 343942352, + "step": 3055, + "train_runtime": 42479.2067, + "train_tokens_per_second": 8096.723 + }, + { + "epoch": 1.5208747514910537, + "grad_norm": 0.6914691275574822, + "learning_rate": 1.353126649982075e-06, + "loss": 0.3196723222732544, + "num_input_tokens_seen": 344476216, + "step": 3060, + "train_runtime": 42544.23, + "train_tokens_per_second": 8096.896 + }, + { + "epoch": 1.5233598409542743, + "grad_norm": 0.6076065686006078, + "learning_rate": 1.3398020278080742e-06, + "loss": 0.31299633979797364, + "num_input_tokens_seen": 345023576, + "step": 3065, + "train_runtime": 42617.3044, + "train_tokens_per_second": 8095.856 + }, + { + "epoch": 1.525844930417495, + "grad_norm": 0.6525409540021935, + "learning_rate": 1.3265331791258008e-06, + "loss": 0.31205224990844727, + "num_input_tokens_seen": 345577072, + "step": 3070, + "train_runtime": 42683.8625, + "train_tokens_per_second": 8096.2 + }, + { + "epoch": 1.5283300198807157, + "grad_norm": 0.6422296311367647, + "learning_rate": 1.3133203061237775e-06, + "loss": 0.3147293567657471, + "num_input_tokens_seen": 346150200, + "step": 3075, + "train_runtime": 42760.4469, + "train_tokens_per_second": 8095.102 + }, + { + "epoch": 1.5308151093439364, + "grad_norm": 0.6282010229703389, + "learning_rate": 1.3001636101375804e-06, + "loss": 0.30835280418395994, + "num_input_tokens_seen": 346673856, + "step": 3080, + "train_runtime": 42830.404, + "train_tokens_per_second": 8094.107 + }, + { + "epoch": 1.533300198807157, + "grad_norm": 0.6824954584861774, + "learning_rate": 1.2870632916467673e-06, + "loss": 0.303128719329834, + "num_input_tokens_seen": 347212208, + "step": 3085, + "train_runtime": 42903.229, + "train_tokens_per_second": 8092.916 + }, + { + "epoch": 1.5357852882703777, + "grad_norm": 0.6120096179741551, + "learning_rate": 1.2740195502718266e-06, + "loss": 0.3151275157928467, + "num_input_tokens_seen": 347746888, + "step": 3090, + "train_runtime": 42966.7784, + "train_tokens_per_second": 8093.39 + }, + { + "epoch": 1.5382703777335984, + "grad_norm": 0.5921461763049634, + "learning_rate": 1.2610325847711336e-06, + "loss": 0.3232539176940918, + "num_input_tokens_seen": 348289368, + "step": 3095, + "train_runtime": 43033.251, + "train_tokens_per_second": 8093.494 + }, + { + "epoch": 1.540755467196819, + "grad_norm": 0.6571697732878194, + "learning_rate": 1.2481025930379214e-06, + "loss": 0.3129811525344849, + "num_input_tokens_seen": 348841200, + "step": 3100, + "train_runtime": 43099.9776, + "train_tokens_per_second": 8093.768 + }, + { + "epoch": 1.5432405566600398, + "grad_norm": 0.6135570921165027, + "learning_rate": 1.2352297720972689e-06, + "loss": 0.31615419387817384, + "num_input_tokens_seen": 349422680, + "step": 3105, + "train_runtime": 43168.8614, + "train_tokens_per_second": 8094.322 + }, + { + "epoch": 1.5457256461232605, + "grad_norm": 0.694058625138879, + "learning_rate": 1.2224143181030907e-06, + "loss": 0.3251053810119629, + "num_input_tokens_seen": 349994696, + "step": 3110, + "train_runtime": 43237.6624, + "train_tokens_per_second": 8094.672 + }, + { + "epoch": 1.5482107355864811, + "grad_norm": 0.5725321060866029, + "learning_rate": 1.209656426335158e-06, + "loss": 0.3102801561355591, + "num_input_tokens_seen": 350540704, + "step": 3115, + "train_runtime": 43300.2187, + "train_tokens_per_second": 8095.587 + }, + { + "epoch": 1.5506958250497018, + "grad_norm": 0.6126863378607672, + "learning_rate": 1.1969562911961168e-06, + "loss": 0.3238799571990967, + "num_input_tokens_seen": 351095816, + "step": 3120, + "train_runtime": 43364.7261, + "train_tokens_per_second": 8096.346 + }, + { + "epoch": 1.5531809145129225, + "grad_norm": 0.6990159971194807, + "learning_rate": 1.1843141062085278e-06, + "loss": 0.31937243938446047, + "num_input_tokens_seen": 351687088, + "step": 3125, + "train_runtime": 43432.567, + "train_tokens_per_second": 8097.313 + }, + { + "epoch": 1.5556660039761432, + "grad_norm": 0.6045485676457566, + "learning_rate": 1.1717300640119162e-06, + "loss": 0.3220341205596924, + "num_input_tokens_seen": 352236616, + "step": 3130, + "train_runtime": 43509.9646, + "train_tokens_per_second": 8095.539 + }, + { + "epoch": 1.5581510934393639, + "grad_norm": 0.5928689194377964, + "learning_rate": 1.1592043563598365e-06, + "loss": 0.3087661027908325, + "num_input_tokens_seen": 352790056, + "step": 3135, + "train_runtime": 43577.4906, + "train_tokens_per_second": 8095.695 + }, + { + "epoch": 1.5606361829025845, + "grad_norm": 0.5785307620510605, + "learning_rate": 1.1467371741169519e-06, + "loss": 0.31176204681396485, + "num_input_tokens_seen": 353377912, + "step": 3140, + "train_runtime": 43642.984, + "train_tokens_per_second": 8097.015 + }, + { + "epoch": 1.5631212723658052, + "grad_norm": 0.64250753929324, + "learning_rate": 1.134328707256125e-06, + "loss": 0.3161168575286865, + "num_input_tokens_seen": 353934680, + "step": 3145, + "train_runtime": 43713.1114, + "train_tokens_per_second": 8096.762 + }, + { + "epoch": 1.5656063618290257, + "grad_norm": 0.5812939017283076, + "learning_rate": 1.1219791448555223e-06, + "loss": 0.3318530559539795, + "num_input_tokens_seen": 354485392, + "step": 3150, + "train_runtime": 43785.626, + "train_tokens_per_second": 8095.931 + }, + { + "epoch": 1.5680914512922466, + "grad_norm": 0.5856080985348556, + "learning_rate": 1.1096886750957347e-06, + "loss": 0.3182845115661621, + "num_input_tokens_seen": 355049000, + "step": 3155, + "train_runtime": 43859.6983, + "train_tokens_per_second": 8095.108 + }, + { + "epoch": 1.570576540755467, + "grad_norm": 0.7020225974783655, + "learning_rate": 1.097457485256911e-06, + "loss": 0.33698220252990724, + "num_input_tokens_seen": 355587696, + "step": 3160, + "train_runtime": 43924.5859, + "train_tokens_per_second": 8095.414 + }, + { + "epoch": 1.573061630218688, + "grad_norm": 0.5969074665471703, + "learning_rate": 1.085285761715897e-06, + "loss": 0.3074070453643799, + "num_input_tokens_seen": 356152560, + "step": 3165, + "train_runtime": 43992.5106, + "train_tokens_per_second": 8095.754 + }, + { + "epoch": 1.5755467196819084, + "grad_norm": 0.6372155855954388, + "learning_rate": 1.0731736899434053e-06, + "loss": 0.31992578506469727, + "num_input_tokens_seen": 356730728, + "step": 3170, + "train_runtime": 44057.9143, + "train_tokens_per_second": 8096.859 + }, + { + "epoch": 1.5780318091451293, + "grad_norm": 0.6739486103613654, + "learning_rate": 1.0611214545011845e-06, + "loss": 0.3220165491104126, + "num_input_tokens_seen": 357317104, + "step": 3175, + "train_runtime": 44128.745, + "train_tokens_per_second": 8097.151 + }, + { + "epoch": 1.5805168986083498, + "grad_norm": 0.6148825360748863, + "learning_rate": 1.0491292390392055e-06, + "loss": 0.3173060417175293, + "num_input_tokens_seen": 357881896, + "step": 3180, + "train_runtime": 44196.5924, + "train_tokens_per_second": 8097.5 + }, + { + "epoch": 1.5830019880715707, + "grad_norm": 0.5615815248877837, + "learning_rate": 1.0371972262928653e-06, + "loss": 0.2923841953277588, + "num_input_tokens_seen": 358461456, + "step": 3185, + "train_runtime": 44265.8158, + "train_tokens_per_second": 8097.93 + }, + { + "epoch": 1.5854870775347911, + "grad_norm": 0.6294903048453709, + "learning_rate": 1.0253255980802029e-06, + "loss": 0.31896076202392576, + "num_input_tokens_seen": 359009032, + "step": 3190, + "train_runtime": 44329.0688, + "train_tokens_per_second": 8098.727 + }, + { + "epoch": 1.587972166998012, + "grad_norm": 0.627920839403491, + "learning_rate": 1.013514535299126e-06, + "loss": 0.3224520683288574, + "num_input_tokens_seen": 359533152, + "step": 3195, + "train_runtime": 44399.1588, + "train_tokens_per_second": 8097.747 + }, + { + "epoch": 1.5904572564612325, + "grad_norm": 0.6127120636628913, + "learning_rate": 1.0017642179246574e-06, + "loss": 0.31302390098571775, + "num_input_tokens_seen": 360118152, + "step": 3200, + "train_runtime": 44476.2584, + "train_tokens_per_second": 8096.863 + }, + { + "epoch": 1.5929423459244534, + "grad_norm": 0.6280010782981149, + "learning_rate": 9.900748250061909e-07, + "loss": 0.3197129249572754, + "num_input_tokens_seen": 360667872, + "step": 3205, + "train_runtime": 44547.2232, + "train_tokens_per_second": 8096.304 + }, + { + "epoch": 1.5954274353876738, + "grad_norm": 0.652600048520917, + "learning_rate": 9.784465346647638e-07, + "loss": 0.31191105842590333, + "num_input_tokens_seen": 361237760, + "step": 3210, + "train_runtime": 44619.0822, + "train_tokens_per_second": 8096.037 + }, + { + "epoch": 1.5979125248508947, + "grad_norm": 0.6795754070001668, + "learning_rate": 9.668795240903423e-07, + "loss": 0.3151463508605957, + "num_input_tokens_seen": 361810952, + "step": 3215, + "train_runtime": 44687.9116, + "train_tokens_per_second": 8096.394 + }, + { + "epoch": 1.6003976143141152, + "grad_norm": 0.6754602162458507, + "learning_rate": 9.553739695391228e-07, + "loss": 0.3232785701751709, + "num_input_tokens_seen": 362343976, + "step": 3220, + "train_runtime": 44755.8889, + "train_tokens_per_second": 8096.007 + }, + { + "epoch": 1.602882703777336, + "grad_norm": 0.5741065095590924, + "learning_rate": 9.439300463308421e-07, + "loss": 0.30777673721313475, + "num_input_tokens_seen": 362928672, + "step": 3225, + "train_runtime": 44825.6016, + "train_tokens_per_second": 8096.46 + }, + { + "epoch": 1.6053677932405566, + "grad_norm": 0.6516898574642163, + "learning_rate": 9.325479288461109e-07, + "loss": 0.30566678047180174, + "num_input_tokens_seen": 363529832, + "step": 3230, + "train_runtime": 44893.8226, + "train_tokens_per_second": 8097.547 + }, + { + "epoch": 1.6078528827037775, + "grad_norm": 0.6426098445115822, + "learning_rate": 9.212277905237549e-07, + "loss": 0.30101923942565917, + "num_input_tokens_seen": 364054904, + "step": 3235, + "train_runtime": 44963.18, + "train_tokens_per_second": 8096.734 + }, + { + "epoch": 1.610337972166998, + "grad_norm": 0.5960549176082046, + "learning_rate": 9.099698038581701e-07, + "loss": 0.3242988109588623, + "num_input_tokens_seen": 364654032, + "step": 3240, + "train_runtime": 45029.7567, + "train_tokens_per_second": 8098.068 + }, + { + "epoch": 1.6128230616302188, + "grad_norm": 0.7387323087959445, + "learning_rate": 8.987741403966971e-07, + "loss": 0.3314228534698486, + "num_input_tokens_seen": 365207032, + "step": 3245, + "train_runtime": 45101.7524, + "train_tokens_per_second": 8097.402 + }, + { + "epoch": 1.6153081510934393, + "grad_norm": 0.5864769612158937, + "learning_rate": 8.876409707370048e-07, + "loss": 0.3272867202758789, + "num_input_tokens_seen": 365780616, + "step": 3250, + "train_runtime": 45172.2915, + "train_tokens_per_second": 8097.455 + }, + { + "epoch": 1.6177932405566602, + "grad_norm": 0.6293663216733479, + "learning_rate": 8.765704645244927e-07, + "loss": 0.31007771492004393, + "num_input_tokens_seen": 366378552, + "step": 3255, + "train_runtime": 45239.3532, + "train_tokens_per_second": 8098.669 + }, + { + "epoch": 1.6202783300198806, + "grad_norm": 0.6148967315668572, + "learning_rate": 8.655627904497043e-07, + "loss": 0.32148215770721433, + "num_input_tokens_seen": 366958960, + "step": 3260, + "train_runtime": 45311.3944, + "train_tokens_per_second": 8098.602 + }, + { + "epoch": 1.6227634194831015, + "grad_norm": 0.6364190346056351, + "learning_rate": 8.546181162457585e-07, + "loss": 0.30961339473724364, + "num_input_tokens_seen": 367503672, + "step": 3265, + "train_runtime": 45385.8271, + "train_tokens_per_second": 8097.322 + }, + { + "epoch": 1.625248508946322, + "grad_norm": 0.6413053717931511, + "learning_rate": 8.437366086857912e-07, + "loss": 0.31174097061157224, + "num_input_tokens_seen": 368049536, + "step": 3270, + "train_runtime": 45454.7228, + "train_tokens_per_second": 8097.058 + }, + { + "epoch": 1.627733598409543, + "grad_norm": 0.6406608373261115, + "learning_rate": 8.329184335804174e-07, + "loss": 0.3215543270111084, + "num_input_tokens_seen": 368622536, + "step": 3275, + "train_runtime": 45517.3147, + "train_tokens_per_second": 8098.512 + }, + { + "epoch": 1.6302186878727634, + "grad_norm": 0.6025291891693101, + "learning_rate": 8.221637557751988e-07, + "loss": 0.31618926525115965, + "num_input_tokens_seen": 369196392, + "step": 3280, + "train_runtime": 45577.172, + "train_tokens_per_second": 8100.467 + }, + { + "epoch": 1.632703777335984, + "grad_norm": 0.6270538180783158, + "learning_rate": 8.114727391481397e-07, + "loss": 0.3081305980682373, + "num_input_tokens_seen": 369791144, + "step": 3285, + "train_runtime": 45641.1318, + "train_tokens_per_second": 8102.147 + }, + { + "epoch": 1.6351888667992047, + "grad_norm": 0.6209426759386114, + "learning_rate": 8.008455466071841e-07, + "loss": 0.32327721118927, + "num_input_tokens_seen": 370328456, + "step": 3290, + "train_runtime": 45699.0559, + "train_tokens_per_second": 8103.635 + }, + { + "epoch": 1.6376739562624254, + "grad_norm": 0.6380293234605218, + "learning_rate": 7.902823400877352e-07, + "loss": 0.3361134767532349, + "num_input_tokens_seen": 370888040, + "step": 3295, + "train_runtime": 45771.4049, + "train_tokens_per_second": 8103.051 + }, + { + "epoch": 1.640159045725646, + "grad_norm": 0.5914642789893445, + "learning_rate": 7.797832805501887e-07, + "loss": 0.3205102443695068, + "num_input_tokens_seen": 371421048, + "step": 3300, + "train_runtime": 45840.4892, + "train_tokens_per_second": 8102.467 + }, + { + "epoch": 1.6426441351888668, + "grad_norm": 1.4968427150473462, + "learning_rate": 7.693485279774782e-07, + "loss": 0.29322023391723634, + "num_input_tokens_seen": 371971152, + "step": 3305, + "train_runtime": 45910.124, + "train_tokens_per_second": 8102.16 + }, + { + "epoch": 1.6451292246520874, + "grad_norm": 0.6415965223679025, + "learning_rate": 7.589782413726387e-07, + "loss": 0.3274980068206787, + "num_input_tokens_seen": 372512192, + "step": 3310, + "train_runtime": 45987.0436, + "train_tokens_per_second": 8100.373 + }, + { + "epoch": 1.6476143141153081, + "grad_norm": 0.6358736833491117, + "learning_rate": 7.486725787563837e-07, + "loss": 0.31942410469055177, + "num_input_tokens_seen": 373111160, + "step": 3315, + "train_runtime": 46046.7063, + "train_tokens_per_second": 8102.885 + }, + { + "epoch": 1.6500994035785288, + "grad_norm": 0.6412517551330648, + "learning_rate": 7.38431697164696e-07, + "loss": 0.33071558475494384, + "num_input_tokens_seen": 373680712, + "step": 3320, + "train_runtime": 46122.4826, + "train_tokens_per_second": 8101.921 + }, + { + "epoch": 1.6525844930417495, + "grad_norm": 0.6284493158730468, + "learning_rate": 7.282557526464373e-07, + "loss": 0.34810700416564944, + "num_input_tokens_seen": 374245120, + "step": 3325, + "train_runtime": 46191.2978, + "train_tokens_per_second": 8102.07 + }, + { + "epoch": 1.6550695825049702, + "grad_norm": 0.5661420673732055, + "learning_rate": 7.181449002609691e-07, + "loss": 0.3048112392425537, + "num_input_tokens_seen": 374779288, + "step": 3330, + "train_runtime": 46252.16, + "train_tokens_per_second": 8102.958 + }, + { + "epoch": 1.6575546719681908, + "grad_norm": 0.5784411447481479, + "learning_rate": 7.080992940757869e-07, + "loss": 0.2945127487182617, + "num_input_tokens_seen": 375328808, + "step": 3335, + "train_runtime": 46316.3496, + "train_tokens_per_second": 8103.592 + }, + { + "epoch": 1.6600397614314115, + "grad_norm": 0.8658889907358424, + "learning_rate": 6.981190871641785e-07, + "loss": 0.32708420753479006, + "num_input_tokens_seen": 375872728, + "step": 3340, + "train_runtime": 46383.2269, + "train_tokens_per_second": 8103.635 + }, + { + "epoch": 1.6625248508946322, + "grad_norm": 0.8606049669378169, + "learning_rate": 6.88204431602888e-07, + "loss": 0.3353096961975098, + "num_input_tokens_seen": 376447480, + "step": 3345, + "train_runtime": 46455.9116, + "train_tokens_per_second": 8103.328 + }, + { + "epoch": 1.6650099403578529, + "grad_norm": 0.5663562336170048, + "learning_rate": 6.78355478469796e-07, + "loss": 0.31557445526123046, + "num_input_tokens_seen": 377024664, + "step": 3350, + "train_runtime": 46528.7879, + "train_tokens_per_second": 8103.041 + }, + { + "epoch": 1.6674950298210736, + "grad_norm": 0.6286806019157742, + "learning_rate": 6.685723778416242e-07, + "loss": 0.3248952865600586, + "num_input_tokens_seen": 377589048, + "step": 3355, + "train_runtime": 46598.8839, + "train_tokens_per_second": 8102.963 + }, + { + "epoch": 1.6699801192842942, + "grad_norm": 0.6111937402982289, + "learning_rate": 6.588552787916419e-07, + "loss": 0.3210245847702026, + "num_input_tokens_seen": 378190200, + "step": 3360, + "train_runtime": 46662.1119, + "train_tokens_per_second": 8104.867 + }, + { + "epoch": 1.672465208747515, + "grad_norm": 0.6101887660941094, + "learning_rate": 6.49204329387399e-07, + "loss": 0.31774330139160156, + "num_input_tokens_seen": 378755376, + "step": 3365, + "train_runtime": 46721.0887, + "train_tokens_per_second": 8106.733 + }, + { + "epoch": 1.6749502982107356, + "grad_norm": 0.5643044680239365, + "learning_rate": 6.396196766884671e-07, + "loss": 0.31786355972290037, + "num_input_tokens_seen": 379354496, + "step": 3370, + "train_runtime": 46788.5025, + "train_tokens_per_second": 8107.857 + }, + { + "epoch": 1.6774353876739563, + "grad_norm": 0.6388497766407458, + "learning_rate": 6.301014667442001e-07, + "loss": 0.3027843713760376, + "num_input_tokens_seen": 379920984, + "step": 3375, + "train_runtime": 46860.2678, + "train_tokens_per_second": 8107.529 + }, + { + "epoch": 1.679920477137177, + "grad_norm": 0.6061320606834356, + "learning_rate": 6.206498445915082e-07, + "loss": 0.2907253265380859, + "num_input_tokens_seen": 380475784, + "step": 3380, + "train_runtime": 46925.192, + "train_tokens_per_second": 8108.135 + }, + { + "epoch": 1.6824055666003976, + "grad_norm": 0.5501212128475166, + "learning_rate": 6.112649542526466e-07, + "loss": 0.30368647575378416, + "num_input_tokens_seen": 381061200, + "step": 3385, + "train_runtime": 46995.3597, + "train_tokens_per_second": 8108.486 + }, + { + "epoch": 1.6848906560636183, + "grad_norm": 0.8813905792149939, + "learning_rate": 6.01946938733024e-07, + "loss": 0.3421525001525879, + "num_input_tokens_seen": 381648512, + "step": 3390, + "train_runtime": 47071.6134, + "train_tokens_per_second": 8107.827 + }, + { + "epoch": 1.687375745526839, + "grad_norm": 0.6389765764640893, + "learning_rate": 5.926959400190207e-07, + "loss": 0.33472723960876466, + "num_input_tokens_seen": 382219256, + "step": 3395, + "train_runtime": 47138.5785, + "train_tokens_per_second": 8108.417 + }, + { + "epoch": 1.6898608349900597, + "grad_norm": 0.6452430352832449, + "learning_rate": 5.835120990758264e-07, + "loss": 0.3110776901245117, + "num_input_tokens_seen": 382767528, + "step": 3400, + "train_runtime": 47211.0486, + "train_tokens_per_second": 8107.584 + }, + { + "epoch": 1.6923459244532804, + "grad_norm": 0.6217394969355298, + "learning_rate": 5.743955558452902e-07, + "loss": 0.3155216217041016, + "num_input_tokens_seen": 383335096, + "step": 3405, + "train_runtime": 47275.85, + "train_tokens_per_second": 8108.476 + }, + { + "epoch": 1.694831013916501, + "grad_norm": 0.5950131578011024, + "learning_rate": 5.653464492437916e-07, + "loss": 0.33304295539855955, + "num_input_tokens_seen": 383904704, + "step": 3410, + "train_runtime": 47341.1198, + "train_tokens_per_second": 8109.329 + }, + { + "epoch": 1.6973161033797217, + "grad_norm": 0.6164514103893838, + "learning_rate": 5.56364917160121e-07, + "loss": 0.297628116607666, + "num_input_tokens_seen": 384470136, + "step": 3415, + "train_runtime": 47410.3151, + "train_tokens_per_second": 8109.42 + }, + { + "epoch": 1.6998011928429424, + "grad_norm": 0.7085854257914194, + "learning_rate": 5.474510964533808e-07, + "loss": 0.3105302333831787, + "num_input_tokens_seen": 385019640, + "step": 3420, + "train_runtime": 47480.4641, + "train_tokens_per_second": 8109.012 + }, + { + "epoch": 1.7022862823061629, + "grad_norm": 0.5669507671560684, + "learning_rate": 5.386051229508965e-07, + "loss": 0.3406516551971436, + "num_input_tokens_seen": 385550008, + "step": 3425, + "train_runtime": 47543.4999, + "train_tokens_per_second": 8109.416 + }, + { + "epoch": 1.7047713717693838, + "grad_norm": 0.6183491573263754, + "learning_rate": 5.298271314461511e-07, + "loss": 0.315033483505249, + "num_input_tokens_seen": 386104624, + "step": 3430, + "train_runtime": 47612.3801, + "train_tokens_per_second": 8109.333 + }, + { + "epoch": 1.7072564612326042, + "grad_norm": 0.7043755691254879, + "learning_rate": 5.211172556967276e-07, + "loss": 0.3052037239074707, + "num_input_tokens_seen": 386655064, + "step": 3435, + "train_runtime": 47677.8704, + "train_tokens_per_second": 8109.739 + }, + { + "epoch": 1.7097415506958251, + "grad_norm": 0.7158766705432269, + "learning_rate": 5.124756284222737e-07, + "loss": 0.312931227684021, + "num_input_tokens_seen": 387200632, + "step": 3440, + "train_runtime": 47750.2408, + "train_tokens_per_second": 8108.873 + }, + { + "epoch": 1.7122266401590456, + "grad_norm": 0.5809883646584292, + "learning_rate": 5.039023813024769e-07, + "loss": 0.317107105255127, + "num_input_tokens_seen": 387773192, + "step": 3445, + "train_runtime": 47817.7385, + "train_tokens_per_second": 8109.4 + }, + { + "epoch": 1.7147117296222665, + "grad_norm": 0.7465857777084137, + "learning_rate": 4.953976449750603e-07, + "loss": 0.3150139093399048, + "num_input_tokens_seen": 388278152, + "step": 3450, + "train_runtime": 47887.381, + "train_tokens_per_second": 8108.152 + }, + { + "epoch": 1.717196819085487, + "grad_norm": 0.5878178099660997, + "learning_rate": 4.869615490337898e-07, + "loss": 0.30156168937683103, + "num_input_tokens_seen": 388857976, + "step": 3455, + "train_runtime": 47966.0437, + "train_tokens_per_second": 8106.943 + }, + { + "epoch": 1.7196819085487078, + "grad_norm": 0.6235447646043073, + "learning_rate": 4.785942220265028e-07, + "loss": 0.3213622570037842, + "num_input_tokens_seen": 389433808, + "step": 3460, + "train_runtime": 48033.4458, + "train_tokens_per_second": 8107.555 + }, + { + "epoch": 1.7221669980119283, + "grad_norm": 0.6514446911845969, + "learning_rate": 4.702957914531425e-07, + "loss": 0.3123664140701294, + "num_input_tokens_seen": 390015968, + "step": 3465, + "train_runtime": 48103.8607, + "train_tokens_per_second": 8107.789 + }, + { + "epoch": 1.7246520874751492, + "grad_norm": 0.6061782790826798, + "learning_rate": 4.6206638376382394e-07, + "loss": 0.32015409469604494, + "num_input_tokens_seen": 390583160, + "step": 3470, + "train_runtime": 48171.5762, + "train_tokens_per_second": 8108.166 + }, + { + "epoch": 1.7271371769383697, + "grad_norm": 0.6288429525761271, + "learning_rate": 4.539061243569004e-07, + "loss": 0.3095594882965088, + "num_input_tokens_seen": 391155216, + "step": 3475, + "train_runtime": 48242.7318, + "train_tokens_per_second": 8108.065 + }, + { + "epoch": 1.7296222664015906, + "grad_norm": 0.594951102621166, + "learning_rate": 4.4581513757705696e-07, + "loss": 0.325455904006958, + "num_input_tokens_seen": 391745440, + "step": 3480, + "train_runtime": 48312.2409, + "train_tokens_per_second": 8108.617 + }, + { + "epoch": 1.732107355864811, + "grad_norm": 0.6597131315437922, + "learning_rate": 4.3779354671341255e-07, + "loss": 0.30549724102020265, + "num_input_tokens_seen": 392310752, + "step": 3485, + "train_runtime": 48372.3238, + "train_tokens_per_second": 8110.232 + }, + { + "epoch": 1.734592445328032, + "grad_norm": 0.6098141310139199, + "learning_rate": 4.2984147399764254e-07, + "loss": 0.3285971641540527, + "num_input_tokens_seen": 392888464, + "step": 3490, + "train_runtime": 48444.7177, + "train_tokens_per_second": 8110.037 + }, + { + "epoch": 1.7370775347912524, + "grad_norm": 0.5628261324620512, + "learning_rate": 4.2195904060211735e-07, + "loss": 0.31354660987854005, + "num_input_tokens_seen": 393448024, + "step": 3495, + "train_runtime": 48518.0871, + "train_tokens_per_second": 8109.306 + }, + { + "epoch": 1.7395626242544733, + "grad_norm": 0.617467375530746, + "learning_rate": 4.1414636663805355e-07, + "loss": 0.3031325340270996, + "num_input_tokens_seen": 393991336, + "step": 3500, + "train_runtime": 48587.0211, + "train_tokens_per_second": 8108.983 + }, + { + "epoch": 1.7420477137176937, + "grad_norm": 0.5633406680136066, + "learning_rate": 4.064035711536862e-07, + "loss": 0.31957159042358396, + "num_input_tokens_seen": 394573368, + "step": 3505, + "train_runtime": 48662.7519, + "train_tokens_per_second": 8108.324 + }, + { + "epoch": 1.7445328031809146, + "grad_norm": 0.59158597609514, + "learning_rate": 3.9873077213245236e-07, + "loss": 0.31620006561279296, + "num_input_tokens_seen": 395160080, + "step": 3510, + "train_runtime": 48733.6918, + "train_tokens_per_second": 8108.56 + }, + { + "epoch": 1.747017892644135, + "grad_norm": 0.6043440760128012, + "learning_rate": 3.9112808649119703e-07, + "loss": 0.3270209312438965, + "num_input_tokens_seen": 395711320, + "step": 3515, + "train_runtime": 48793.5843, + "train_tokens_per_second": 8109.905 + }, + { + "epoch": 1.749502982107356, + "grad_norm": 0.6378112336797943, + "learning_rate": 3.8359563007838496e-07, + "loss": 0.32279019355773925, + "num_input_tokens_seen": 396251040, + "step": 3520, + "train_runtime": 48854.789, + "train_tokens_per_second": 8110.792 + }, + { + "epoch": 1.7519880715705765, + "grad_norm": 0.6369477982846655, + "learning_rate": 3.7613351767234354e-07, + "loss": 0.31135079860687254, + "num_input_tokens_seen": 396820104, + "step": 3525, + "train_runtime": 48917.7511, + "train_tokens_per_second": 8111.986 + }, + { + "epoch": 1.7544731610337974, + "grad_norm": 0.954112307845048, + "learning_rate": 3.687418629795081e-07, + "loss": 0.3224629402160645, + "num_input_tokens_seen": 397336776, + "step": 3530, + "train_runtime": 48988.9925, + "train_tokens_per_second": 8110.736 + }, + { + "epoch": 1.7569582504970178, + "grad_norm": 0.6158217718007925, + "learning_rate": 3.6142077863269066e-07, + "loss": 0.31424548625946047, + "num_input_tokens_seen": 397893768, + "step": 3535, + "train_runtime": 49063.7727, + "train_tokens_per_second": 8109.726 + }, + { + "epoch": 1.7594433399602387, + "grad_norm": 0.6008782708684459, + "learning_rate": 3.541703761893645e-07, + "loss": 0.3181297302246094, + "num_input_tokens_seen": 398467408, + "step": 3540, + "train_runtime": 49133.5932, + "train_tokens_per_second": 8109.877 + }, + { + "epoch": 1.7619284294234592, + "grad_norm": 0.575324387955075, + "learning_rate": 3.4699076612996427e-07, + "loss": 0.3093132495880127, + "num_input_tokens_seen": 399031392, + "step": 3545, + "train_runtime": 49199.7287, + "train_tokens_per_second": 8110.439 + }, + { + "epoch": 1.76441351888668, + "grad_norm": 0.6580463615307782, + "learning_rate": 3.3988205785620064e-07, + "loss": 0.29964122772216795, + "num_input_tokens_seen": 399600416, + "step": 3550, + "train_runtime": 49274.7328, + "train_tokens_per_second": 8109.641 + }, + { + "epoch": 1.7668986083499005, + "grad_norm": 0.5796225285788789, + "learning_rate": 3.32844359689396e-07, + "loss": 0.31161746978759763, + "num_input_tokens_seen": 400149088, + "step": 3555, + "train_runtime": 49349.3871, + "train_tokens_per_second": 8108.492 + }, + { + "epoch": 1.7693836978131214, + "grad_norm": 0.7020480430779626, + "learning_rate": 3.2587777886883165e-07, + "loss": 0.30706286430358887, + "num_input_tokens_seen": 400729624, + "step": 3560, + "train_runtime": 49419.2744, + "train_tokens_per_second": 8108.772 + }, + { + "epoch": 1.771868787276342, + "grad_norm": 0.6084397848592192, + "learning_rate": 3.189824215501142e-07, + "loss": 0.3256417751312256, + "num_input_tokens_seen": 401300176, + "step": 3565, + "train_runtime": 49490.4511, + "train_tokens_per_second": 8108.638 + }, + { + "epoch": 1.7743538767395626, + "grad_norm": 0.6472171692593617, + "learning_rate": 3.1215839280355877e-07, + "loss": 0.3079630613327026, + "num_input_tokens_seen": 401854248, + "step": 3570, + "train_runtime": 49555.7031, + "train_tokens_per_second": 8109.142 + }, + { + "epoch": 1.7768389662027833, + "grad_norm": 0.6074019559547069, + "learning_rate": 3.05405796612589e-07, + "loss": 0.30272536277770995, + "num_input_tokens_seen": 402407832, + "step": 3575, + "train_runtime": 49622.8388, + "train_tokens_per_second": 8109.327 + }, + { + "epoch": 1.779324055666004, + "grad_norm": 0.5789364339576805, + "learning_rate": 2.987247358721479e-07, + "loss": 0.31909980773925783, + "num_input_tokens_seen": 402986560, + "step": 3580, + "train_runtime": 49684.3161, + "train_tokens_per_second": 8110.941 + }, + { + "epoch": 1.7818091451292246, + "grad_norm": 0.6410885463242871, + "learning_rate": 2.9211531238713454e-07, + "loss": 0.3228114604949951, + "num_input_tokens_seen": 403508752, + "step": 3585, + "train_runtime": 49747.7261, + "train_tokens_per_second": 8111.099 + }, + { + "epoch": 1.7842942345924453, + "grad_norm": 0.5797060718294117, + "learning_rate": 2.855776268708527e-07, + "loss": 0.31345381736755373, + "num_input_tokens_seen": 404045592, + "step": 3590, + "train_runtime": 49817.2988, + "train_tokens_per_second": 8110.548 + }, + { + "epoch": 1.786779324055666, + "grad_norm": 0.6542959818567567, + "learning_rate": 2.7911177894347274e-07, + "loss": 0.3365085363388062, + "num_input_tokens_seen": 404605104, + "step": 3595, + "train_runtime": 49884.9066, + "train_tokens_per_second": 8110.772 + }, + { + "epoch": 1.7892644135188867, + "grad_norm": 1.2815024607675278, + "learning_rate": 2.727178671305175e-07, + "loss": 0.3168356895446777, + "num_input_tokens_seen": 405187464, + "step": 3600, + "train_runtime": 49963.2468, + "train_tokens_per_second": 8109.71 + }, + { + "epoch": 1.7917495029821073, + "grad_norm": 0.6051938797925253, + "learning_rate": 2.663959888613582e-07, + "loss": 0.3264192581176758, + "num_input_tokens_seen": 405763312, + "step": 3605, + "train_runtime": 50033.6017, + "train_tokens_per_second": 8109.816 + }, + { + "epoch": 1.794234592445328, + "grad_norm": 0.6193660951985073, + "learning_rate": 2.601462404677313e-07, + "loss": 0.3080559730529785, + "num_input_tokens_seen": 406331704, + "step": 3610, + "train_runtime": 50105.1471, + "train_tokens_per_second": 8109.58 + }, + { + "epoch": 1.7967196819085487, + "grad_norm": 0.5969535709886561, + "learning_rate": 2.539687171822691e-07, + "loss": 0.32686400413513184, + "num_input_tokens_seen": 406939152, + "step": 3615, + "train_runtime": 50177.5848, + "train_tokens_per_second": 8109.979 + }, + { + "epoch": 1.7992047713717694, + "grad_norm": 0.6525875511722198, + "learning_rate": 2.478635131370499e-07, + "loss": 0.31873512268066406, + "num_input_tokens_seen": 407511680, + "step": 3620, + "train_runtime": 50251.5998, + "train_tokens_per_second": 8109.427 + }, + { + "epoch": 1.80168986083499, + "grad_norm": 0.6096303980317259, + "learning_rate": 2.4183072136216515e-07, + "loss": 0.3121727466583252, + "num_input_tokens_seen": 408111912, + "step": 3625, + "train_runtime": 50325.5541, + "train_tokens_per_second": 8109.437 + }, + { + "epoch": 1.8041749502982107, + "grad_norm": 0.7304718041065009, + "learning_rate": 2.3587043378429818e-07, + "loss": 0.3239481210708618, + "num_input_tokens_seen": 408686512, + "step": 3630, + "train_runtime": 50397.4741, + "train_tokens_per_second": 8109.266 + }, + { + "epoch": 1.8066600397614314, + "grad_norm": 0.6044788674646036, + "learning_rate": 2.2998274122532581e-07, + "loss": 0.3261247158050537, + "num_input_tokens_seen": 409263600, + "step": 3635, + "train_runtime": 50468.3486, + "train_tokens_per_second": 8109.312 + }, + { + "epoch": 1.809145129224652, + "grad_norm": 0.5848935709269248, + "learning_rate": 2.2416773340093456e-07, + "loss": 0.31814463138580323, + "num_input_tokens_seen": 409832744, + "step": 3640, + "train_runtime": 50537.4353, + "train_tokens_per_second": 8109.488 + }, + { + "epoch": 1.8116302186878728, + "grad_norm": 0.6227853093555468, + "learning_rate": 2.1842549891925302e-07, + "loss": 0.3035116672515869, + "num_input_tokens_seen": 410393864, + "step": 3645, + "train_runtime": 50606.2912, + "train_tokens_per_second": 8109.542 + }, + { + "epoch": 1.8141153081510935, + "grad_norm": 0.5447675559696712, + "learning_rate": 2.1275612527949907e-07, + "loss": 0.3397731065750122, + "num_input_tokens_seen": 410978720, + "step": 3650, + "train_runtime": 50678.2734, + "train_tokens_per_second": 8109.564 + }, + { + "epoch": 1.8166003976143141, + "grad_norm": 0.6012647187882215, + "learning_rate": 2.0715969887065312e-07, + "loss": 0.33132114410400393, + "num_input_tokens_seen": 411557448, + "step": 3655, + "train_runtime": 50750.2668, + "train_tokens_per_second": 8109.464 + }, + { + "epoch": 1.8190854870775348, + "grad_norm": 0.6836385736858126, + "learning_rate": 2.016363049701353e-07, + "loss": 0.34366092681884763, + "num_input_tokens_seen": 412136216, + "step": 3660, + "train_runtime": 50826.6009, + "train_tokens_per_second": 8108.672 + }, + { + "epoch": 1.8215705765407555, + "grad_norm": 0.6593978354726521, + "learning_rate": 1.9618602774250815e-07, + "loss": 0.29894418716430665, + "num_input_tokens_seen": 412658480, + "step": 3665, + "train_runtime": 50892.4073, + "train_tokens_per_second": 8108.449 + }, + { + "epoch": 1.8240556660039762, + "grad_norm": 0.69340388073698, + "learning_rate": 1.9080895023819657e-07, + "loss": 0.31429295539855956, + "num_input_tokens_seen": 413228216, + "step": 3670, + "train_runtime": 50957.5303, + "train_tokens_per_second": 8109.267 + }, + { + "epoch": 1.8265407554671969, + "grad_norm": 0.6847454161706115, + "learning_rate": 1.8550515439221762e-07, + "loss": 0.3116039752960205, + "num_input_tokens_seen": 413777432, + "step": 3675, + "train_runtime": 51035.9932, + "train_tokens_per_second": 8107.561 + }, + { + "epoch": 1.8290258449304175, + "grad_norm": 0.5957346347384566, + "learning_rate": 1.8027472102293663e-07, + "loss": 0.3293259620666504, + "num_input_tokens_seen": 414362544, + "step": 3680, + "train_runtime": 51109.4523, + "train_tokens_per_second": 8107.356 + }, + { + "epoch": 1.8315109343936382, + "grad_norm": 0.6202225580773381, + "learning_rate": 1.7511772983083208e-07, + "loss": 0.31985466480255126, + "num_input_tokens_seen": 414907208, + "step": 3685, + "train_runtime": 51176.1469, + "train_tokens_per_second": 8107.434 + }, + { + "epoch": 1.833996023856859, + "grad_norm": 0.6613830249696063, + "learning_rate": 1.7003425939728425e-07, + "loss": 0.30423722267150877, + "num_input_tokens_seen": 415470920, + "step": 3690, + "train_runtime": 51244.6428, + "train_tokens_per_second": 8107.597 + }, + { + "epoch": 1.8364811133200796, + "grad_norm": 0.662347103466514, + "learning_rate": 1.6502438718337454e-07, + "loss": 0.33441877365112305, + "num_input_tokens_seen": 416045344, + "step": 3695, + "train_runtime": 51302.7375, + "train_tokens_per_second": 8109.613 + }, + { + "epoch": 1.8389662027833003, + "grad_norm": 0.5723349570850523, + "learning_rate": 1.600881895287082e-07, + "loss": 0.3084590435028076, + "num_input_tokens_seen": 416607008, + "step": 3700, + "train_runtime": 51383.2485, + "train_tokens_per_second": 8107.837 + }, + { + "epoch": 1.841451292246521, + "grad_norm": 0.70682167344373, + "learning_rate": 1.5522574165024896e-07, + "loss": 0.32577691078186033, + "num_input_tokens_seen": 417111792, + "step": 3705, + "train_runtime": 51445.3761, + "train_tokens_per_second": 8107.858 + }, + { + "epoch": 1.8439363817097414, + "grad_norm": 0.638387994913655, + "learning_rate": 1.5043711764117396e-07, + "loss": 0.32839040756225585, + "num_input_tokens_seen": 417668808, + "step": 3710, + "train_runtime": 51524.497, + "train_tokens_per_second": 8106.218 + }, + { + "epoch": 1.8464214711729623, + "grad_norm": 0.6320838305497825, + "learning_rate": 1.4572239046974412e-07, + "loss": 0.326462984085083, + "num_input_tokens_seen": 418242704, + "step": 3715, + "train_runtime": 51597.2625, + "train_tokens_per_second": 8105.909 + }, + { + "epoch": 1.8489065606361827, + "grad_norm": 0.6172767261414444, + "learning_rate": 1.4108163197819325e-07, + "loss": 0.3121402502059937, + "num_input_tokens_seen": 418813368, + "step": 3720, + "train_runtime": 51670.7086, + "train_tokens_per_second": 8105.431 + }, + { + "epoch": 1.8513916500994037, + "grad_norm": 0.5885042274412391, + "learning_rate": 1.365149128816323e-07, + "loss": 0.3086491584777832, + "num_input_tokens_seen": 419363464, + "step": 3725, + "train_runtime": 51749.031, + "train_tokens_per_second": 8103.794 + }, + { + "epoch": 1.853876739562624, + "grad_norm": 0.6464968176266476, + "learning_rate": 1.3202230276697147e-07, + "loss": 0.308359432220459, + "num_input_tokens_seen": 419904248, + "step": 3730, + "train_runtime": 51818.3929, + "train_tokens_per_second": 8103.382 + }, + { + "epoch": 1.856361829025845, + "grad_norm": 0.6203401801102979, + "learning_rate": 1.2760387009186137e-07, + "loss": 0.3032233238220215, + "num_input_tokens_seen": 420469704, + "step": 3735, + "train_runtime": 51892.2055, + "train_tokens_per_second": 8102.753 + }, + { + "epoch": 1.8588469184890655, + "grad_norm": 0.578193416508872, + "learning_rate": 1.2325968218364904e-07, + "loss": 0.3259268760681152, + "num_input_tokens_seen": 421048464, + "step": 3740, + "train_runtime": 51965.3997, + "train_tokens_per_second": 8102.477 + }, + { + "epoch": 1.8613320079522864, + "grad_norm": 0.6564446926594277, + "learning_rate": 1.1898980523835146e-07, + "loss": 0.31359918117523194, + "num_input_tokens_seen": 421590720, + "step": 3745, + "train_runtime": 52036.9232, + "train_tokens_per_second": 8101.761 + }, + { + "epoch": 1.8638170974155068, + "grad_norm": 0.6723632415016808, + "learning_rate": 1.1479430431964811e-07, + "loss": 0.3024056196212769, + "num_input_tokens_seen": 422166152, + "step": 3750, + "train_runtime": 52106.9209, + "train_tokens_per_second": 8101.921 + }, + { + "epoch": 1.8663021868787277, + "grad_norm": 0.7010487948506207, + "learning_rate": 1.1067324335788832e-07, + "loss": 0.3258035182952881, + "num_input_tokens_seen": 422704976, + "step": 3755, + "train_runtime": 52179.0455, + "train_tokens_per_second": 8101.048 + }, + { + "epoch": 1.8687872763419482, + "grad_norm": 0.5259056957094247, + "learning_rate": 1.0662668514911878e-07, + "loss": 0.30820698738098146, + "num_input_tokens_seen": 423266112, + "step": 3760, + "train_runtime": 52245.6002, + "train_tokens_per_second": 8101.469 + }, + { + "epoch": 1.871272365805169, + "grad_norm": 0.5894398781201959, + "learning_rate": 1.0265469135412264e-07, + "loss": 0.32106695175170896, + "num_input_tokens_seen": 423829936, + "step": 3765, + "train_runtime": 52317.0821, + "train_tokens_per_second": 8101.177 + }, + { + "epoch": 1.8737574552683895, + "grad_norm": 0.6182488266998701, + "learning_rate": 9.875732249748637e-08, + "loss": 0.2933901309967041, + "num_input_tokens_seen": 424381848, + "step": 3770, + "train_runtime": 52383.3014, + "train_tokens_per_second": 8101.472 + }, + { + "epoch": 1.8762425447316105, + "grad_norm": 0.7357251206054523, + "learning_rate": 9.493463796667213e-08, + "loss": 0.31568107604980467, + "num_input_tokens_seen": 424915000, + "step": 3775, + "train_runtime": 52436.3143, + "train_tokens_per_second": 8103.449 + }, + { + "epoch": 1.878727634194831, + "grad_norm": 0.6239025227590759, + "learning_rate": 9.118669601111518e-08, + "loss": 0.3163498878479004, + "num_input_tokens_seen": 425492016, + "step": 3780, + "train_runtime": 52512.7499, + "train_tokens_per_second": 8102.642 + }, + { + "epoch": 1.8812127236580518, + "grad_norm": 0.635501912995116, + "learning_rate": 8.751355374133464e-08, + "loss": 0.32223618030548096, + "num_input_tokens_seen": 426063792, + "step": 3785, + "train_runtime": 52591.2186, + "train_tokens_per_second": 8101.425 + }, + { + "epoch": 1.8836978131212723, + "grad_norm": 0.6017722497997513, + "learning_rate": 8.39152671280663e-08, + "loss": 0.30569891929626464, + "num_input_tokens_seen": 426632272, + "step": 3790, + "train_runtime": 52658.884, + "train_tokens_per_second": 8101.81 + }, + { + "epoch": 1.8861829025844932, + "grad_norm": 0.6258726032477966, + "learning_rate": 8.039189100140676e-08, + "loss": 0.3324729442596436, + "num_input_tokens_seen": 427193136, + "step": 3795, + "train_runtime": 52725.8006, + "train_tokens_per_second": 8102.165 + }, + { + "epoch": 1.8886679920477136, + "grad_norm": 0.7077504219313828, + "learning_rate": 7.694347904997846e-08, + "loss": 0.3002906799316406, + "num_input_tokens_seen": 427724536, + "step": 3800, + "train_runtime": 52788.6314, + "train_tokens_per_second": 8102.588 + }, + { + "epoch": 1.8911530815109345, + "grad_norm": 0.6541487515084592, + "learning_rate": 7.35700838201131e-08, + "loss": 0.31304187774658204, + "num_input_tokens_seen": 428321072, + "step": 3805, + "train_runtime": 52856.1143, + "train_tokens_per_second": 8103.529 + }, + { + "epoch": 1.893638170974155, + "grad_norm": 0.5756554898673986, + "learning_rate": 7.027175671505015e-08, + "loss": 0.2898926019668579, + "num_input_tokens_seen": 428865760, + "step": 3810, + "train_runtime": 52928.9631, + "train_tokens_per_second": 8102.667 + }, + { + "epoch": 1.896123260437376, + "grad_norm": 0.5978032709617255, + "learning_rate": 6.70485479941524e-08, + "loss": 0.3094179630279541, + "num_input_tokens_seen": 429437824, + "step": 3815, + "train_runtime": 53003.0774, + "train_tokens_per_second": 8102.13 + }, + { + "epoch": 1.8986083499005963, + "grad_norm": 0.6324533668488967, + "learning_rate": 6.390050677214099e-08, + "loss": 0.31641445159912107, + "num_input_tokens_seen": 429984680, + "step": 3820, + "train_runtime": 53080.8996, + "train_tokens_per_second": 8100.554 + }, + { + "epoch": 1.9010934393638173, + "grad_norm": 0.6281322370085697, + "learning_rate": 6.082768101834835e-08, + "loss": 0.3233936786651611, + "num_input_tokens_seen": 430552104, + "step": 3825, + "train_runtime": 53149.2843, + "train_tokens_per_second": 8100.807 + }, + { + "epoch": 1.9035785288270377, + "grad_norm": 0.6084220666068709, + "learning_rate": 5.783011755598533e-08, + "loss": 0.3093274593353271, + "num_input_tokens_seen": 431082544, + "step": 3830, + "train_runtime": 53222.2549, + "train_tokens_per_second": 8099.667 + }, + { + "epoch": 1.9060636182902586, + "grad_norm": 0.6412867843786071, + "learning_rate": 5.490786206142851e-08, + "loss": 0.3198287010192871, + "num_input_tokens_seen": 431664288, + "step": 3835, + "train_runtime": 53297.683, + "train_tokens_per_second": 8099.119 + }, + { + "epoch": 1.908548707753479, + "grad_norm": 0.6584250619499648, + "learning_rate": 5.2060959063524595e-08, + "loss": 0.31406712532043457, + "num_input_tokens_seen": 432216792, + "step": 3840, + "train_runtime": 53369.0577, + "train_tokens_per_second": 8098.64 + }, + { + "epoch": 1.9110337972166997, + "grad_norm": 0.5977325891706663, + "learning_rate": 4.928945194291046e-08, + "loss": 0.31024718284606934, + "num_input_tokens_seen": 432818848, + "step": 3845, + "train_runtime": 53437.1322, + "train_tokens_per_second": 8099.59 + }, + { + "epoch": 1.9135188866799204, + "grad_norm": 0.6239562994508425, + "learning_rate": 4.6593382931353074e-08, + "loss": 0.3158505916595459, + "num_input_tokens_seen": 433377496, + "step": 3850, + "train_runtime": 53513.4003, + "train_tokens_per_second": 8098.485 + }, + { + "epoch": 1.916003976143141, + "grad_norm": 0.6058128962441687, + "learning_rate": 4.397279311110725e-08, + "loss": 0.33526182174682617, + "num_input_tokens_seen": 433969152, + "step": 3855, + "train_runtime": 53588.162, + "train_tokens_per_second": 8098.228 + }, + { + "epoch": 1.9184890656063618, + "grad_norm": 0.5664050787730245, + "learning_rate": 4.142772241428727e-08, + "loss": 0.29236435890197754, + "num_input_tokens_seen": 434536456, + "step": 3860, + "train_runtime": 53660.5347, + "train_tokens_per_second": 8097.878 + }, + { + "epoch": 1.9209741550695825, + "grad_norm": 0.5784360930172237, + "learning_rate": 3.8958209622259044e-08, + "loss": 0.3196124076843262, + "num_input_tokens_seen": 435084600, + "step": 3865, + "train_runtime": 53736.1325, + "train_tokens_per_second": 8096.686 + }, + { + "epoch": 1.9234592445328031, + "grad_norm": 0.5839648770294493, + "learning_rate": 3.6564292365051655e-08, + "loss": 0.3151907205581665, + "num_input_tokens_seen": 435653488, + "step": 3870, + "train_runtime": 53796.1997, + "train_tokens_per_second": 8098.221 + }, + { + "epoch": 1.9259443339960238, + "grad_norm": 0.6456021802730554, + "learning_rate": 3.424600712077952e-08, + "loss": 0.3286634922027588, + "num_input_tokens_seen": 436181976, + "step": 3875, + "train_runtime": 53866.8479, + "train_tokens_per_second": 8097.41 + }, + { + "epoch": 1.9284294234592445, + "grad_norm": 0.6400705179023234, + "learning_rate": 3.200338921509061e-08, + "loss": 0.3336653709411621, + "num_input_tokens_seen": 436715752, + "step": 3880, + "train_runtime": 53936.6618, + "train_tokens_per_second": 8096.826 + }, + { + "epoch": 1.9309145129224652, + "grad_norm": 0.5760788911560203, + "learning_rate": 2.983647282062574e-08, + "loss": 0.2989939212799072, + "num_input_tokens_seen": 437297232, + "step": 3885, + "train_runtime": 54007.8973, + "train_tokens_per_second": 8096.913 + }, + { + "epoch": 1.9333996023856859, + "grad_norm": 0.6848560180363625, + "learning_rate": 2.774529095649847e-08, + "loss": 0.31984658241271974, + "num_input_tokens_seen": 437850160, + "step": 3890, + "train_runtime": 54078.7558, + "train_tokens_per_second": 8096.528 + }, + { + "epoch": 1.9358846918489065, + "grad_norm": 0.5800435059818678, + "learning_rate": 2.572987548779271e-08, + "loss": 0.31107923984527586, + "num_input_tokens_seen": 438414800, + "step": 3895, + "train_runtime": 54153.565, + "train_tokens_per_second": 8095.77 + }, + { + "epoch": 1.9383697813121272, + "grad_norm": 0.6461170118148103, + "learning_rate": 2.379025712507588e-08, + "loss": 0.29071803092956544, + "num_input_tokens_seen": 438939800, + "step": 3900, + "train_runtime": 54218.3773, + "train_tokens_per_second": 8095.775 + }, + { + "epoch": 1.940854870775348, + "grad_norm": 2.174113690716615, + "learning_rate": 2.1926465423932618e-08, + "loss": 0.3068253755569458, + "num_input_tokens_seen": 439451128, + "step": 3905, + "train_runtime": 54275.7897, + "train_tokens_per_second": 8096.633 + }, + { + "epoch": 1.9433399602385686, + "grad_norm": 0.6259125537863874, + "learning_rate": 2.013852878451239e-08, + "loss": 0.3135919332504272, + "num_input_tokens_seen": 440013664, + "step": 3910, + "train_runtime": 54344.0126, + "train_tokens_per_second": 8096.82 + }, + { + "epoch": 1.9458250497017893, + "grad_norm": 0.6403485689413579, + "learning_rate": 1.842647445109813e-08, + "loss": 0.30338706970214846, + "num_input_tokens_seen": 440560408, + "step": 3915, + "train_runtime": 54412.8301, + "train_tokens_per_second": 8096.627 + }, + { + "epoch": 1.94831013916501, + "grad_norm": 0.6869898078408394, + "learning_rate": 1.6790328511691044e-08, + "loss": 0.3187436580657959, + "num_input_tokens_seen": 441127552, + "step": 3920, + "train_runtime": 54491.5364, + "train_tokens_per_second": 8095.341 + }, + { + "epoch": 1.9507952286282306, + "grad_norm": 0.6499456526952925, + "learning_rate": 1.5230115897612585e-08, + "loss": 0.32091774940490725, + "num_input_tokens_seen": 441681320, + "step": 3925, + "train_runtime": 54557.3983, + "train_tokens_per_second": 8095.718 + }, + { + "epoch": 1.9532803180914513, + "grad_norm": 0.6271896828779471, + "learning_rate": 1.3745860383124198e-08, + "loss": 0.3291923999786377, + "num_input_tokens_seen": 442230784, + "step": 3930, + "train_runtime": 54634.1668, + "train_tokens_per_second": 8094.4 + }, + { + "epoch": 1.955765407554672, + "grad_norm": 0.6356305890792343, + "learning_rate": 1.2337584585067063e-08, + "loss": 0.3415233850479126, + "num_input_tokens_seen": 442800488, + "step": 3935, + "train_runtime": 54704.5607, + "train_tokens_per_second": 8094.398 + }, + { + "epoch": 1.9582504970178927, + "grad_norm": 1.1531063186115513, + "learning_rate": 1.1005309962514032e-08, + "loss": 0.2975748300552368, + "num_input_tokens_seen": 443363544, + "step": 3940, + "train_runtime": 54765.6129, + "train_tokens_per_second": 8095.656 + }, + { + "epoch": 1.9607355864811133, + "grad_norm": 0.6294282912498614, + "learning_rate": 9.749056816446002e-09, + "loss": 0.3283940315246582, + "num_input_tokens_seen": 443953512, + "step": 3945, + "train_runtime": 54829.4928, + "train_tokens_per_second": 8096.984 + }, + { + "epoch": 1.963220675944334, + "grad_norm": 0.6880085169187019, + "learning_rate": 8.568844289440492e-09, + "loss": 0.31568002700805664, + "num_input_tokens_seen": 444481688, + "step": 3950, + "train_runtime": 54898.7301, + "train_tokens_per_second": 8096.393 + }, + { + "epoch": 1.9657057654075547, + "grad_norm": 0.6644794816377358, + "learning_rate": 7.464690365381333e-09, + "loss": 0.31000685691833496, + "num_input_tokens_seen": 445082712, + "step": 3955, + "train_runtime": 54971.9576, + "train_tokens_per_second": 8096.541 + }, + { + "epoch": 1.9681908548707754, + "grad_norm": 0.6000395164233143, + "learning_rate": 6.436611869183873e-09, + "loss": 0.3138153314590454, + "num_input_tokens_seen": 445656144, + "step": 3960, + "train_runtime": 55052.8974, + "train_tokens_per_second": 8095.053 + }, + { + "epoch": 1.970675944333996, + "grad_norm": 0.6301126005040312, + "learning_rate": 5.4846244665368585e-09, + "loss": 0.3102699279785156, + "num_input_tokens_seen": 446179808, + "step": 3965, + "train_runtime": 55116.6747, + "train_tokens_per_second": 8095.187 + }, + { + "epoch": 1.9731610337972167, + "grad_norm": 0.5582579784540138, + "learning_rate": 4.60874266366873e-09, + "loss": 0.30883145332336426, + "num_input_tokens_seen": 446768608, + "step": 3970, + "train_runtime": 55193.9879, + "train_tokens_per_second": 8094.516 + }, + { + "epoch": 1.9756461232604374, + "grad_norm": 0.6622077397215657, + "learning_rate": 3.808979807120028e-09, + "loss": 0.3112310886383057, + "num_input_tokens_seen": 447330808, + "step": 3975, + "train_runtime": 55270.2682, + "train_tokens_per_second": 8093.516 + }, + { + "epoch": 1.978131212723658, + "grad_norm": 0.5965657910872623, + "learning_rate": 3.0853480835463247e-09, + "loss": 0.30730018615722654, + "num_input_tokens_seen": 447906288, + "step": 3980, + "train_runtime": 55332.5303, + "train_tokens_per_second": 8094.809 + }, + { + "epoch": 1.9806163021868786, + "grad_norm": 0.706372130063471, + "learning_rate": 2.4378585195278248e-09, + "loss": 0.3241956949234009, + "num_input_tokens_seen": 448454032, + "step": 3985, + "train_runtime": 55399.071, + "train_tokens_per_second": 8094.974 + }, + { + "epoch": 1.9831013916500995, + "grad_norm": 0.6189513348476302, + "learning_rate": 1.8665209814028306e-09, + "loss": 0.3116312980651855, + "num_input_tokens_seen": 449014192, + "step": 3990, + "train_runtime": 55473.7007, + "train_tokens_per_second": 8094.181 + }, + { + "epoch": 1.98558648111332, + "grad_norm": 0.6102850220738272, + "learning_rate": 1.3713441751189716e-09, + "loss": 0.31803319454193113, + "num_input_tokens_seen": 449553072, + "step": 3995, + "train_runtime": 55554.6096, + "train_tokens_per_second": 8092.093 + }, + { + "epoch": 1.9880715705765408, + "grad_norm": 0.6398210113456437, + "learning_rate": 9.523356460983124e-10, + "loss": 0.3097801208496094, + "num_input_tokens_seen": 450114400, + "step": 4000, + "train_runtime": 55623.4485, + "train_tokens_per_second": 8092.17 + }, + { + "epoch": 1.9905566600397613, + "grad_norm": 0.6408413664259602, + "learning_rate": 6.095017791235558e-10, + "loss": 0.31991100311279297, + "num_input_tokens_seen": 450689584, + "step": 4005, + "train_runtime": 55698.2716, + "train_tokens_per_second": 8091.626 + }, + { + "epoch": 1.9930417495029822, + "grad_norm": 0.6049507005761956, + "learning_rate": 3.428477982408973e-10, + "loss": 0.31697897911071776, + "num_input_tokens_seen": 451282784, + "step": 4010, + "train_runtime": 55762.4852, + "train_tokens_per_second": 8092.946 + }, + { + "epoch": 1.9955268389662026, + "grad_norm": 0.6516719328014087, + "learning_rate": 1.5237776667953452e-10, + "loss": 0.3139082908630371, + "num_input_tokens_seen": 451809152, + "step": 4015, + "train_runtime": 55817.6448, + "train_tokens_per_second": 8094.379 + }, + { + "epoch": 1.9980119284294235, + "grad_norm": 0.6167992849095614, + "learning_rate": 3.8094586789494405e-11, + "loss": 0.32825708389282227, + "num_input_tokens_seen": 452362664, + "step": 4020, + "train_runtime": 55886.0311, + "train_tokens_per_second": 8094.378 + }, + { + "epoch": 2.0, + "num_input_tokens_seen": 452805312, + "step": 4024, + "total_flos": 837392786456576.0, + "train_loss": 0.35741217940986514, + "train_runtime": 55966.3522, + "train_samples_per_second": 0.575, + "train_steps_per_second": 0.072 + } + ], + "logging_steps": 5, + "max_steps": 4024, + "num_input_tokens_seen": 452805312, + "num_train_epochs": 2, + "save_steps": 0, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 837392786456576.0, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +}