[data] 31 shards from data/fwedu [data] skipping 1 shard(s) smaller than seq_len+1 [run] model=768d/12L steps=5639 bs=40x13 seq=1024 compile=True ckpt=checkpoints/base_124m [warn] wandb init failed: No API key configured. Use `wandb login` to log in. step 10 | loss 9.8265 | lr 2.20e-05 | gnorm 3.50 | 174745 tok/s | mfu 31.7% step 20 | loss 9.3561 | lr 4.20e-05 | gnorm 2.16 | 174359 tok/s | mfu 31.7% step 30 | loss 8.7749 | lr 6.20e-05 | gnorm 1.92 | 174038 tok/s | mfu 31.7% step 40 | loss 8.2169 | lr 8.20e-05 | gnorm 1.45 | 173957 tok/s | mfu 31.6% step 50 | loss 7.6476 | lr 1.02e-04 | gnorm 1.50 | 173862 tok/s | mfu 31.6% step 60 | loss 7.2370 | lr 1.22e-04 | gnorm 0.85 | 174070 tok/s | mfu 31.6% step 70 | loss 6.9458 | lr 1.42e-04 | gnorm 0.91 | 173818 tok/s | mfu 31.6% step 80 | loss 6.7593 | lr 1.62e-04 | gnorm 1.03 | 173627 tok/s | mfu 31.6% step 90 | loss 6.5086 | lr 1.82e-04 | gnorm 0.87 | 173518 tok/s | mfu 31.6% step 100 | loss 6.3741 | lr 2.02e-04 | gnorm 0.83 | 173456 tok/s | mfu 31.5% step 110 | loss 6.2535 | lr 2.22e-04 | gnorm 1.20 | 173503 tok/s | mfu 31.5% step 120 | loss 6.1320 | lr 2.42e-04 | gnorm 0.62 | 173432 tok/s | mfu 31.5% step 130 | loss 6.0450 | lr 2.62e-04 | gnorm 0.96 | 173492 tok/s | mfu 31.5% step 140 | loss 5.9466 | lr 2.82e-04 | gnorm 0.93 | 173449 tok/s | mfu 31.5% step 150 | loss 5.8292 | lr 3.02e-04 | gnorm 0.91 | 173305 tok/s | mfu 31.5% step 160 | loss 5.7645 | lr 3.22e-04 | gnorm 0.79 | 173344 tok/s | mfu 31.5% step 170 | loss 5.6466 | lr 3.42e-04 | gnorm 0.92 | 173378 tok/s | mfu 31.5% step 180 | loss 5.5709 | lr 3.62e-04 | gnorm 1.22 | 173368 tok/s | mfu 31.5% step 190 | loss 5.5270 | lr 3.82e-04 | gnorm 1.13 | 173423 tok/s | mfu 31.5% step 200 | loss 5.4553 | lr 4.02e-04 | gnorm 1.62 | 173320 tok/s | mfu 31.5% step 210 | loss 5.4066 | lr 4.22e-04 | gnorm 1.23 | 173388 tok/s | mfu 31.5% step 220 | loss 5.3553 | lr 4.42e-04 | gnorm 1.09 | 173262 tok/s | mfu 31.5% step 230 | loss 5.2903 | lr 4.62e-04 | gnorm 0.80 | 173156 tok/s | mfu 31.5% step 240 | loss 5.2215 | lr 4.82e-04 | gnorm 0.62 | 173211 tok/s | mfu 31.5% step 250 | loss 5.1072 | lr 5.02e-04 | gnorm 0.63 | 173466 tok/s | mfu 31.5% step 260 | loss 5.1226 | lr 5.22e-04 | gnorm 1.12 | 173371 tok/s | mfu 31.5% step 270 | loss 5.1226 | lr 5.42e-04 | gnorm 1.17 | 173296 tok/s | mfu 31.5% step 280 | loss 5.1561 | lr 5.62e-04 | gnorm 2.14 | 173372 tok/s | mfu 31.5% step 290 | loss 5.1147 | lr 5.82e-04 | gnorm 1.22 | 173463 tok/s | mfu 31.5% step 300 | loss 5.0332 | lr 6.00e-04 | gnorm 0.64 | 173534 tok/s | mfu 31.5% step 310 | loss 4.9140 | lr 6.00e-04 | gnorm 0.35 | 173597 tok/s | mfu 31.5% step 320 | loss 4.8308 | lr 6.00e-04 | gnorm 0.45 | 173398 tok/s | mfu 31.5% step 330 | loss 4.8269 | lr 6.00e-04 | gnorm 0.54 | 173454 tok/s | mfu 31.5% step 340 | loss 4.7694 | lr 6.00e-04 | gnorm 0.46 | 173435 tok/s | mfu 31.5% step 350 | loss 4.7318 | lr 6.00e-04 | gnorm 0.45 | 173487 tok/s | mfu 31.5% step 360 | loss 4.6773 | lr 6.00e-04 | gnorm 0.61 | 173461 tok/s | mfu 31.5% step 370 | loss 4.5925 | lr 6.00e-04 | gnorm 0.40 | 173479 tok/s | mfu 31.5% step 380 | loss 4.6017 | lr 6.00e-04 | gnorm 0.36 | 173425 tok/s | mfu 31.5% step 390 | loss 4.5676 | lr 6.00e-04 | gnorm 0.51 | 173429 tok/s | mfu 31.5% step 400 | loss 4.5094 | lr 6.00e-04 | gnorm 0.55 | 173391 tok/s | mfu 31.5% step 410 | loss 4.5039 | lr 5.99e-04 | gnorm 0.36 | 173382 tok/s | mfu 31.5% step 420 | loss 4.4344 | lr 5.99e-04 | gnorm 0.47 | 173436 tok/s | mfu 31.5% step 430 | loss 4.4581 | lr 5.99e-04 | gnorm 0.54 | 173357 tok/s | mfu 31.5% step 440 | loss 4.4103 | lr 5.99e-04 | gnorm 0.34 | 173409 tok/s | mfu 31.5% step 450 | loss 4.3761 | lr 5.99e-04 | gnorm 0.32 | 173429 tok/s | mfu 31.5% step 460 | loss 4.3263 | lr 5.99e-04 | gnorm 0.48 | 173425 tok/s | mfu 31.5% step 470 | loss 4.3295 | lr 5.99e-04 | gnorm 0.31 | 173237 tok/s | mfu 31.5% step 480 | loss 4.2446 | lr 5.98e-04 | gnorm 0.39 | 173365 tok/s | mfu 31.5% step 490 | loss 4.2233 | lr 5.98e-04 | gnorm 0.35 | 173552 tok/s | mfu 31.5% step 500 | loss 4.2209 | lr 5.98e-04 | gnorm 0.48 | 173487 tok/s | mfu 31.5% step 510 | loss 4.1671 | lr 5.98e-04 | gnorm 0.30 | 173657 tok/s | mfu 31.3% step 520 | loss 4.1593 | lr 5.98e-04 | gnorm 0.32 | 173431 tok/s | mfu 31.3% step 530 | loss 4.1421 | lr 5.98e-04 | gnorm 0.39 | 173606 tok/s | mfu 31.3% step 540 | loss 4.1394 | lr 5.97e-04 | gnorm 0.50 | 173374 tok/s | mfu 31.3% step 550 | loss 4.0783 | lr 5.97e-04 | gnorm 0.35 | 173594 tok/s | mfu 31.3% step 560 | loss 4.0745 | lr 5.97e-04 | gnorm 0.32 | 173710 tok/s | mfu 31.5% step 570 | loss 4.0410 | lr 5.97e-04 | gnorm 0.37 | 173641 tok/s | mfu 31.5% step 580 | loss 4.0625 | lr 5.96e-04 | gnorm 0.31 | 173553 tok/s | mfu 31.5% step 590 | loss 4.0133 | lr 5.96e-04 | gnorm 0.35 | 173529 tok/s | mfu 31.5% step 600 | loss 3.9927 | lr 5.96e-04 | gnorm 0.36 | 173447 tok/s | mfu 31.5% step 610 | loss 3.9924 | lr 5.96e-04 | gnorm 0.31 | 173635 tok/s | mfu 31.5% step 620 | loss 4.0056 | lr 5.95e-04 | gnorm 0.41 | 173398 tok/s | mfu 31.5% step 630 | loss 3.9355 | lr 5.95e-04 | gnorm 0.35 | 173222 tok/s | mfu 31.5% step 640 | loss 3.9328 | lr 5.95e-04 | gnorm 0.29 | 173493 tok/s | mfu 31.5% step 650 | loss 3.9270 | lr 5.94e-04 | gnorm 0.30 | 173366 tok/s | mfu 31.5% step 660 | loss 3.9115 | lr 5.94e-04 | gnorm 0.37 | 173363 tok/s | mfu 31.5% step 670 | loss 3.9089 | lr 5.94e-04 | gnorm 0.30 | 173170 tok/s | mfu 31.5% step 680 | loss 3.8828 | lr 5.93e-04 | gnorm 0.29 | 173608 tok/s | mfu 31.5% step 690 | loss 3.8999 | lr 5.93e-04 | gnorm 0.29 | 173190 tok/s | mfu 31.5% step 700 | loss 3.8645 | lr 5.93e-04 | gnorm 0.31 | 173409 tok/s | mfu 31.5% step 710 | loss 3.8645 | lr 5.92e-04 | gnorm 0.27 | 173419 tok/s | mfu 31.5% step 720 | loss 3.8769 | lr 5.92e-04 | gnorm 0.32 | 173360 tok/s | mfu 31.5% step 730 | loss 3.8379 | lr 5.91e-04 | gnorm 0.31 | 173121 tok/s | mfu 31.5% step 740 | loss 3.8498 | lr 5.91e-04 | gnorm 0.30 | 173292 tok/s | mfu 31.5% step 750 | loss 3.8398 | lr 5.91e-04 | gnorm 0.27 | 173145 tok/s | mfu 31.5% step 760 | loss 3.8061 | lr 5.90e-04 | gnorm 0.27 | 173230 tok/s | mfu 31.4% step 770 | loss 3.8217 | lr 5.90e-04 | gnorm 0.32 | 173326 tok/s | mfu 31.4% step 780 | loss 3.8219 | lr 5.89e-04 | gnorm 0.26 | 173392 tok/s | mfu 31.4% step 790 | loss 3.7749 | lr 5.89e-04 | gnorm 0.27 | 173433 tok/s | mfu 31.5% step 800 | loss 3.8009 | lr 5.88e-04 | gnorm 0.36 | 173398 tok/s | mfu 31.5% step 810 | loss 3.7763 | lr 5.88e-04 | gnorm 0.31 | 173267 tok/s | mfu 31.5% step 820 | loss 3.7880 | lr 5.87e-04 | gnorm 0.31 | 173422 tok/s | mfu 31.5% step 830 | loss 3.7902 | lr 5.87e-04 | gnorm 0.28 | 173448 tok/s | mfu 31.5% step 840 | loss 3.7731 | lr 5.86e-04 | gnorm 0.31 | 173472 tok/s | mfu 31.5% step 850 | loss 3.7557 | lr 5.86e-04 | gnorm 0.27 | 173364 tok/s | mfu 31.5% step 860 | loss 3.7446 | lr 5.85e-04 | gnorm 0.55 | 173255 tok/s | mfu 31.5% step 870 | loss 3.7218 | lr 5.85e-04 | gnorm 0.27 | 173191 tok/s | mfu 31.5% step 880 | loss 3.7398 | lr 5.84e-04 | gnorm 0.24 | 173182 tok/s | mfu 31.5% step 890 | loss 3.7320 | lr 5.84e-04 | gnorm 0.30 | 173247 tok/s | mfu 31.5% step 900 | loss 3.7105 | lr 5.83e-04 | gnorm 0.27 | 173355 tok/s | mfu 31.5% step 910 | loss 3.7190 | lr 5.83e-04 | gnorm 0.23 | 173240 tok/s | mfu 31.4% step 920 | loss 3.7390 | lr 5.82e-04 | gnorm 0.28 | 173179 tok/s | mfu 31.4% step 930 | loss 3.7001 | lr 5.82e-04 | gnorm 0.25 | 173309 tok/s | mfu 31.4% step 940 | loss 3.6808 | lr 5.81e-04 | gnorm 0.32 | 173300 tok/s | mfu 31.5% step 950 | loss 3.7354 | lr 5.80e-04 | gnorm 0.28 | 173389 tok/s | mfu 31.5% step 960 | loss 3.6919 | lr 5.80e-04 | gnorm 0.23 | 173433 tok/s | mfu 31.5% step 970 | loss 3.6790 | lr 5.79e-04 | gnorm 0.24 | 173161 tok/s | mfu 31.5% step 980 | loss 3.6655 | lr 5.79e-04 | gnorm 0.27 | 173404 tok/s | mfu 31.5% step 990 | loss 3.6723 | lr 5.78e-04 | gnorm 0.25 | 173246 tok/s | mfu 31.5% step 1000 | loss 3.6707 | lr 5.77e-04 | gnorm 0.29 | 173113 tok/s | mfu 31.5% step 1010 | loss 3.6919 | lr 5.77e-04 | gnorm 0.27 | 173272 tok/s | mfu 31.2% step 1020 | loss 3.6719 | lr 5.76e-04 | gnorm 0.26 | 173076 tok/s | mfu 31.2% step 1030 | loss 3.6632 | lr 5.75e-04 | gnorm 0.30 | 173383 tok/s | mfu 31.2% step 1040 | loss 3.6569 | lr 5.75e-04 | gnorm 0.24 | 173342 tok/s | mfu 31.2% step 1050 | loss 3.6864 | lr 5.74e-04 | gnorm 0.30 | 173282 tok/s | mfu 31.2% step 1060 | loss 3.6643 | lr 5.73e-04 | gnorm 0.23 | 173025 tok/s | mfu 31.5% step 1070 | loss 3.6426 | lr 5.73e-04 | gnorm 0.27 | 173281 tok/s | mfu 31.5% step 1080 | loss 3.6540 | lr 5.72e-04 | gnorm 0.30 | 173245 tok/s | mfu 31.4% step 1090 | loss 3.6333 | lr 5.71e-04 | gnorm 0.23 | 172975 tok/s | mfu 31.4% step 1100 | loss 3.6321 | lr 5.71e-04 | gnorm 0.23 | 173288 tok/s | mfu 31.4% step 1110 | loss 3.5915 | lr 5.70e-04 | gnorm 0.25 | 173264 tok/s | mfu 31.4% step 1120 | loss 3.6097 | lr 5.69e-04 | gnorm 0.25 | 173350 tok/s | mfu 31.4% step 1130 | loss 3.6190 | lr 5.68e-04 | gnorm 0.23 | 173422 tok/s | mfu 31.4% step 1140 | loss 3.6197 | lr 5.68e-04 | gnorm 0.23 | 173341 tok/s | mfu 31.5% step 1150 | loss 3.5962 | lr 5.67e-04 | gnorm 0.28 | 173281 tok/s | mfu 31.5% step 1160 | loss 3.6102 | lr 5.66e-04 | gnorm 0.22 | 173398 tok/s | mfu 31.5% step 1170 | loss 3.5972 | lr 5.65e-04 | gnorm 0.28 | 173281 tok/s | mfu 31.5% step 1180 | loss 3.5943 | lr 5.65e-04 | gnorm 0.24 | 173272 tok/s | mfu 31.5% step 1190 | loss 3.5970 | lr 5.64e-04 | gnorm 0.21 | 173309 tok/s | mfu 31.5% step 1200 | loss 3.6065 | lr 5.63e-04 | gnorm 0.26 | 173294 tok/s | mfu 31.5% step 1210 | loss 3.6015 | lr 5.62e-04 | gnorm 0.23 | 173217 tok/s | mfu 31.5% step 1220 | loss 3.5580 | lr 5.61e-04 | gnorm 0.24 | 173184 tok/s | mfu 31.5% step 1230 | loss 3.6094 | lr 5.61e-04 | gnorm 0.23 | 173217 tok/s | mfu 31.4% step 1240 | loss 3.5467 | lr 5.60e-04 | gnorm 0.25 | 173140 tok/s | mfu 31.4% step 1250 | loss 3.5512 | lr 5.59e-04 | gnorm 0.24 | 173218 tok/s | mfu 31.4% step 1260 | loss 3.5617 | lr 5.58e-04 | gnorm 0.21 | 172992 tok/s | mfu 31.4% step 1270 | loss 3.5578 | lr 5.57e-04 | gnorm 0.26 | 173212 tok/s | mfu 31.4% step 1280 | loss 3.5593 | lr 5.56e-04 | gnorm 0.23 | 173294 tok/s | mfu 31.4% step 1290 | loss 3.5549 | lr 5.55e-04 | gnorm 0.25 | 173159 tok/s | mfu 31.4% step 1300 | loss 3.5470 | lr 5.55e-04 | gnorm 0.22 | 173187 tok/s | mfu 31.4% step 1310 | loss 3.5672 | lr 5.54e-04 | gnorm 0.23 | 173447 tok/s | mfu 31.4% step 1320 | loss 3.5758 | lr 5.53e-04 | gnorm 0.23 | 173149 tok/s | mfu 31.4% step 1330 | loss 3.5461 | lr 5.52e-04 | gnorm 0.23 | 173116 tok/s | mfu 31.4% step 1340 | loss 3.5787 | lr 5.51e-04 | gnorm 0.24 | 173231 tok/s | mfu 31.4% step 1350 | loss 3.5375 | lr 5.50e-04 | gnorm 0.23 | 173110 tok/s | mfu 31.4% step 1360 | loss 3.5343 | lr 5.49e-04 | gnorm 0.22 | 173155 tok/s | mfu 31.4% step 1370 | loss 3.5253 | lr 5.48e-04 | gnorm 0.22 | 173118 tok/s | mfu 31.4% step 1380 | loss 3.5693 | lr 5.47e-04 | gnorm 0.28 | 173175 tok/s | mfu 31.4% step 1390 | loss 3.5406 | lr 5.46e-04 | gnorm 0.21 | 173170 tok/s | mfu 31.4% step 1400 | loss 3.5251 | lr 5.45e-04 | gnorm 0.23 | 173094 tok/s | mfu 31.4% step 1410 | loss 3.5486 | lr 5.44e-04 | gnorm 0.25 | 172953 tok/s | mfu 31.4% step 1420 | loss 3.5318 | lr 5.43e-04 | gnorm 0.25 | 173470 tok/s | mfu 31.5% step 1430 | loss 3.5348 | lr 5.42e-04 | gnorm 0.21 | 173400 tok/s | mfu 31.5% step 1440 | loss 3.5422 | lr 5.41e-04 | gnorm 0.23 | 173526 tok/s | mfu 31.5% step 1450 | loss 3.5321 | lr 5.41e-04 | gnorm 0.23 | 173429 tok/s | mfu 31.5% step 1460 | loss 3.5378 | lr 5.40e-04 | gnorm 0.28 | 173447 tok/s | mfu 31.5% step 1470 | loss 3.5094 | lr 5.39e-04 | gnorm 0.21 | 173478 tok/s | mfu 31.5% step 1480 | loss 3.5191 | lr 5.37e-04 | gnorm 0.23 | 173346 tok/s | mfu 31.5% step 1490 | loss 3.5223 | lr 5.36e-04 | gnorm 0.21 | 173434 tok/s | mfu 31.5% step 1500 | loss 3.5071 | lr 5.35e-04 | gnorm 0.24 | 173438 tok/s | mfu 31.5% step 1510 | loss 3.5212 | lr 5.34e-04 | gnorm 0.21 | 173312 tok/s | mfu 31.3% step 1520 | loss 3.5115 | lr 5.33e-04 | gnorm 0.24 | 173327 tok/s | mfu 31.3% step 1530 | loss 3.4894 | lr 5.32e-04 | gnorm 0.23 | 173377 tok/s | mfu 31.3% step 1540 | loss 3.5617 | lr 5.31e-04 | gnorm 0.23 | 173414 tok/s | mfu 31.3% step 1550 | loss 3.4803 | lr 5.30e-04 | gnorm 0.22 | 173360 tok/s | mfu 31.3% step 1560 | loss 3.5059 | lr 5.29e-04 | gnorm 0.22 | 173603 tok/s | mfu 31.5% step 1570 | loss 3.5131 | lr 5.28e-04 | gnorm 0.21 | 173510 tok/s | mfu 31.5% step 1580 | loss 3.4945 | lr 5.27e-04 | gnorm 0.23 | 173526 tok/s | mfu 31.5% step 1590 | loss 3.5082 | lr 5.26e-04 | gnorm 0.21 | 173517 tok/s | mfu 31.5% step 1600 | loss 3.5015 | lr 5.25e-04 | gnorm 0.23 | 173472 tok/s | mfu 31.5% step 1610 | loss 3.4867 | lr 5.24e-04 | gnorm 0.21 | 173246 tok/s | mfu 31.5% step 1620 | loss 3.4703 | lr 5.23e-04 | gnorm 0.21 | 173451 tok/s | mfu 31.5% step 1630 | loss 3.4757 | lr 5.21e-04 | gnorm 0.21 | 173354 tok/s | mfu 31.5% step 1640 | loss 3.4794 | lr 5.20e-04 | gnorm 0.20 | 173460 tok/s | mfu 31.5% step 1650 | loss 3.4699 | lr 5.19e-04 | gnorm 0.21 | 173315 tok/s | mfu 31.5% step 1660 | loss 3.5003 | lr 5.18e-04 | gnorm 0.22 | 173425 tok/s | mfu 31.5% step 1670 | loss 3.4627 | lr 5.17e-04 | gnorm 0.21 | 173451 tok/s | mfu 31.5% step 1680 | loss 3.4796 | lr 5.16e-04 | gnorm 0.23 | 173195 tok/s | mfu 31.5% step 1690 | loss 3.4555 | lr 5.15e-04 | gnorm 0.21 | 173398 tok/s | mfu 31.5% step 1700 | loss 3.4557 | lr 5.13e-04 | gnorm 0.23 | 173301 tok/s | mfu 31.5% step 1710 | loss 3.4219 | lr 5.12e-04 | gnorm 0.22 | 173172 tok/s | mfu 31.5% step 1720 | loss 3.4306 | lr 5.11e-04 | gnorm 0.21 | 173143 tok/s | mfu 31.4% step 1730 | loss 3.4649 | lr 5.10e-04 | gnorm 0.21 | 173320 tok/s | mfu 31.4% step 1740 | loss 3.4554 | lr 5.09e-04 | gnorm 0.22 | 173387 tok/s | mfu 31.4% step 1750 | loss 3.4467 | lr 5.08e-04 | gnorm 0.21 | 173319 tok/s | mfu 31.5% step 1760 | loss 3.4451 | lr 5.06e-04 | gnorm 0.22 | 173402 tok/s | mfu 31.5% step 1770 | loss 3.4620 | lr 5.05e-04 | gnorm 0.22 | 173340 tok/s | mfu 31.5% step 1780 | loss 3.4418 | lr 5.04e-04 | gnorm 0.22 | 173343 tok/s | mfu 31.5% step 1790 | loss 3.4301 | lr 5.03e-04 | gnorm 0.20 | 173318 tok/s | mfu 31.5% step 1800 | loss 3.4875 | lr 5.01e-04 | gnorm 0.23 | 173389 tok/s | mfu 31.5% step 1810 | loss 3.4508 | lr 5.00e-04 | gnorm 0.20 | 173235 tok/s | mfu 31.5% step 1820 | loss 3.4505 | lr 4.99e-04 | gnorm 0.21 | 173167 tok/s | mfu 31.5% step 1830 | loss 3.4722 | lr 4.98e-04 | gnorm 0.21 | 173326 tok/s | mfu 31.5% step 1840 | loss 3.4375 | lr 4.97e-04 | gnorm 0.25 | 173324 tok/s | mfu 31.5% step 1850 | loss 3.4270 | lr 4.95e-04 | gnorm 0.20 | 173350 tok/s | mfu 31.5% step 1860 | loss 3.4341 | lr 4.94e-04 | gnorm 0.21 | 172890 tok/s | mfu 31.5% step 1870 | loss 3.4196 | lr 4.93e-04 | gnorm 0.20 | 173142 tok/s | mfu 31.4% step 1880 | loss 3.4318 | lr 4.91e-04 | gnorm 0.21 | 173240 tok/s | mfu 31.5% step 1890 | loss 3.4439 | lr 4.90e-04 | gnorm 0.21 | 173284 tok/s | mfu 31.5% step 1900 | loss 3.4320 | lr 4.89e-04 | gnorm 0.23 | 173235 tok/s | mfu 31.4% step 1910 | loss 3.4377 | lr 4.88e-04 | gnorm 0.23 | 173102 tok/s | mfu 31.4% step 1920 | loss 3.4252 | lr 4.86e-04 | gnorm 0.19 | 173275 tok/s | mfu 31.5% step 1930 | loss 3.4208 | lr 4.85e-04 | gnorm 0.19 | 173474 tok/s | mfu 31.5% step 1940 | loss 3.4060 | lr 4.84e-04 | gnorm 0.22 | 173369 tok/s | mfu 31.5% step 1950 | loss 3.4449 | lr 4.82e-04 | gnorm 0.23 | 173517 tok/s | mfu 31.5% step 1960 | loss 3.4516 | lr 4.81e-04 | gnorm 0.23 | 173476 tok/s | mfu 31.5% step 1970 | loss 3.4312 | lr 4.80e-04 | gnorm 0.21 | 173411 tok/s | mfu 31.5% step 1980 | loss 3.4232 | lr 4.78e-04 | gnorm 0.23 | 173401 tok/s | mfu 31.5% step 1990 | loss 3.4016 | lr 4.77e-04 | gnorm 0.21 | 173309 tok/s | mfu 31.5% step 2000 | loss 3.4198 | lr 4.76e-04 | gnorm 0.20 | 173183 tok/s | mfu 31.5% step 2010 | loss 3.4313 | lr 4.74e-04 | gnorm 0.21 | 173166 tok/s | mfu 31.2% step 2020 | loss 3.4258 | lr 4.73e-04 | gnorm 0.22 | 173301 tok/s | mfu 31.2% step 2030 | loss 3.4127 | lr 4.72e-04 | gnorm 0.21 | 173362 tok/s | mfu 31.2% step 2040 | loss 3.4259 | lr 4.70e-04 | gnorm 0.22 | 173476 tok/s | mfu 31.2% step 2050 | loss 3.3991 | lr 4.69e-04 | gnorm 0.20 | 173440 tok/s | mfu 31.2% step 2060 | loss 3.3866 | lr 4.68e-04 | gnorm 0.20 | 173462 tok/s | mfu 31.5% step 2070 | loss 3.4130 | lr 4.66e-04 | gnorm 0.20 | 173601 tok/s | mfu 31.5% step 2080 | loss 3.4073 | lr 4.65e-04 | gnorm 0.22 | 173714 tok/s | mfu 31.5% step 2090 | loss 3.3792 | lr 4.64e-04 | gnorm 0.20 | 173579 tok/s | mfu 31.5% step 2100 | loss 3.3780 | lr 4.62e-04 | gnorm 0.20 | 173370 tok/s | mfu 31.5% step 2110 | loss 3.3982 | lr 4.61e-04 | gnorm 0.20 | 173413 tok/s | mfu 31.5% step 2120 | loss 3.3971 | lr 4.59e-04 | gnorm 0.21 | 173546 tok/s | mfu 31.5% step 2130 | loss 3.3960 | lr 4.58e-04 | gnorm 0.21 | 173444 tok/s | mfu 31.5% step 2140 | loss 3.3959 | lr 4.57e-04 | gnorm 0.20 | 173446 tok/s | mfu 31.5% step 2150 | loss 3.3887 | lr 4.55e-04 | gnorm 0.20 | 173506 tok/s | mfu 31.5% step 2160 | loss 3.3938 | lr 4.54e-04 | gnorm 0.21 | 173594 tok/s | mfu 31.5% step 2170 | loss 3.3842 | lr 4.52e-04 | gnorm 0.23 | 173392 tok/s | mfu 31.5% step 2180 | loss 3.3795 | lr 4.51e-04 | gnorm 0.21 | 173510 tok/s | mfu 31.5% step 2190 | loss 3.3773 | lr 4.50e-04 | gnorm 0.24 | 173523 tok/s | mfu 31.5% step 2200 | loss 3.3665 | lr 4.48e-04 | gnorm 0.22 | 173374 tok/s | mfu 31.5% step 2210 | loss 3.4096 | lr 4.47e-04 | gnorm 0.19 | 173380 tok/s | mfu 31.5% step 2220 | loss 3.3912 | lr 4.45e-04 | gnorm 0.19 | 173424 tok/s | mfu 31.5% step 2230 | loss 3.3869 | lr 4.44e-04 | gnorm 0.20 | 173398 tok/s | mfu 31.5% step 2240 | loss 3.3900 | lr 4.42e-04 | gnorm 0.24 | 173365 tok/s | mfu 31.5% step 2250 | loss 3.4083 | lr 4.41e-04 | gnorm 0.21 | 173397 tok/s | mfu 31.5% step 2260 | loss 3.3854 | lr 4.39e-04 | gnorm 0.21 | 173373 tok/s | mfu 31.5% step 2270 | loss 3.3973 | lr 4.38e-04 | gnorm 0.20 | 173171 tok/s | mfu 31.5% step 2280 | loss 3.3575 | lr 4.37e-04 | gnorm 0.20 | 173185 tok/s | mfu 31.5% step 2290 | loss 3.3709 | lr 4.35e-04 | gnorm 0.20 | 173285 tok/s | mfu 31.5% step 2300 | loss 3.3684 | lr 4.34e-04 | gnorm 0.19 | 173390 tok/s | mfu 31.5% step 2310 | loss 3.3814 | lr 4.32e-04 | gnorm 0.20 | 173405 tok/s | mfu 31.5% step 2320 | loss 3.3596 | lr 4.31e-04 | gnorm 0.21 | 173530 tok/s | mfu 31.5% step 2330 | loss 3.3476 | lr 4.29e-04 | gnorm 0.20 | 173446 tok/s | mfu 31.5% step 2340 | loss 3.3755 | lr 4.28e-04 | gnorm 0.22 | 173352 tok/s | mfu 31.5% step 2350 | loss 3.3807 | lr 4.26e-04 | gnorm 0.19 | 173381 tok/s | mfu 31.5% step 2360 | loss 3.3774 | lr 4.25e-04 | gnorm 0.21 | 173436 tok/s | mfu 31.5% step 2370 | loss 3.3622 | lr 4.23e-04 | gnorm 0.20 | 173388 tok/s | mfu 31.5% step 2380 | loss 3.3864 | lr 4.22e-04 | gnorm 0.21 | 173345 tok/s | mfu 31.5% step 2390 | loss 3.3722 | lr 4.20e-04 | gnorm 0.20 | 173508 tok/s | mfu 31.5% step 2400 | loss 3.3541 | lr 4.19e-04 | gnorm 0.20 | 173355 tok/s | mfu 31.5% step 2410 | loss 3.3399 | lr 4.17e-04 | gnorm 0.20 | 173423 tok/s | mfu 31.5% step 2420 | loss 3.3718 | lr 4.16e-04 | gnorm 0.19 | 173480 tok/s | mfu 31.5% step 2430 | loss 3.3641 | lr 4.14e-04 | gnorm 0.19 | 173561 tok/s | mfu 31.5% step 2440 | loss 3.3331 | lr 4.13e-04 | gnorm 0.21 | 173481 tok/s | mfu 31.5% step 2450 | loss 3.3824 | lr 4.11e-04 | gnorm 0.21 | 173426 tok/s | mfu 31.5% step 2460 | loss 3.3480 | lr 4.10e-04 | gnorm 0.20 | 173545 tok/s | mfu 31.5% step 2470 | loss 3.3602 | lr 4.08e-04 | gnorm 0.22 | 173493 tok/s | mfu 31.5% step 2480 | loss 3.3413 | lr 4.07e-04 | gnorm 0.19 | 173364 tok/s | mfu 31.5% step 2490 | loss 3.3450 | lr 4.05e-04 | gnorm 0.19 | 173420 tok/s | mfu 31.5% step 2500 | loss 3.3764 | lr 4.04e-04 | gnorm 0.20 | 173436 tok/s | mfu 31.5% step 2510 | loss 3.3458 | lr 4.02e-04 | gnorm 0.20 | 173536 tok/s | mfu 31.2% step 2520 | loss 3.3471 | lr 4.01e-04 | gnorm 0.20 | 173627 tok/s | mfu 31.3% step 2530 | loss 3.3508 | lr 3.99e-04 | gnorm 0.19 | 173413 tok/s | mfu 31.3% step 2540 | loss 3.3398 | lr 3.98e-04 | gnorm 0.20 | 173595 tok/s | mfu 31.3% step 2550 | loss 3.3155 | lr 3.96e-04 | gnorm 0.20 | 173479 tok/s | mfu 31.3% step 2560 | loss 3.3260 | lr 3.94e-04 | gnorm 0.21 | 173623 tok/s | mfu 31.5% step 2570 | loss 3.3472 | lr 3.93e-04 | gnorm 0.20 | 173509 tok/s | mfu 31.5% step 2580 | loss 3.3607 | lr 3.91e-04 | gnorm 0.19 | 173516 tok/s | mfu 31.5% step 2590 | loss 3.3477 | lr 3.90e-04 | gnorm 0.20 | 173797 tok/s | mfu 31.5% step 2600 | loss 3.3074 | lr 3.88e-04 | gnorm 0.21 | 173437 tok/s | mfu 31.5% step 2610 | loss 3.3338 | lr 3.87e-04 | gnorm 0.22 | 173413 tok/s | mfu 31.5% step 2620 | loss 3.3134 | lr 3.85e-04 | gnorm 0.20 | 173313 tok/s | mfu 31.5% step 2630 | loss 3.3328 | lr 3.84e-04 | gnorm 0.19 | 173340 tok/s | mfu 31.5% step 2640 | loss 3.3212 | lr 3.82e-04 | gnorm 0.21 | 173594 tok/s | mfu 31.5% step 2650 | loss 3.3577 | lr 3.80e-04 | gnorm 0.20 | 173538 tok/s | mfu 31.5% step 2660 | loss 3.3172 | lr 3.79e-04 | gnorm 0.19 | 173231 tok/s | mfu 31.5% step 2670 | loss 3.3489 | lr 3.77e-04 | gnorm 0.21 | 173389 tok/s | mfu 31.5% step 2680 | loss 3.3207 | lr 3.76e-04 | gnorm 0.24 | 173463 tok/s | mfu 31.5% step 2690 | loss 3.3348 | lr 3.74e-04 | gnorm 0.20 | 173468 tok/s | mfu 31.5% step 2700 | loss 3.3496 | lr 3.73e-04 | gnorm 0.20 | 172999 tok/s | mfu 31.5% step 2710 | loss 3.3516 | lr 3.71e-04 | gnorm 0.21 | 173302 tok/s | mfu 31.5% step 2720 | loss 3.3138 | lr 3.69e-04 | gnorm 0.21 | 173450 tok/s | mfu 31.5% step 2730 | loss 3.3345 | lr 3.68e-04 | gnorm 0.21 | 173250 tok/s | mfu 31.5% step 2740 | loss 3.3209 | lr 3.66e-04 | gnorm 0.19 | 173364 tok/s | mfu 31.5% step 2750 | loss 3.3199 | lr 3.65e-04 | gnorm 0.20 | 173284 tok/s | mfu 31.5% step 2760 | loss 3.3352 | lr 3.63e-04 | gnorm 0.19 | 173424 tok/s | mfu 31.5% step 2770 | loss 3.3204 | lr 3.62e-04 | gnorm 0.21 | 173546 tok/s | mfu 31.5% step 2780 | loss 3.3004 | lr 3.60e-04 | gnorm 0.19 | 173396 tok/s | mfu 31.5% step 2790 | loss 3.3357 | lr 3.58e-04 | gnorm 0.19 | 173317 tok/s | mfu 31.5% step 2800 | loss 3.3209 | lr 3.57e-04 | gnorm 0.20 | 173571 tok/s | mfu 31.5% step 2810 | loss 3.3008 | lr 3.55e-04 | gnorm 0.19 | 173493 tok/s | mfu 31.5% step 2820 | loss 3.3230 | lr 3.54e-04 | gnorm 0.20 | 173373 tok/s | mfu 31.5% step 2830 | loss 3.3059 | lr 3.52e-04 | gnorm 0.19 | 173477 tok/s | mfu 31.5% step 2840 | loss 3.3403 | lr 3.51e-04 | gnorm 0.19 | 173275 tok/s | mfu 31.5% step 2850 | loss 3.3162 | lr 3.49e-04 | gnorm 0.20 | 173246 tok/s | mfu 31.5% step 2860 | loss 3.3435 | lr 3.47e-04 | gnorm 0.20 | 173228 tok/s | mfu 31.5% step 2870 | loss 3.3176 | lr 3.46e-04 | gnorm 0.21 | 173156 tok/s | mfu 31.5% step 2880 | loss 3.2848 | lr 3.44e-04 | gnorm 0.21 | 173264 tok/s | mfu 31.4% step 2890 | loss 3.3212 | lr 3.43e-04 | gnorm 0.19 | 173365 tok/s | mfu 31.4% step 2900 | loss 3.3116 | lr 3.41e-04 | gnorm 0.21 | 173569 tok/s | mfu 31.5% step 2910 | loss 3.2962 | lr 3.39e-04 | gnorm 0.20 | 173457 tok/s | mfu 31.5% step 2920 | loss 3.2953 | lr 3.38e-04 | gnorm 0.22 | 173498 tok/s | mfu 31.5% step 2930 | loss 3.3306 | lr 3.36e-04 | gnorm 0.21 | 173598 tok/s | mfu 31.5% step 2940 | loss 3.2850 | lr 3.35e-04 | gnorm 0.19 | 173326 tok/s | mfu 31.5% step 2950 | loss 3.3189 | lr 3.33e-04 | gnorm 0.20 | 173293 tok/s | mfu 31.5% step 2960 | loss 3.3164 | lr 3.32e-04 | gnorm 0.19 | 173493 tok/s | mfu 31.5% step 2970 | loss 3.3196 | lr 3.30e-04 | gnorm 0.20 | 173435 tok/s | mfu 31.5% step 2980 | loss 3.2855 | lr 3.28e-04 | gnorm 0.20 | 173507 tok/s | mfu 31.5% step 2990 | loss 3.2866 | lr 3.27e-04 | gnorm 0.19 | 173413 tok/s | mfu 31.5% step 3000 | loss 3.2853 | lr 3.25e-04 | gnorm 0.19 | 173497 tok/s | mfu 31.5% step 3010 | loss 3.3124 | lr 3.24e-04 | gnorm 0.20 | 173319 tok/s | mfu 31.3% step 3020 | loss 3.2861 | lr 3.22e-04 | gnorm 0.19 | 173294 tok/s | mfu 31.3% step 3030 | loss 3.2973 | lr 3.20e-04 | gnorm 0.19 | 173336 tok/s | mfu 31.2% step 3040 | loss 3.2995 | lr 3.19e-04 | gnorm 0.20 | 173426 tok/s | mfu 31.2% step 3050 | loss 3.3066 | lr 3.17e-04 | gnorm 0.21 | 173460 tok/s | mfu 31.2% step 3060 | loss 3.3150 | lr 3.16e-04 | gnorm 0.20 | 173473 tok/s | mfu 31.5% step 3070 | loss 3.3102 | lr 3.14e-04 | gnorm 0.19 | 173609 tok/s | mfu 31.5% step 3080 | loss 3.2874 | lr 3.12e-04 | gnorm 0.21 | 173613 tok/s | mfu 31.5% step 3090 | loss 3.2888 | lr 3.11e-04 | gnorm 0.20 | 173466 tok/s | mfu 31.5% step 3100 | loss 3.2726 | lr 3.09e-04 | gnorm 0.21 | 173445 tok/s | mfu 31.5% step 3110 | loss 3.2998 | lr 3.08e-04 | gnorm 0.21 | 173455 tok/s | mfu 31.5% step 3120 | loss 3.2531 | lr 3.06e-04 | gnorm 0.19 | 173571 tok/s | mfu 31.5% step 3130 | loss 3.2508 | lr 3.05e-04 | gnorm 0.20 | 173624 tok/s | mfu 31.5% step 3140 | loss 3.2965 | lr 3.03e-04 | gnorm 0.20 | 173593 tok/s | mfu 31.5% step 3150 | loss 3.3263 | lr 3.01e-04 | gnorm 0.19 | 173705 tok/s | mfu 31.5% step 3160 | loss 3.2834 | lr 3.00e-04 | gnorm 0.20 | 173602 tok/s | mfu 31.5% step 3170 | loss 3.2951 | lr 2.98e-04 | gnorm 0.19 | 173464 tok/s | mfu 31.5% step 3180 | loss 3.2835 | lr 2.97e-04 | gnorm 0.19 | 173472 tok/s | mfu 31.5% step 3190 | loss 3.3044 | lr 2.95e-04 | gnorm 0.19 | 173483 tok/s | mfu 31.5% step 3200 | loss 3.2897 | lr 2.93e-04 | gnorm 0.20 | 173554 tok/s | mfu 31.5% step 3210 | loss 3.2823 | lr 2.92e-04 | gnorm 0.19 | 173670 tok/s | mfu 31.5% step 3220 | loss 3.2681 | lr 2.90e-04 | gnorm 0.20 | 173670 tok/s | mfu 31.5% step 3230 | loss 3.2782 | lr 2.89e-04 | gnorm 0.18 | 173605 tok/s | mfu 31.5% step 3240 | loss 3.2643 | lr 2.87e-04 | gnorm 0.19 | 173592 tok/s | mfu 31.5% step 3250 | loss 3.2378 | lr 2.86e-04 | gnorm 0.20 | 173589 tok/s | mfu 31.5% step 3260 | loss 3.2620 | lr 2.84e-04 | gnorm 0.19 | 173692 tok/s | mfu 31.5% step 3270 | loss 3.2796 | lr 2.83e-04 | gnorm 0.20 | 173574 tok/s | mfu 31.5% step 3280 | loss 3.2617 | lr 2.81e-04 | gnorm 0.20 | 173669 tok/s | mfu 31.5% step 3290 | loss 3.2691 | lr 2.79e-04 | gnorm 0.19 | 173744 tok/s | mfu 31.5% step 3300 | loss 3.2681 | lr 2.78e-04 | gnorm 0.19 | 173532 tok/s | mfu 31.5% step 3310 | loss 3.2654 | lr 2.76e-04 | gnorm 0.20 | 173581 tok/s | mfu 31.5% step 3320 | loss 3.2482 | lr 2.75e-04 | gnorm 0.19 | 173428 tok/s | mfu 31.5% step 3330 | loss 3.2949 | lr 2.73e-04 | gnorm 0.20 | 173569 tok/s | mfu 31.5% step 3340 | loss 3.2568 | lr 2.72e-04 | gnorm 0.19 | 173613 tok/s | mfu 31.5% step 3350 | loss 3.2659 | lr 2.70e-04 | gnorm 0.18 | 173745 tok/s | mfu 31.5% step 3360 | loss 3.2335 | lr 2.69e-04 | gnorm 0.19 | 173615 tok/s | mfu 31.5% step 3370 | loss 3.2691 | lr 2.67e-04 | gnorm 0.20 | 173593 tok/s | mfu 31.5% step 3380 | loss 3.2490 | lr 2.65e-04 | gnorm 0.21 | 173423 tok/s | mfu 31.5% step 3390 | loss 3.2742 | lr 2.64e-04 | gnorm 0.19 | 173526 tok/s | mfu 31.5% step 3400 | loss 3.2681 | lr 2.62e-04 | gnorm 0.19 | 173494 tok/s | mfu 31.5% step 3410 | loss 3.2566 | lr 2.61e-04 | gnorm 0.19 | 173509 tok/s | mfu 31.5% step 3420 | loss 3.2518 | lr 2.59e-04 | gnorm 0.19 | 173571 tok/s | mfu 31.5% step 3430 | loss 3.2532 | lr 2.58e-04 | gnorm 0.18 | 173478 tok/s | mfu 31.5% step 3440 | loss 3.2514 | lr 2.56e-04 | gnorm 0.20 | 173606 tok/s | mfu 31.5% step 3450 | loss 3.2499 | lr 2.55e-04 | gnorm 0.19 | 173462 tok/s | mfu 31.5% step 3460 | loss 3.2622 | lr 2.53e-04 | gnorm 0.19 | 173408 tok/s | mfu 31.5% step 3470 | loss 3.2867 | lr 2.52e-04 | gnorm 0.21 | 173573 tok/s | mfu 31.5% step 3480 | loss 3.2434 | lr 2.50e-04 | gnorm 0.19 | 173671 tok/s | mfu 31.5% step 3490 | loss 3.2586 | lr 2.49e-04 | gnorm 0.19 | 173595 tok/s | mfu 31.5% step 3500 | loss 3.2308 | lr 2.47e-04 | gnorm 0.19 | 173623 tok/s | mfu 31.5% step 3510 | loss 3.2597 | lr 2.46e-04 | gnorm 0.20 | 173566 tok/s | mfu 31.3% step 3520 | loss 3.2636 | lr 2.44e-04 | gnorm 0.20 | 173725 tok/s | mfu 31.3% step 3530 | loss 3.2957 | lr 2.43e-04 | gnorm 0.20 | 173656 tok/s | mfu 31.3% step 3540 | loss 3.2580 | lr 2.41e-04 | gnorm 0.20 | 173603 tok/s | mfu 31.3% step 3550 | loss 3.2491 | lr 2.40e-04 | gnorm 0.20 | 173669 tok/s | mfu 31.3% step 3560 | loss 3.2439 | lr 2.38e-04 | gnorm 0.19 | 173558 tok/s | mfu 31.5% step 3570 | loss 3.2229 | lr 2.37e-04 | gnorm 0.19 | 173629 tok/s | mfu 31.5% step 3580 | loss 3.2422 | lr 2.35e-04 | gnorm 0.20 | 173499 tok/s | mfu 31.5% step 3590 | loss 3.2357 | lr 2.34e-04 | gnorm 0.19 | 173425 tok/s | mfu 31.5% step 3600 | loss 3.2406 | lr 2.32e-04 | gnorm 0.19 | 173525 tok/s | mfu 31.5% step 3610 | loss 3.2161 | lr 2.31e-04 | gnorm 0.20 | 173590 tok/s | mfu 31.5% step 3620 | loss 3.2372 | lr 2.29e-04 | gnorm 0.19 | 173540 tok/s | mfu 31.5% step 3630 | loss 3.2645 | lr 2.28e-04 | gnorm 0.21 | 173395 tok/s | mfu 31.5% step 3640 | loss 3.2570 | lr 2.26e-04 | gnorm 0.19 | 173615 tok/s | mfu 31.5% step 3650 | loss 3.2566 | lr 2.25e-04 | gnorm 0.19 | 173665 tok/s | mfu 31.5% step 3660 | loss 3.2463 | lr 2.23e-04 | gnorm 0.21 | 173599 tok/s | mfu 31.5% step 3670 | loss 3.2605 | lr 2.22e-04 | gnorm 0.20 | 173437 tok/s | mfu 31.5% step 3680 | loss 3.2617 | lr 2.20e-04 | gnorm 0.20 | 173707 tok/s | mfu 31.5% step 3690 | loss 3.2151 | lr 2.19e-04 | gnorm 0.19 | 173704 tok/s | mfu 31.5% step 3700 | loss 3.2263 | lr 2.17e-04 | gnorm 0.20 | 173713 tok/s | mfu 31.5% step 3710 | loss 3.2601 | lr 2.16e-04 | gnorm 0.20 | 173658 tok/s | mfu 31.5% step 3720 | loss 3.2400 | lr 2.15e-04 | gnorm 0.20 | 173700 tok/s | mfu 31.5% step 3730 | loss 3.2230 | lr 2.13e-04 | gnorm 0.18 | 173694 tok/s | mfu 31.5% step 3740 | loss 3.2683 | lr 2.12e-04 | gnorm 0.19 | 173701 tok/s | mfu 31.5% step 3750 | loss 3.2187 | lr 2.10e-04 | gnorm 0.19 | 173635 tok/s | mfu 31.5% step 3760 | loss 3.2416 | lr 2.09e-04 | gnorm 0.19 | 173361 tok/s | mfu 31.5% step 3770 | loss 3.2136 | lr 2.07e-04 | gnorm 0.19 | 173504 tok/s | mfu 31.5% step 3780 | loss 3.2460 | lr 2.06e-04 | gnorm 0.19 | 173634 tok/s | mfu 31.5% step 3790 | loss 3.2415 | lr 2.05e-04 | gnorm 0.19 | 173577 tok/s | mfu 31.5% step 3800 | loss 3.2406 | lr 2.03e-04 | gnorm 0.19 | 173590 tok/s | mfu 31.5% step 3810 | loss 3.2348 | lr 2.02e-04 | gnorm 0.19 | 173410 tok/s | mfu 31.5% step 3820 | loss 3.2178 | lr 2.00e-04 | gnorm 0.18 | 173521 tok/s | mfu 31.5% step 3830 | loss 3.2467 | lr 1.99e-04 | gnorm 0.18 | 173441 tok/s | mfu 31.5% step 3840 | loss 3.2442 | lr 1.98e-04 | gnorm 0.20 | 173473 tok/s | mfu 31.5% step 3850 | loss 3.2140 | lr 1.96e-04 | gnorm 0.20 | 173470 tok/s | mfu 31.5% step 3860 | loss 3.2523 | lr 1.95e-04 | gnorm 0.20 | 173603 tok/s | mfu 31.5% step 3870 | loss 3.2386 | lr 1.94e-04 | gnorm 0.19 | 173682 tok/s | mfu 31.5% step 3880 | loss 3.2189 | lr 1.92e-04 | gnorm 0.19 | 173566 tok/s | mfu 31.5% step 3890 | loss 3.2397 | lr 1.91e-04 | gnorm 0.19 | 173476 tok/s | mfu 31.5% step 3900 | loss 3.2628 | lr 1.89e-04 | gnorm 0.19 | 173536 tok/s | mfu 31.5% step 3910 | loss 3.1880 | lr 1.88e-04 | gnorm 0.20 | 173535 tok/s | mfu 31.5% step 3920 | loss 3.2220 | lr 1.87e-04 | gnorm 0.19 | 173331 tok/s | mfu 31.5% step 3930 | loss 3.2312 | lr 1.85e-04 | gnorm 0.19 | 173361 tok/s | mfu 31.5% step 3940 | loss 3.2482 | lr 1.84e-04 | gnorm 0.19 | 173393 tok/s | mfu 31.5% step 3950 | loss 3.2000 | lr 1.83e-04 | gnorm 0.19 | 173462 tok/s | mfu 31.5% step 3960 | loss 3.1801 | lr 1.81e-04 | gnorm 0.21 | 173440 tok/s | mfu 31.5% step 3970 | loss 3.2333 | lr 1.80e-04 | gnorm 0.20 | 173561 tok/s | mfu 31.5% step 3980 | loss 3.2353 | lr 1.79e-04 | gnorm 0.20 | 173491 tok/s | mfu 31.5% step 3990 | loss 3.2336 | lr 1.77e-04 | gnorm 0.19 | 173168 tok/s | mfu 31.5% step 4000 | loss 3.2444 | lr 1.76e-04 | gnorm 0.19 | 173432 tok/s | mfu 31.5% step 4010 | loss 3.1846 | lr 1.75e-04 | gnorm 0.19 | 173511 tok/s | mfu 31.3% step 4020 | loss 3.2165 | lr 1.74e-04 | gnorm 0.20 | 173474 tok/s | mfu 31.3% step 4030 | loss 3.2179 | lr 1.72e-04 | gnorm 0.20 | 173617 tok/s | mfu 31.3% step 4040 | loss 3.2178 | lr 1.71e-04 | gnorm 0.19 | 173435 tok/s | mfu 31.3% step 4050 | loss 3.2321 | lr 1.70e-04 | gnorm 0.19 | 173316 tok/s | mfu 31.3% step 4060 | loss 3.2504 | lr 1.68e-04 | gnorm 0.19 | 173214 tok/s | mfu 31.5% step 4070 | loss 3.2070 | lr 1.67e-04 | gnorm 0.20 | 173457 tok/s | mfu 31.5% step 4080 | loss 3.2338 | lr 1.66e-04 | gnorm 0.18 | 173440 tok/s | mfu 31.5% step 4090 | loss 3.2106 | lr 1.65e-04 | gnorm 0.19 | 173343 tok/s | mfu 31.5% step 4100 | loss 3.2117 | lr 1.63e-04 | gnorm 0.18 | 173374 tok/s | mfu 31.5% step 4110 | loss 3.1994 | lr 1.62e-04 | gnorm 0.18 | 173333 tok/s | mfu 31.5% step 4120 | loss 3.2024 | lr 1.61e-04 | gnorm 0.19 | 173421 tok/s | mfu 31.5% step 4130 | loss 3.2342 | lr 1.60e-04 | gnorm 0.20 | 173602 tok/s | mfu 31.5% step 4140 | loss 3.2027 | lr 1.58e-04 | gnorm 0.20 | 173420 tok/s | mfu 31.5% step 4150 | loss 3.2129 | lr 1.57e-04 | gnorm 0.20 | 173507 tok/s | mfu 31.5% step 4160 | loss 3.2133 | lr 1.56e-04 | gnorm 0.19 | 173426 tok/s | mfu 31.5% step 4170 | loss 3.2259 | lr 1.55e-04 | gnorm 0.20 | 173559 tok/s | mfu 31.5% step 4180 | loss 3.2158 | lr 1.54e-04 | gnorm 0.19 | 173710 tok/s | mfu 31.5% step 4190 | loss 3.2310 | lr 1.52e-04 | gnorm 0.18 | 173676 tok/s | mfu 31.5% step 4200 | loss 3.2063 | lr 1.51e-04 | gnorm 0.19 | 173582 tok/s | mfu 31.5% step 4210 | loss 3.2190 | lr 1.50e-04 | gnorm 0.19 | 173659 tok/s | mfu 31.5% step 4220 | loss 3.2273 | lr 1.49e-04 | gnorm 0.20 | 173598 tok/s | mfu 31.5% step 4230 | loss 3.2383 | lr 1.48e-04 | gnorm 0.20 | 173637 tok/s | mfu 31.5% step 4240 | loss 3.2230 | lr 1.46e-04 | gnorm 0.19 | 173661 tok/s | mfu 31.5% step 4250 | loss 3.2289 | lr 1.45e-04 | gnorm 0.19 | 173694 tok/s | mfu 31.5% step 4260 | loss 3.2011 | lr 1.44e-04 | gnorm 0.19 | 173651 tok/s | mfu 31.5% step 4270 | loss 3.1918 | lr 1.43e-04 | gnorm 0.20 | 173477 tok/s | mfu 31.5% step 4280 | loss 3.2195 | lr 1.42e-04 | gnorm 0.19 | 173436 tok/s | mfu 31.5% step 4290 | loss 3.1896 | lr 1.41e-04 | gnorm 0.18 | 173402 tok/s | mfu 31.5% step 4300 | loss 3.2234 | lr 1.40e-04 | gnorm 0.19 | 173426 tok/s | mfu 31.5% step 4310 | loss 3.1801 | lr 1.38e-04 | gnorm 0.19 | 173517 tok/s | mfu 31.5% step 4320 | loss 3.2140 | lr 1.37e-04 | gnorm 0.19 | 173672 tok/s | mfu 31.5% step 4330 | loss 3.1806 | lr 1.36e-04 | gnorm 0.18 | 173518 tok/s | mfu 31.5% step 4340 | loss 3.2080 | lr 1.35e-04 | gnorm 0.19 | 173421 tok/s | mfu 31.5% step 4350 | loss 3.2802 | lr 1.34e-04 | gnorm 0.20 | 173691 tok/s | mfu 31.5% step 4360 | loss 3.2153 | lr 1.33e-04 | gnorm 0.20 | 173627 tok/s | mfu 31.5% step 4370 | loss 3.1813 | lr 1.32e-04 | gnorm 0.21 | 173389 tok/s | mfu 31.5% step 4380 | loss 3.2056 | lr 1.31e-04 | gnorm 0.19 | 173469 tok/s | mfu 31.5% step 4390 | loss 3.2353 | lr 1.30e-04 | gnorm 0.19 | 173412 tok/s | mfu 31.5% step 4400 | loss 3.2049 | lr 1.29e-04 | gnorm 0.19 | 173396 tok/s | mfu 31.5% step 4410 | loss 3.2020 | lr 1.28e-04 | gnorm 0.19 | 173398 tok/s | mfu 31.5% step 4420 | loss 3.2010 | lr 1.27e-04 | gnorm 0.19 | 173359 tok/s | mfu 31.5% step 4430 | loss 3.2310 | lr 1.25e-04 | gnorm 0.20 | 173459 tok/s | mfu 31.5% step 4440 | loss 3.1881 | lr 1.24e-04 | gnorm 0.19 | 173469 tok/s | mfu 31.5% step 4450 | loss 3.1733 | lr 1.23e-04 | gnorm 0.19 | 173396 tok/s | mfu 31.5% step 4460 | loss 3.2051 | lr 1.22e-04 | gnorm 0.19 | 173328 tok/s | mfu 31.5% step 4470 | loss 3.1920 | lr 1.21e-04 | gnorm 0.19 | 173299 tok/s | mfu 31.5% step 4480 | loss 3.1857 | lr 1.20e-04 | gnorm 0.19 | 173468 tok/s | mfu 31.5% step 4490 | loss 3.1725 | lr 1.19e-04 | gnorm 0.19 | 173374 tok/s | mfu 31.5% step 4500 | loss 3.2318 | lr 1.18e-04 | gnorm 0.19 | 173441 tok/s | mfu 31.5% step 4510 | loss 3.1859 | lr 1.17e-04 | gnorm 0.20 | 173538 tok/s | mfu 31.3% step 4520 | loss 3.2216 | lr 1.16e-04 | gnorm 0.21 | 173559 tok/s | mfu 31.3% step 4530 | loss 3.2296 | lr 1.15e-04 | gnorm 0.19 | 173556 tok/s | mfu 31.3% step 4540 | loss 3.1807 | lr 1.15e-04 | gnorm 0.19 | 173546 tok/s | mfu 31.3% step 4550 | loss 3.1837 | lr 1.14e-04 | gnorm 0.19 | 173329 tok/s | mfu 31.3% step 4560 | loss 3.1836 | lr 1.13e-04 | gnorm 0.19 | 173334 tok/s | mfu 31.5% step 4570 | loss 3.2265 | lr 1.12e-04 | gnorm 0.20 | 173389 tok/s | mfu 31.5% step 4580 | loss 3.1843 | lr 1.11e-04 | gnorm 0.19 | 173591 tok/s | mfu 31.5% step 4590 | loss 3.2051 | lr 1.10e-04 | gnorm 0.19 | 173557 tok/s | mfu 31.5% step 4600 | loss 3.1911 | lr 1.09e-04 | gnorm 0.19 | 173540 tok/s | mfu 31.5% step 4610 | loss 3.1943 | lr 1.08e-04 | gnorm 0.20 | 173471 tok/s | mfu 31.5% step 4620 | loss 3.1796 | lr 1.07e-04 | gnorm 0.19 | 173474 tok/s | mfu 31.5% step 4630 | loss 3.1899 | lr 1.06e-04 | gnorm 0.18 | 173349 tok/s | mfu 31.5% step 4640 | loss 3.1691 | lr 1.05e-04 | gnorm 0.20 | 173390 tok/s | mfu 31.5% step 4650 | loss 3.2310 | lr 1.04e-04 | gnorm 0.19 | 173479 tok/s | mfu 31.5% step 4660 | loss 3.1614 | lr 1.04e-04 | gnorm 0.21 | 173452 tok/s | mfu 31.5% step 4670 | loss 3.1935 | lr 1.03e-04 | gnorm 0.19 | 173549 tok/s | mfu 31.5% step 4680 | loss 3.1737 | lr 1.02e-04 | gnorm 0.19 | 173521 tok/s | mfu 31.5% step 4690 | loss 3.1584 | lr 1.01e-04 | gnorm 0.19 | 173437 tok/s | mfu 31.5% step 4700 | loss 3.2051 | lr 1.00e-04 | gnorm 0.19 | 173547 tok/s | mfu 31.5% step 4710 | loss 3.1819 | lr 9.93e-05 | gnorm 0.19 | 173481 tok/s | mfu 31.5% step 4720 | loss 3.1644 | lr 9.85e-05 | gnorm 0.19 | 173482 tok/s | mfu 31.5% step 4730 | loss 3.1794 | lr 9.77e-05 | gnorm 0.21 | 173342 tok/s | mfu 31.5% step 4740 | loss 3.1685 | lr 9.69e-05 | gnorm 0.19 | 173661 tok/s | mfu 31.5% step 4750 | loss 3.2076 | lr 9.61e-05 | gnorm 0.19 | 173520 tok/s | mfu 31.5% step 4760 | loss 3.1797 | lr 9.53e-05 | gnorm 0.19 | 173507 tok/s | mfu 31.5% step 4770 | loss 3.1824 | lr 9.45e-05 | gnorm 0.19 | 173612 tok/s | mfu 31.5% step 4780 | loss 3.1888 | lr 9.38e-05 | gnorm 0.19 | 173503 tok/s | mfu 31.5% step 4790 | loss 3.2233 | lr 9.30e-05 | gnorm 0.20 | 173478 tok/s | mfu 31.5% step 4800 | loss 3.2044 | lr 9.22e-05 | gnorm 0.19 | 173636 tok/s | mfu 31.5% step 4810 | loss 3.1642 | lr 9.15e-05 | gnorm 0.19 | 173631 tok/s | mfu 31.5% step 4820 | loss 3.1872 | lr 9.08e-05 | gnorm 0.21 | 173636 tok/s | mfu 31.5% step 4830 | loss 3.2115 | lr 9.00e-05 | gnorm 0.19 | 173631 tok/s | mfu 31.5% step 4840 | loss 3.1797 | lr 8.93e-05 | gnorm 0.19 | 173458 tok/s | mfu 31.5% step 4850 | loss 3.1713 | lr 8.86e-05 | gnorm 0.20 | 173408 tok/s | mfu 31.5% step 4860 | loss 3.1921 | lr 8.79e-05 | gnorm 0.19 | 173347 tok/s | mfu 31.5% step 4870 | loss 3.1722 | lr 8.72e-05 | gnorm 0.18 | 173453 tok/s | mfu 31.5% step 4880 | loss 3.1669 | lr 8.65e-05 | gnorm 0.19 | 173418 tok/s | mfu 31.5% step 4890 | loss 3.1684 | lr 8.58e-05 | gnorm 0.19 | 173505 tok/s | mfu 31.5% step 4900 | loss 3.1986 | lr 8.51e-05 | gnorm 0.20 | 173553 tok/s | mfu 31.5% step 4910 | loss 3.1714 | lr 8.45e-05 | gnorm 0.20 | 173571 tok/s | mfu 31.5% step 4920 | loss 3.1685 | lr 8.38e-05 | gnorm 0.20 | 173592 tok/s | mfu 31.5% step 4930 | loss 3.2111 | lr 8.32e-05 | gnorm 0.19 | 173560 tok/s | mfu 31.5% step 4940 | loss 3.1813 | lr 8.25e-05 | gnorm 0.19 | 173659 tok/s | mfu 31.5% step 4950 | loss 3.1777 | lr 8.19e-05 | gnorm 0.19 | 173567 tok/s | mfu 31.5% step 4960 | loss 3.1819 | lr 8.13e-05 | gnorm 0.19 | 173744 tok/s | mfu 31.5% step 4970 | loss 3.1950 | lr 8.07e-05 | gnorm 0.19 | 173581 tok/s | mfu 31.5% step 4980 | loss 3.1918 | lr 8.00e-05 | gnorm 0.18 | 173602 tok/s | mfu 31.5% step 4990 | loss 3.1868 | lr 7.94e-05 | gnorm 0.19 | 173576 tok/s | mfu 31.5% step 5000 | loss 3.1787 | lr 7.89e-05 | gnorm 0.19 | 173612 tok/s | mfu 31.5% step 5010 | loss 3.1721 | lr 7.83e-05 | gnorm 0.19 | 173468 tok/s | mfu 31.3% step 5020 | loss 3.1755 | lr 7.77e-05 | gnorm 0.19 | 173428 tok/s | mfu 31.3% step 5030 | loss 3.1838 | lr 7.72e-05 | gnorm 0.19 | 173423 tok/s | mfu 31.3% step 5040 | loss 3.2032 | lr 7.66e-05 | gnorm 0.19 | 173581 tok/s | mfu 31.3% step 5050 | loss 3.1893 | lr 7.61e-05 | gnorm 0.19 | 173243 tok/s | mfu 31.3% step 5060 | loss 3.1631 | lr 7.55e-05 | gnorm 0.19 | 173505 tok/s | mfu 31.5% step 5070 | loss 3.1716 | lr 7.50e-05 | gnorm 0.19 | 173384 tok/s | mfu 31.5% step 5080 | loss 3.1927 | lr 7.45e-05 | gnorm 0.18 | 173312 tok/s | mfu 31.5% step 5090 | loss 3.1970 | lr 7.40e-05 | gnorm 0.19 | 173321 tok/s | mfu 31.5% step 5100 | loss 3.1728 | lr 7.35e-05 | gnorm 0.19 | 173271 tok/s | mfu 31.5% step 5110 | loss 3.1724 | lr 7.30e-05 | gnorm 0.19 | 173343 tok/s | mfu 31.5% step 5120 | loss 3.1558 | lr 7.25e-05 | gnorm 0.20 | 173394 tok/s | mfu 31.5% step 5130 | loss 3.1810 | lr 7.20e-05 | gnorm 0.18 | 173582 tok/s | mfu 31.5% step 5140 | loss 3.1945 | lr 7.16e-05 | gnorm 0.19 | 173507 tok/s | mfu 31.5% step 5150 | loss 3.1813 | lr 7.11e-05 | gnorm 0.19 | 173520 tok/s | mfu 31.5% step 5160 | loss 3.1908 | lr 7.07e-05 | gnorm 0.19 | 173466 tok/s | mfu 31.5% step 5170 | loss 3.1597 | lr 7.02e-05 | gnorm 0.19 | 173542 tok/s | mfu 31.5% step 5180 | loss 3.1968 | lr 6.98e-05 | gnorm 0.19 | 173378 tok/s | mfu 31.5% step 5190 | loss 3.1578 | lr 6.94e-05 | gnorm 0.19 | 173481 tok/s | mfu 31.5% step 5200 | loss 3.1742 | lr 6.90e-05 | gnorm 0.19 | 173457 tok/s | mfu 31.5% step 5210 | loss 3.1798 | lr 6.86e-05 | gnorm 0.19 | 173385 tok/s | mfu 31.5% step 5220 | loss 3.1707 | lr 6.82e-05 | gnorm 0.19 | 173487 tok/s | mfu 31.5% step 5230 | loss 3.1500 | lr 6.78e-05 | gnorm 0.21 | 173595 tok/s | mfu 31.5% step 5240 | loss 3.1494 | lr 6.74e-05 | gnorm 0.19 | 173476 tok/s | mfu 31.5% step 5250 | loss 3.1727 | lr 6.70e-05 | gnorm 0.20 | 173531 tok/s | mfu 31.5% step 5260 | loss 3.1645 | lr 6.67e-05 | gnorm 0.19 | 173575 tok/s | mfu 31.5% step 5270 | loss 3.1716 | lr 6.63e-05 | gnorm 0.19 | 173596 tok/s | mfu 31.5% step 5280 | loss 3.1835 | lr 6.60e-05 | gnorm 0.19 | 173427 tok/s | mfu 31.5% step 5290 | loss 3.2172 | lr 6.57e-05 | gnorm 0.19 | 173495 tok/s | mfu 31.5% step 5300 | loss 3.1692 | lr 6.54e-05 | gnorm 0.19 | 173660 tok/s | mfu 31.5% step 5310 | loss 3.1846 | lr 6.50e-05 | gnorm 0.19 | 173707 tok/s | mfu 31.5% step 5320 | loss 3.1490 | lr 6.47e-05 | gnorm 0.20 | 173661 tok/s | mfu 31.5% step 5330 | loss 3.1596 | lr 6.45e-05 | gnorm 0.19 | 173748 tok/s | mfu 31.5% step 5340 | loss 3.1813 | lr 6.42e-05 | gnorm 0.19 | 173567 tok/s | mfu 31.5% step 5350 | loss 3.1909 | lr 6.39e-05 | gnorm 0.19 | 173662 tok/s | mfu 31.5% step 5360 | loss 3.1712 | lr 6.36e-05 | gnorm 0.19 | 173386 tok/s | mfu 31.5% step 5370 | loss 3.1687 | lr 6.34e-05 | gnorm 0.19 | 173647 tok/s | mfu 31.5% step 5380 | loss 3.1587 | lr 6.31e-05 | gnorm 0.19 | 173280 tok/s | mfu 31.5% step 5390 | loss 3.1798 | lr 6.29e-05 | gnorm 0.19 | 173574 tok/s | mfu 31.5% step 5400 | loss 3.1720 | lr 6.27e-05 | gnorm 0.19 | 173503 tok/s | mfu 31.5% step 5410 | loss 3.1670 | lr 6.24e-05 | gnorm 0.19 | 173501 tok/s | mfu 31.5% step 5420 | loss 3.1505 | lr 6.22e-05 | gnorm 0.19 | 173454 tok/s | mfu 31.5% step 5430 | loss 3.1922 | lr 6.20e-05 | gnorm 0.20 | 173458 tok/s | mfu 31.5% step 5440 | loss 3.1512 | lr 6.18e-05 | gnorm 0.20 | 173288 tok/s | mfu 31.5% step 5450 | loss 3.1386 | lr 6.17e-05 | gnorm 0.19 | 173453 tok/s | mfu 31.5% step 5460 | loss 3.1678 | lr 6.15e-05 | gnorm 0.19 | 173474 tok/s | mfu 31.5% step 5470 | loss 3.1616 | lr 6.13e-05 | gnorm 0.20 | 173478 tok/s | mfu 31.5% step 5480 | loss 3.1820 | lr 6.12e-05 | gnorm 0.20 | 173544 tok/s | mfu 31.5% step 5490 | loss 3.1498 | lr 6.10e-05 | gnorm 0.19 | 173507 tok/s | mfu 31.5% step 5500 | loss 3.1690 | lr 6.09e-05 | gnorm 0.19 | 173352 tok/s | mfu 31.5% step 5510 | loss 3.1546 | lr 6.08e-05 | gnorm 0.20 | 173649 tok/s | mfu 31.3% step 5520 | loss 3.1815 | lr 6.07e-05 | gnorm 0.19 | 173718 tok/s | mfu 31.3% step 5530 | loss 3.1713 | lr 6.06e-05 | gnorm 0.19 | 173750 tok/s | mfu 31.3% step 5540 | loss 3.1585 | lr 6.05e-05 | gnorm 0.19 | 173624 tok/s | mfu 31.3% step 5550 | loss 3.1865 | lr 6.04e-05 | gnorm 0.19 | 173729 tok/s | mfu 31.3% step 5560 | loss 3.1664 | lr 6.03e-05 | gnorm 0.20 | 173727 tok/s | mfu 31.5% step 5570 | loss 3.1399 | lr 6.02e-05 | gnorm 0.19 | 173569 tok/s | mfu 31.5% step 5580 | loss 3.1673 | lr 6.02e-05 | gnorm 0.20 | 173697 tok/s | mfu 31.5% step 5590 | loss 3.1603 | lr 6.01e-05 | gnorm 0.19 | 173671 tok/s | mfu 31.5% step 5600 | loss 3.1794 | lr 6.01e-05 | gnorm 0.19 | 173661 tok/s | mfu 31.5% step 5610 | loss 3.1728 | lr 6.00e-05 | gnorm 0.19 | 173389 tok/s | mfu 31.5% step 5620 | loss 3.1530 | lr 6.00e-05 | gnorm 0.19 | 173784 tok/s | mfu 31.5% step 5630 | loss 3.1587 | lr 6.00e-05 | gnorm 0.19 | 173604 tok/s | mfu 31.5%