| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 1258, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.007950705625124229, |
| "grad_norm": 219.54016494389, |
| "learning_rate": 7.142857142857143e-07, |
| "loss": 1.7018884658813476, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.015901411250248458, |
| "grad_norm": 25.5689285955615, |
| "learning_rate": 1.507936507936508e-06, |
| "loss": 0.8416803359985352, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.02385211687537269, |
| "grad_norm": 12.567087787318425, |
| "learning_rate": 2.301587301587302e-06, |
| "loss": 0.42536497116088867, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.031802822500496916, |
| "grad_norm": 9.923579832187075, |
| "learning_rate": 3.0952380952380957e-06, |
| "loss": 0.2580261707305908, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.03975352812562115, |
| "grad_norm": 6.740346825469367, |
| "learning_rate": 3.88888888888889e-06, |
| "loss": 0.21076388359069825, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.04770423375074538, |
| "grad_norm": 4.415378230152806, |
| "learning_rate": 4.682539682539683e-06, |
| "loss": 0.20317745208740234, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.05565493937586961, |
| "grad_norm": 4.542282886024745, |
| "learning_rate": 5.476190476190477e-06, |
| "loss": 0.2051560878753662, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.06360564500099383, |
| "grad_norm": 4.140509241218129, |
| "learning_rate": 6.26984126984127e-06, |
| "loss": 0.20307469367980957, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.07155635062611806, |
| "grad_norm": 3.4397313426405622, |
| "learning_rate": 7.063492063492064e-06, |
| "loss": 0.19124785661697388, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.0795070562512423, |
| "grad_norm": 3.1396243716144037, |
| "learning_rate": 7.857142857142858e-06, |
| "loss": 0.18574079275131225, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.08745776187636653, |
| "grad_norm": 4.446541510757395, |
| "learning_rate": 8.650793650793651e-06, |
| "loss": 0.18685503005981446, |
| "step": 110 |
| }, |
| { |
| "epoch": 0.09540846750149076, |
| "grad_norm": 5.735669031144987, |
| "learning_rate": 9.444444444444445e-06, |
| "loss": 0.19291961193084717, |
| "step": 120 |
| }, |
| { |
| "epoch": 0.10335917312661498, |
| "grad_norm": 3.874442862167701, |
| "learning_rate": 9.999826704653376e-06, |
| "loss": 0.19922959804534912, |
| "step": 130 |
| }, |
| { |
| "epoch": 0.11130987875173921, |
| "grad_norm": 2.9212784210722593, |
| "learning_rate": 9.996746232654956e-06, |
| "loss": 0.1822590112686157, |
| "step": 140 |
| }, |
| { |
| "epoch": 0.11926058437686345, |
| "grad_norm": 3.191386814899392, |
| "learning_rate": 9.989817483785946e-06, |
| "loss": 0.19430736303329468, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.12721129000198766, |
| "grad_norm": 2.6740381318443553, |
| "learning_rate": 9.979045794267713e-06, |
| "loss": 0.18704640865325928, |
| "step": 160 |
| }, |
| { |
| "epoch": 0.1351619956271119, |
| "grad_norm": 2.9782125248494005, |
| "learning_rate": 9.964439459987563e-06, |
| "loss": 0.18864736557006836, |
| "step": 170 |
| }, |
| { |
| "epoch": 0.14311270125223613, |
| "grad_norm": 2.710615719782371, |
| "learning_rate": 9.9460097301096e-06, |
| "loss": 0.17778729200363158, |
| "step": 180 |
| }, |
| { |
| "epoch": 0.15106340687736036, |
| "grad_norm": 3.7180731721983404, |
| "learning_rate": 9.923770798411121e-06, |
| "loss": 0.1797666072845459, |
| "step": 190 |
| }, |
| { |
| "epoch": 0.1590141125024846, |
| "grad_norm": 2.632392593320569, |
| "learning_rate": 9.897739792351174e-06, |
| "loss": 0.20337791442871095, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.16696481812760883, |
| "grad_norm": 2.30889007019316, |
| "learning_rate": 9.86793675987974e-06, |
| "loss": 0.1722043991088867, |
| "step": 210 |
| }, |
| { |
| "epoch": 0.17491552375273306, |
| "grad_norm": 2.2788339334399894, |
| "learning_rate": 9.83438465399767e-06, |
| "loss": 0.1788644552230835, |
| "step": 220 |
| }, |
| { |
| "epoch": 0.1828662293778573, |
| "grad_norm": 4.09430348910076, |
| "learning_rate": 9.797109315079285e-06, |
| "loss": 0.18456263542175294, |
| "step": 230 |
| }, |
| { |
| "epoch": 0.19081693500298152, |
| "grad_norm": 2.5088383006458956, |
| "learning_rate": 9.756139450971234e-06, |
| "loss": 0.18333306312561035, |
| "step": 240 |
| }, |
| { |
| "epoch": 0.19876764062810576, |
| "grad_norm": 2.281918898605021, |
| "learning_rate": 9.711506614882965e-06, |
| "loss": 0.17209144830703735, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.20671834625322996, |
| "grad_norm": 1.9125682021670853, |
| "learning_rate": 9.663245181085812e-06, |
| "loss": 0.16443560123443604, |
| "step": 260 |
| }, |
| { |
| "epoch": 0.2146690518783542, |
| "grad_norm": 1.6339937971275729, |
| "learning_rate": 9.611392318439427e-06, |
| "loss": 0.18000526428222657, |
| "step": 270 |
| }, |
| { |
| "epoch": 0.22261975750347843, |
| "grad_norm": 2.9164760702944186, |
| "learning_rate": 9.555987961765929e-06, |
| "loss": 0.17941322326660156, |
| "step": 280 |
| }, |
| { |
| "epoch": 0.23057046312860266, |
| "grad_norm": 2.040893165255211, |
| "learning_rate": 9.497074781093851e-06, |
| "loss": 0.18319871425628662, |
| "step": 290 |
| }, |
| { |
| "epoch": 0.2385211687537269, |
| "grad_norm": 3.3491731821507256, |
| "learning_rate": 9.434698148795534e-06, |
| "loss": 0.17000328302383422, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.24647187437885112, |
| "grad_norm": 2.616304309363476, |
| "learning_rate": 9.368906104643295e-06, |
| "loss": 0.16073766946792603, |
| "step": 310 |
| }, |
| { |
| "epoch": 0.25442258000397533, |
| "grad_norm": 3.036419074712572, |
| "learning_rate": 9.29974931881128e-06, |
| "loss": 0.14889835119247435, |
| "step": 320 |
| }, |
| { |
| "epoch": 0.26237328562909956, |
| "grad_norm": 2.235034315825349, |
| "learning_rate": 9.2272810528515e-06, |
| "loss": 0.18206405639648438, |
| "step": 330 |
| }, |
| { |
| "epoch": 0.2703239912542238, |
| "grad_norm": 2.0387554029286297, |
| "learning_rate": 9.1515571186741e-06, |
| "loss": 0.1785102128982544, |
| "step": 340 |
| }, |
| { |
| "epoch": 0.278274696879348, |
| "grad_norm": 2.3223395798285718, |
| "learning_rate": 9.072635835563446e-06, |
| "loss": 0.16460620164871215, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.28622540250447226, |
| "grad_norm": 2.1765980392575104, |
| "learning_rate": 8.990577985263144e-06, |
| "loss": 0.18332879543304442, |
| "step": 360 |
| }, |
| { |
| "epoch": 0.2941761081295965, |
| "grad_norm": 2.82432523037995, |
| "learning_rate": 8.90544676516459e-06, |
| "loss": 0.16783593893051146, |
| "step": 370 |
| }, |
| { |
| "epoch": 0.3021268137547207, |
| "grad_norm": 3.062275075339145, |
| "learning_rate": 8.817307739635079e-06, |
| "loss": 0.17636168003082275, |
| "step": 380 |
| }, |
| { |
| "epoch": 0.31007751937984496, |
| "grad_norm": 2.02878205344946, |
| "learning_rate": 8.726228789522974e-06, |
| "loss": 0.1478671669960022, |
| "step": 390 |
| }, |
| { |
| "epoch": 0.3180282250049692, |
| "grad_norm": 2.3162091329963324, |
| "learning_rate": 8.632280059878835e-06, |
| "loss": 0.17146013975143432, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.3259789306300934, |
| "grad_norm": 2.2289945520335226, |
| "learning_rate": 8.535533905932739e-06, |
| "loss": 0.16558258533477782, |
| "step": 410 |
| }, |
| { |
| "epoch": 0.33392963625521765, |
| "grad_norm": 2.157454374800103, |
| "learning_rate": 8.436064837369433e-06, |
| "loss": 0.17361073493957518, |
| "step": 420 |
| }, |
| { |
| "epoch": 0.3418803418803419, |
| "grad_norm": 2.038678467516046, |
| "learning_rate": 8.333949460944206e-06, |
| "loss": 0.17476344108581543, |
| "step": 430 |
| }, |
| { |
| "epoch": 0.3498310475054661, |
| "grad_norm": 2.0417260830139, |
| "learning_rate": 8.2292664214837e-06, |
| "loss": 0.17707545757293702, |
| "step": 440 |
| }, |
| { |
| "epoch": 0.35778175313059035, |
| "grad_norm": 2.130327795287117, |
| "learning_rate": 8.122096341317071e-06, |
| "loss": 0.1499488115310669, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.3657324587557146, |
| "grad_norm": 2.1750357372101754, |
| "learning_rate": 8.01252175818419e-06, |
| "loss": 0.16685774326324462, |
| "step": 460 |
| }, |
| { |
| "epoch": 0.3736831643808388, |
| "grad_norm": 2.205396355836447, |
| "learning_rate": 7.90062706166864e-06, |
| "loss": 0.15516927242279052, |
| "step": 470 |
| }, |
| { |
| "epoch": 0.38163387000596305, |
| "grad_norm": 1.6893210650312653, |
| "learning_rate": 7.786498428204547e-06, |
| "loss": 0.16797863245010375, |
| "step": 480 |
| }, |
| { |
| "epoch": 0.3895845756310873, |
| "grad_norm": 1.787973908712826, |
| "learning_rate": 7.67022375470722e-06, |
| "loss": 0.17217243909835817, |
| "step": 490 |
| }, |
| { |
| "epoch": 0.3975352812562115, |
| "grad_norm": 2.12364662089689, |
| "learning_rate": 7.551892590878766e-06, |
| "loss": 0.16128942966461182, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.4054859868813357, |
| "grad_norm": 1.5940988121150617, |
| "learning_rate": 7.431596070240812e-06, |
| "loss": 0.15976805686950685, |
| "step": 510 |
| }, |
| { |
| "epoch": 0.4134366925064599, |
| "grad_norm": 2.5063127610357716, |
| "learning_rate": 7.309426839947407e-06, |
| "loss": 0.15266921520233154, |
| "step": 520 |
| }, |
| { |
| "epoch": 0.42138739813158416, |
| "grad_norm": 2.512553781152212, |
| "learning_rate": 7.185478989432225e-06, |
| "loss": 0.16032958030700684, |
| "step": 530 |
| }, |
| { |
| "epoch": 0.4293381037567084, |
| "grad_norm": 1.971625874355437, |
| "learning_rate": 7.059847977944962e-06, |
| "loss": 0.15459973812103273, |
| "step": 540 |
| }, |
| { |
| "epoch": 0.4372888093818326, |
| "grad_norm": 2.667141664710735, |
| "learning_rate": 6.932630561032776e-06, |
| "loss": 0.16032447814941406, |
| "step": 550 |
| }, |
| { |
| "epoch": 0.44523951500695685, |
| "grad_norm": 1.6533861969612529, |
| "learning_rate": 6.803924716023358e-06, |
| "loss": 0.1557694911956787, |
| "step": 560 |
| }, |
| { |
| "epoch": 0.4531902206320811, |
| "grad_norm": 1.7546319660173646, |
| "learning_rate": 6.673829566567059e-06, |
| "loss": 0.154685115814209, |
| "step": 570 |
| }, |
| { |
| "epoch": 0.4611409262572053, |
| "grad_norm": 2.156448823222435, |
| "learning_rate": 6.542445306296163e-06, |
| "loss": 0.1502530336380005, |
| "step": 580 |
| }, |
| { |
| "epoch": 0.46909163188232955, |
| "grad_norm": 1.7417715170150208, |
| "learning_rate": 6.409873121660088e-06, |
| "loss": 0.15132423639297485, |
| "step": 590 |
| }, |
| { |
| "epoch": 0.4770423375074538, |
| "grad_norm": 2.1319670561684747, |
| "learning_rate": 6.2762151139960026e-06, |
| "loss": 0.16759523153305053, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.484993043132578, |
| "grad_norm": 1.9598640513542338, |
| "learning_rate": 6.141574220894793e-06, |
| "loss": 0.16453773975372316, |
| "step": 610 |
| }, |
| { |
| "epoch": 0.49294374875770225, |
| "grad_norm": 1.70760086196145, |
| "learning_rate": 6.0060541369230055e-06, |
| "loss": 0.16313371658325196, |
| "step": 620 |
| }, |
| { |
| "epoch": 0.5008944543828264, |
| "grad_norm": 2.081992626179426, |
| "learning_rate": 5.869759233761807e-06, |
| "loss": 0.16269201040267944, |
| "step": 630 |
| }, |
| { |
| "epoch": 0.5088451600079507, |
| "grad_norm": 1.652904029133773, |
| "learning_rate": 5.73279447982444e-06, |
| "loss": 0.14864494800567626, |
| "step": 640 |
| }, |
| { |
| "epoch": 0.5167958656330749, |
| "grad_norm": 2.0691755916985297, |
| "learning_rate": 5.59526535941411e-06, |
| "loss": 0.1553126811981201, |
| "step": 650 |
| }, |
| { |
| "epoch": 0.5247465712581991, |
| "grad_norm": 1.902151511653258, |
| "learning_rate": 5.45727779148457e-06, |
| "loss": 0.15448397397994995, |
| "step": 660 |
| }, |
| { |
| "epoch": 0.5326972768833234, |
| "grad_norm": 1.8740749730610198, |
| "learning_rate": 5.318938048065926e-06, |
| "loss": 0.15808215141296386, |
| "step": 670 |
| }, |
| { |
| "epoch": 0.5406479825084476, |
| "grad_norm": 2.2468608784907627, |
| "learning_rate": 5.180352672418553e-06, |
| "loss": 0.1532147169113159, |
| "step": 680 |
| }, |
| { |
| "epoch": 0.5485986881335718, |
| "grad_norm": 1.5095202848596316, |
| "learning_rate": 5.041628396978093e-06, |
| "loss": 0.1534576892852783, |
| "step": 690 |
| }, |
| { |
| "epoch": 0.556549393758696, |
| "grad_norm": 1.451471074686343, |
| "learning_rate": 4.902872061154779e-06, |
| "loss": 0.16522419452667236, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.5645000993838203, |
| "grad_norm": 2.089090552764648, |
| "learning_rate": 4.7641905290503535e-06, |
| "loss": 0.1640038013458252, |
| "step": 710 |
| }, |
| { |
| "epoch": 0.5724508050089445, |
| "grad_norm": 1.5604030352322342, |
| "learning_rate": 4.625690607155987e-06, |
| "loss": 0.1599220633506775, |
| "step": 720 |
| }, |
| { |
| "epoch": 0.5804015106340688, |
| "grad_norm": 1.7098303052089416, |
| "learning_rate": 4.487478962094554e-06, |
| "loss": 0.15147488117218016, |
| "step": 730 |
| }, |
| { |
| "epoch": 0.588352216259193, |
| "grad_norm": 2.0090244106821946, |
| "learning_rate": 4.349662038470639e-06, |
| "loss": 0.1502910852432251, |
| "step": 740 |
| }, |
| { |
| "epoch": 0.5963029218843172, |
| "grad_norm": 1.7117397000999754, |
| "learning_rate": 4.21234597689152e-06, |
| "loss": 0.1602674126625061, |
| "step": 750 |
| }, |
| { |
| "epoch": 0.6042536275094414, |
| "grad_norm": 1.484756155862593, |
| "learning_rate": 4.0756365322222915e-06, |
| "loss": 0.1462937355041504, |
| "step": 760 |
| }, |
| { |
| "epoch": 0.6122043331345657, |
| "grad_norm": 2.11277258147162, |
| "learning_rate": 3.939638992138053e-06, |
| "loss": 0.15924113988876343, |
| "step": 770 |
| }, |
| { |
| "epoch": 0.6201550387596899, |
| "grad_norm": 2.453081338652199, |
| "learning_rate": 3.804458096035928e-06, |
| "loss": 0.14493399858474731, |
| "step": 780 |
| }, |
| { |
| "epoch": 0.6281057443848141, |
| "grad_norm": 1.7558673371889524, |
| "learning_rate": 3.670197954369311e-06, |
| "loss": 0.14885088205337524, |
| "step": 790 |
| }, |
| { |
| "epoch": 0.6360564500099384, |
| "grad_norm": 2.06447084029734, |
| "learning_rate": 3.5369619684665456e-06, |
| "loss": 0.1475375771522522, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.6440071556350626, |
| "grad_norm": 2.175240558099584, |
| "learning_rate": 3.4048527508956896e-06, |
| "loss": 0.1498485326766968, |
| "step": 810 |
| }, |
| { |
| "epoch": 0.6519578612601868, |
| "grad_norm": 1.697869444240965, |
| "learning_rate": 3.2739720464367848e-06, |
| "loss": 0.15282604694366456, |
| "step": 820 |
| }, |
| { |
| "epoch": 0.6599085668853111, |
| "grad_norm": 2.2627041073717353, |
| "learning_rate": 3.1444206537224477e-06, |
| "loss": 0.14622220993041993, |
| "step": 830 |
| }, |
| { |
| "epoch": 0.6678592725104353, |
| "grad_norm": 1.9502614996369985, |
| "learning_rate": 3.016298347607136e-06, |
| "loss": 0.14957189559936523, |
| "step": 840 |
| }, |
| { |
| "epoch": 0.6758099781355595, |
| "grad_norm": 1.6007670757083179, |
| "learning_rate": 2.889703802324897e-06, |
| "loss": 0.14099133014678955, |
| "step": 850 |
| }, |
| { |
| "epoch": 0.6837606837606838, |
| "grad_norm": 1.5805214374840055, |
| "learning_rate": 2.7647345154947557e-06, |
| "loss": 0.13873375654220582, |
| "step": 860 |
| }, |
| { |
| "epoch": 0.691711389385808, |
| "grad_norm": 1.8640779256147106, |
| "learning_rate": 2.6414867330322835e-06, |
| "loss": 0.14327547550201417, |
| "step": 870 |
| }, |
| { |
| "epoch": 0.6996620950109322, |
| "grad_norm": 1.7752317091411747, |
| "learning_rate": 2.520055375025178e-06, |
| "loss": 0.1548085927963257, |
| "step": 880 |
| }, |
| { |
| "epoch": 0.7076128006360565, |
| "grad_norm": 1.840951910844571, |
| "learning_rate": 2.4005339626299285e-06, |
| "loss": 0.13864609003067016, |
| "step": 890 |
| }, |
| { |
| "epoch": 0.7155635062611807, |
| "grad_norm": 1.9080116695594433, |
| "learning_rate": 2.283014546045899e-06, |
| "loss": 0.14006226062774657, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.7235142118863049, |
| "grad_norm": 1.7480792858075453, |
| "learning_rate": 2.1675876336222475e-06, |
| "loss": 0.14048845767974855, |
| "step": 910 |
| }, |
| { |
| "epoch": 0.7314649175114292, |
| "grad_norm": 2.3999895150411827, |
| "learning_rate": 2.054342122152334e-06, |
| "loss": 0.13474048376083375, |
| "step": 920 |
| }, |
| { |
| "epoch": 0.7394156231365534, |
| "grad_norm": 2.146278824012164, |
| "learning_rate": 1.943365228409273e-06, |
| "loss": 0.1452116847038269, |
| "step": 930 |
| }, |
| { |
| "epoch": 0.7473663287616776, |
| "grad_norm": 1.8039072971240762, |
| "learning_rate": 1.8347424219753674e-06, |
| "loss": 0.1531359076499939, |
| "step": 940 |
| }, |
| { |
| "epoch": 0.7553170343868019, |
| "grad_norm": 2.3035422636250846, |
| "learning_rate": 1.728557359417139e-06, |
| "loss": 0.14287930727005005, |
| "step": 950 |
| }, |
| { |
| "epoch": 0.7632677400119261, |
| "grad_norm": 1.70615453468217, |
| "learning_rate": 1.6248918198566832e-06, |
| "loss": 0.13460946083068848, |
| "step": 960 |
| }, |
| { |
| "epoch": 0.7712184456370503, |
| "grad_norm": 1.7733701208529742, |
| "learning_rate": 1.5238256419889447e-06, |
| "loss": 0.1354965090751648, |
| "step": 970 |
| }, |
| { |
| "epoch": 0.7791691512621746, |
| "grad_norm": 1.58623156014517, |
| "learning_rate": 1.425436662593413e-06, |
| "loss": 0.1442467212677002, |
| "step": 980 |
| }, |
| { |
| "epoch": 0.7871198568872988, |
| "grad_norm": 1.9404770203070638, |
| "learning_rate": 1.3298006565876249e-06, |
| "loss": 0.14258553981781005, |
| "step": 990 |
| }, |
| { |
| "epoch": 0.795070562512423, |
| "grad_norm": 1.4985066433130616, |
| "learning_rate": 1.23699127866861e-06, |
| "loss": 0.1358756184577942, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.8030212681375473, |
| "grad_norm": 2.230516031014272, |
| "learning_rate": 1.1470800065872489e-06, |
| "loss": 0.14560563564300538, |
| "step": 1010 |
| }, |
| { |
| "epoch": 0.8109719737626714, |
| "grad_norm": 1.9677432880405237, |
| "learning_rate": 1.060136086099206e-06, |
| "loss": 0.14896394014358522, |
| "step": 1020 |
| }, |
| { |
| "epoch": 0.8189226793877956, |
| "grad_norm": 1.6488156025523868, |
| "learning_rate": 9.76226477634869e-07, |
| "loss": 0.1371925115585327, |
| "step": 1030 |
| }, |
| { |
| "epoch": 0.8268733850129198, |
| "grad_norm": 2.0952600022358245, |
| "learning_rate": 8.954158047293172e-07, |
| "loss": 0.1430977463722229, |
| "step": 1040 |
| }, |
| { |
| "epoch": 0.8348240906380441, |
| "grad_norm": 1.797986127138805, |
| "learning_rate": 8.177663042520972e-07, |
| "loss": 0.14644017219543456, |
| "step": 1050 |
| }, |
| { |
| "epoch": 0.8427747962631683, |
| "grad_norm": 1.8628800359080484, |
| "learning_rate": 7.433377784750878e-07, |
| "loss": 0.13615297079086303, |
| "step": 1060 |
| }, |
| { |
| "epoch": 0.8507255018882925, |
| "grad_norm": 1.6347540056244632, |
| "learning_rate": 6.721875490153795e-07, |
| "loss": 0.13189680576324464, |
| "step": 1070 |
| }, |
| { |
| "epoch": 0.8586762075134168, |
| "grad_norm": 1.7110066201751546, |
| "learning_rate": 6.043704126886663e-07, |
| "loss": 0.14127652645111083, |
| "step": 1080 |
| }, |
| { |
| "epoch": 0.866626913138541, |
| "grad_norm": 1.6851082657954224, |
| "learning_rate": 5.399385993071093e-07, |
| "loss": 0.13686509132385255, |
| "step": 1090 |
| }, |
| { |
| "epoch": 0.8745776187636652, |
| "grad_norm": 1.8851496069288767, |
| "learning_rate": 4.789417314542099e-07, |
| "loss": 0.1334637999534607, |
| "step": 1100 |
| }, |
| { |
| "epoch": 0.8825283243887895, |
| "grad_norm": 1.9001545891850717, |
| "learning_rate": 4.214267862676508e-07, |
| "loss": 0.1359429121017456, |
| "step": 1110 |
| }, |
| { |
| "epoch": 0.8904790300139137, |
| "grad_norm": 2.0810985631909187, |
| "learning_rate": 3.6743805925954446e-07, |
| "loss": 0.13524262905120848, |
| "step": 1120 |
| }, |
| { |
| "epoch": 0.8984297356390379, |
| "grad_norm": 1.8909665966078972, |
| "learning_rate": 3.1701713020195757e-07, |
| "loss": 0.14022443294525147, |
| "step": 1130 |
| }, |
| { |
| "epoch": 0.9063804412641622, |
| "grad_norm": 2.129450992005257, |
| "learning_rate": 2.702028311039695e-07, |
| "loss": 0.12600768804550172, |
| "step": 1140 |
| }, |
| { |
| "epoch": 0.9143311468892864, |
| "grad_norm": 1.6810344399825308, |
| "learning_rate": 2.270312163049554e-07, |
| "loss": 0.13326795101165773, |
| "step": 1150 |
| }, |
| { |
| "epoch": 0.9222818525144106, |
| "grad_norm": 1.6826853060384712, |
| "learning_rate": 1.8753553470709063e-07, |
| "loss": 0.1351729989051819, |
| "step": 1160 |
| }, |
| { |
| "epoch": 0.9302325581395349, |
| "grad_norm": 2.1054384948186615, |
| "learning_rate": 1.5174620416849972e-07, |
| "loss": 0.12672573328018188, |
| "step": 1170 |
| }, |
| { |
| "epoch": 0.9381832637646591, |
| "grad_norm": 1.8093860593013356, |
| "learning_rate": 1.1969078807673728e-07, |
| "loss": 0.13967084884643555, |
| "step": 1180 |
| }, |
| { |
| "epoch": 0.9461339693897833, |
| "grad_norm": 1.637370708037661, |
| "learning_rate": 9.139397412066709e-08, |
| "loss": 0.1367415189743042, |
| "step": 1190 |
| }, |
| { |
| "epoch": 0.9540846750149076, |
| "grad_norm": 1.803095163089871, |
| "learning_rate": 6.687755527707673e-08, |
| "loss": 0.1392979145050049, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.9620353806400318, |
| "grad_norm": 1.7194583746309609, |
| "learning_rate": 4.6160413026671844e-08, |
| "loss": 0.13963855504989625, |
| "step": 1210 |
| }, |
| { |
| "epoch": 0.969986086265156, |
| "grad_norm": 1.6360162993347653, |
| "learning_rate": 2.925850281238152e-08, |
| "loss": 0.13864167928695678, |
| "step": 1220 |
| }, |
| { |
| "epoch": 0.9779367918902803, |
| "grad_norm": 2.0059460736807604, |
| "learning_rate": 1.6184841751168966e-08, |
| "loss": 0.1250883936882019, |
| "step": 1230 |
| }, |
| { |
| "epoch": 0.9858874975154045, |
| "grad_norm": 1.7799683961554524, |
| "learning_rate": 6.94949860881422e-09, |
| "loss": 0.13083313703536986, |
| "step": 1240 |
| }, |
| { |
| "epoch": 0.9938382031405287, |
| "grad_norm": 1.8264994970720634, |
| "learning_rate": 1.5595860453859745e-09, |
| "loss": 0.1367112636566162, |
| "step": 1250 |
| }, |
| { |
| "epoch": 1.0, |
| "step": 1258, |
| "total_flos": 45379846963200.0, |
| "train_loss": 0.1789945290084863, |
| "train_runtime": 18755.0654, |
| "train_samples_per_second": 4.292, |
| "train_steps_per_second": 0.067 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 1258, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 2000, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 45379846963200.0, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
