| { | |
| "best_metric": 9.4224620303757, | |
| "best_model_checkpoint": "kotoba_v2_enc_logs_epoch2_2/checkpoint-500", | |
| "epoch": 0.0034266994716029417, | |
| "eval_steps": 500, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 3.4266994716029415e-05, | |
| "grad_norm": 1.0561553239822388, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2361, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 6.853398943205883e-05, | |
| "grad_norm": 1.1626238822937012, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2265, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.00010280098414808825, | |
| "grad_norm": 0.9845689535140991, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2279, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.00013706797886411766, | |
| "grad_norm": 1.142356276512146, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2382, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.00017133497358014707, | |
| "grad_norm": 1.0053240060806274, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2473, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0002056019682961765, | |
| "grad_norm": 1.1098105907440186, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2438, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.0002398689630122059, | |
| "grad_norm": 1.191983699798584, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2293, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0002741359577282353, | |
| "grad_norm": 1.1295104026794434, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2362, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0003084029524442647, | |
| "grad_norm": 1.037972092628479, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2455, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.00034266994716029413, | |
| "grad_norm": 1.1975648403167725, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2459, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.00037693694187632354, | |
| "grad_norm": 1.0676342248916626, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2271, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.000411203936592353, | |
| "grad_norm": 1.0749495029449463, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2417, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.0004454709313083824, | |
| "grad_norm": 1.094260811805725, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2354, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.0004797379260244118, | |
| "grad_norm": 1.0395853519439697, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2381, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.0005140049207404412, | |
| "grad_norm": 1.2008885145187378, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2354, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.0005482719154564706, | |
| "grad_norm": 1.0647832155227661, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2321, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.0005825389101725, | |
| "grad_norm": 1.327071189880371, | |
| "learning_rate": 1e-05, | |
| "loss": 0.238, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.0006168059048885295, | |
| "grad_norm": 1.1184055805206299, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2242, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.0006510728996045589, | |
| "grad_norm": 1.2512784004211426, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2437, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.0006853398943205883, | |
| "grad_norm": 1.0614465475082397, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2382, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.0007196068890366177, | |
| "grad_norm": 1.0607149600982666, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2381, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.0007538738837526471, | |
| "grad_norm": 1.0422028303146362, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2294, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.0007881408784686765, | |
| "grad_norm": 1.0162984132766724, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2275, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.000822407873184706, | |
| "grad_norm": 1.1085543632507324, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2161, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.0008566748679007354, | |
| "grad_norm": 1.1854636669158936, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2382, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.0008909418626167648, | |
| "grad_norm": 1.40137779712677, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2579, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.0009252088573327942, | |
| "grad_norm": 1.0814112424850464, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2612, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.0009594758520488236, | |
| "grad_norm": 1.083736538887024, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2711, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.000993742846764853, | |
| "grad_norm": 1.0861411094665527, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2642, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.0010280098414808825, | |
| "grad_norm": 1.1141265630722046, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2585, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.0010622768361969119, | |
| "grad_norm": 1.326241374015808, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2858, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.0010965438309129413, | |
| "grad_norm": 1.393750786781311, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2635, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.0011308108256289707, | |
| "grad_norm": 1.0851459503173828, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2565, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.001165077820345, | |
| "grad_norm": 1.2323757410049438, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2465, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.0011993448150610295, | |
| "grad_norm": 1.376953125, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2671, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.001233611809777059, | |
| "grad_norm": 1.084592580795288, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2643, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.0012678788044930883, | |
| "grad_norm": 1.2907005548477173, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2584, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.0013021457992091177, | |
| "grad_norm": 1.0698130130767822, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2526, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.0013364127939251471, | |
| "grad_norm": 1.1399807929992676, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2759, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.0013706797886411765, | |
| "grad_norm": 1.1480791568756104, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2499, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.001404946783357206, | |
| "grad_norm": 1.3095237016677856, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2536, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.0014392137780732353, | |
| "grad_norm": 1.068246841430664, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2604, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.0014734807727892648, | |
| "grad_norm": 1.2310419082641602, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2632, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.0015077477675052942, | |
| "grad_norm": 1.161867380142212, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2584, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.0015420147622213236, | |
| "grad_norm": 1.1461217403411865, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2592, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.001576281756937353, | |
| "grad_norm": 1.3006030321121216, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2607, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.0016105487516533824, | |
| "grad_norm": 1.1223125457763672, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2433, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.001644815746369412, | |
| "grad_norm": 1.2909380197525024, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2693, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.0016790827410854414, | |
| "grad_norm": 1.2270597219467163, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2661, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.0017133497358014708, | |
| "grad_norm": 1.1439770460128784, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2517, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.0017133497358014708, | |
| "eval_cer": 13.0358087846181, | |
| "eval_loss": 0.25224336981773376, | |
| "eval_normalized_cer": 9.4224620303757, | |
| "eval_runtime": 227.2174, | |
| "eval_samples_per_second": 2.253, | |
| "eval_steps_per_second": 0.035, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.0017476167305175002, | |
| "grad_norm": 1.1377454996109009, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2579, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.0017818837252335296, | |
| "grad_norm": 1.2096498012542725, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2727, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.001816150719949559, | |
| "grad_norm": 1.187213659286499, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2562, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.0018504177146655885, | |
| "grad_norm": 0.969393253326416, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2378, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.0018846847093816179, | |
| "grad_norm": 0.9745528697967529, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2774, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.0019189517040976473, | |
| "grad_norm": 1.0725352764129639, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2541, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.0019532186988136767, | |
| "grad_norm": 1.217871904373169, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2395, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.001987485693529706, | |
| "grad_norm": 1.3582627773284912, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2594, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.0020217526882457355, | |
| "grad_norm": 1.2415379285812378, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2582, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.002056019682961765, | |
| "grad_norm": 0.9810131192207336, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2284, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.0020902866776777943, | |
| "grad_norm": 0.9806564450263977, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2688, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.0021245536723938237, | |
| "grad_norm": 1.2755467891693115, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2591, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.002158820667109853, | |
| "grad_norm": 0.9300326704978943, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2444, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.0021930876618258825, | |
| "grad_norm": 1.1276524066925049, | |
| "learning_rate": 1e-05, | |
| "loss": 0.236, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.002227354656541912, | |
| "grad_norm": 1.1786876916885376, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2443, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.0022616216512579414, | |
| "grad_norm": 1.1702712774276733, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2627, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.0022958886459739708, | |
| "grad_norm": 1.2837899923324585, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2378, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.00233015564069, | |
| "grad_norm": 1.0623608827590942, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2491, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.0023644226354060296, | |
| "grad_norm": 1.1288243532180786, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2773, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.002398689630122059, | |
| "grad_norm": 1.0192692279815674, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2492, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.0024329566248380884, | |
| "grad_norm": 1.2274680137634277, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2345, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.002467223619554118, | |
| "grad_norm": 1.240645170211792, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2624, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.002501490614270147, | |
| "grad_norm": 1.0681366920471191, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2553, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.0025357576089861766, | |
| "grad_norm": 1.0161867141723633, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2547, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.002570024603702206, | |
| "grad_norm": 1.2384017705917358, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2449, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.0026042915984182354, | |
| "grad_norm": 1.1739261150360107, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2523, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.002638558593134265, | |
| "grad_norm": 1.0396535396575928, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2535, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.0026728255878502943, | |
| "grad_norm": 1.14767324924469, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2594, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.0027070925825663237, | |
| "grad_norm": 1.1783303022384644, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2546, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.002741359577282353, | |
| "grad_norm": 1.1065645217895508, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2547, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.0027756265719983825, | |
| "grad_norm": 1.256645917892456, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2548, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.002809893566714412, | |
| "grad_norm": 1.058158278465271, | |
| "learning_rate": 1e-05, | |
| "loss": 0.257, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.0028441605614304413, | |
| "grad_norm": 1.0647656917572021, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2479, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.0028784275561464707, | |
| "grad_norm": 1.1984691619873047, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2503, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.0029126945508625, | |
| "grad_norm": 1.1380070447921753, | |
| "learning_rate": 1e-05, | |
| "loss": 0.245, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.0029469615455785295, | |
| "grad_norm": 1.2131065130233765, | |
| "learning_rate": 1e-05, | |
| "loss": 0.242, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.002981228540294559, | |
| "grad_norm": 1.1822234392166138, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2613, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.0030154955350105883, | |
| "grad_norm": 1.0591018199920654, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2654, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.0030497625297266177, | |
| "grad_norm": 1.2318428754806519, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2525, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.003084029524442647, | |
| "grad_norm": 1.0146839618682861, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2609, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.0031182965191586766, | |
| "grad_norm": 1.1508561372756958, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2541, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.003152563513874706, | |
| "grad_norm": 1.1494849920272827, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2461, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.0031868305085907354, | |
| "grad_norm": 1.2423807382583618, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2573, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.0032210975033067648, | |
| "grad_norm": 1.2714438438415527, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2545, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.0032553644980227946, | |
| "grad_norm": 1.2088007926940918, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2773, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.003289631492738824, | |
| "grad_norm": 1.0737963914871216, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2495, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.0033238984874548534, | |
| "grad_norm": 1.0942472219467163, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2401, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.003358165482170883, | |
| "grad_norm": 1.1282986402511597, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2638, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.0033924324768869123, | |
| "grad_norm": 1.0762425661087036, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2619, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.0034266994716029417, | |
| "grad_norm": 1.09200119972229, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2464, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.0034266994716029417, | |
| "eval_cer": 13.80313988357735, | |
| "eval_loss": 0.25397512316703796, | |
| "eval_normalized_cer": 9.952038369304557, | |
| "eval_runtime": 227.5088, | |
| "eval_samples_per_second": 2.25, | |
| "eval_steps_per_second": 0.035, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 291826, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9223372036854775807, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.0275903070208e+20, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |