| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.56, | |
| "eval_steps": 500, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 8.571518898010254, | |
| "learning_rate": 0.0003965811965811966, | |
| "loss": 6.0892, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.477086067199707, | |
| "learning_rate": 0.00039316239316239317, | |
| "loss": 1.4485, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.7137540578842163, | |
| "learning_rate": 0.00038974358974358975, | |
| "loss": 0.9732, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.5362057685852051, | |
| "learning_rate": 0.0003863247863247863, | |
| "loss": 0.7804, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.6810179948806763, | |
| "learning_rate": 0.00038290598290598296, | |
| "loss": 0.665, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.4986821115016937, | |
| "learning_rate": 0.0003794871794871795, | |
| "loss": 0.6091, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 0.4309682250022888, | |
| "learning_rate": 0.00037606837606837606, | |
| "loss": 0.5502, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.3824257552623749, | |
| "learning_rate": 0.0003726495726495727, | |
| "loss": 0.5164, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 0.3188970386981964, | |
| "learning_rate": 0.00036923076923076927, | |
| "loss": 0.4883, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.3359103202819824, | |
| "learning_rate": 0.00036581196581196584, | |
| "loss": 0.4612, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 0.4327464699745178, | |
| "learning_rate": 0.0003623931623931624, | |
| "loss": 0.4351, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.4424777030944824, | |
| "learning_rate": 0.000358974358974359, | |
| "loss": 0.4217, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 0.520322322845459, | |
| "learning_rate": 0.00035555555555555557, | |
| "loss": 0.4076, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 0.48572778701782227, | |
| "learning_rate": 0.00035213675213675215, | |
| "loss": 0.3948, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 0.2985605001449585, | |
| "learning_rate": 0.0003487179487179487, | |
| "loss": 0.3823, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 0.28738752007484436, | |
| "learning_rate": 0.00034529914529914536, | |
| "loss": 0.375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 0.29423144459724426, | |
| "learning_rate": 0.0003418803418803419, | |
| "loss": 0.3591, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 0.26430046558380127, | |
| "learning_rate": 0.00033846153846153846, | |
| "loss": 0.3494, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 0.2734215259552002, | |
| "learning_rate": 0.0003350427350427351, | |
| "loss": 0.3396, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 0.3005197048187256, | |
| "learning_rate": 0.00033162393162393166, | |
| "loss": 0.3352, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 0.2822723686695099, | |
| "learning_rate": 0.0003282051282051282, | |
| "loss": 0.3241, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 0.2792316973209381, | |
| "learning_rate": 0.0003247863247863248, | |
| "loss": 0.3208, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 0.2761669158935547, | |
| "learning_rate": 0.0003213675213675214, | |
| "loss": 0.3148, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 0.2733113467693329, | |
| "learning_rate": 0.0003179487179487179, | |
| "loss": 0.311, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 0.3393694758415222, | |
| "learning_rate": 0.00031452991452991455, | |
| "loss": 0.3056, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.29316985607147217, | |
| "learning_rate": 0.0003111111111111111, | |
| "loss": 0.2975, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 0.29134783148765564, | |
| "learning_rate": 0.0003076923076923077, | |
| "loss": 0.2898, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 0.41234660148620605, | |
| "learning_rate": 0.0003042735042735043, | |
| "loss": 0.2895, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 0.26693716645240784, | |
| "learning_rate": 0.00030085470085470086, | |
| "loss": 0.2835, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 0.2862294614315033, | |
| "learning_rate": 0.00029743589743589743, | |
| "loss": 0.2747, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 0.2596249282360077, | |
| "learning_rate": 0.00029401709401709406, | |
| "loss": 0.2752, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 0.2555866241455078, | |
| "learning_rate": 0.0002905982905982906, | |
| "loss": 0.2662, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 0.3845195472240448, | |
| "learning_rate": 0.0002871794871794872, | |
| "loss": 0.2625, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 0.23550209403038025, | |
| "learning_rate": 0.0002837606837606838, | |
| "loss": 0.256, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 0.2434936910867691, | |
| "learning_rate": 0.0002803418803418803, | |
| "loss": 0.2545, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 0.23562268912792206, | |
| "learning_rate": 0.00027692307692307695, | |
| "loss": 0.2536, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 0.3110085427761078, | |
| "learning_rate": 0.0002735042735042735, | |
| "loss": 0.2497, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 0.2646142244338989, | |
| "learning_rate": 0.0002700854700854701, | |
| "loss": 0.2448, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.22812116146087646, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.2409, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 0.21481893956661224, | |
| "learning_rate": 0.00026324786324786326, | |
| "loss": 0.225, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 0.2561526298522949, | |
| "learning_rate": 0.00025982905982905983, | |
| "loss": 0.2194, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 0.2297515720129013, | |
| "learning_rate": 0.00025641025641025646, | |
| "loss": 0.216, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 0.25526463985443115, | |
| "learning_rate": 0.000252991452991453, | |
| "loss": 0.2171, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 0.24202637374401093, | |
| "learning_rate": 0.00024957264957264956, | |
| "loss": 0.2149, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 0.20644807815551758, | |
| "learning_rate": 0.0002461538461538462, | |
| "loss": 0.209, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 0.2795998454093933, | |
| "learning_rate": 0.00024273504273504272, | |
| "loss": 0.2071, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.306149959564209, | |
| "learning_rate": 0.00023931623931623932, | |
| "loss": 0.208, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 0.2355523407459259, | |
| "learning_rate": 0.00023589743589743593, | |
| "loss": 0.2051, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 0.2909263074398041, | |
| "learning_rate": 0.0002324786324786325, | |
| "loss": 0.2023, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 0.5298261642456055, | |
| "learning_rate": 0.00022905982905982905, | |
| "loss": 0.2018, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_cer": 0.9319083335386478, | |
| "eval_loss": 0.17038685083389282, | |
| "eval_runtime": 138.0008, | |
| "eval_samples_per_second": 14.493, | |
| "eval_steps_per_second": 0.457, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 0.23548871278762817, | |
| "learning_rate": 0.00022564102564102566, | |
| "loss": 0.2008, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 0.20162977278232574, | |
| "learning_rate": 0.00022222222222222223, | |
| "loss": 0.1975, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 0.2593408524990082, | |
| "learning_rate": 0.00021880341880341884, | |
| "loss": 0.1959, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 0.21452312171459198, | |
| "learning_rate": 0.0002153846153846154, | |
| "loss": 0.194, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 0.2637544274330139, | |
| "learning_rate": 0.000211965811965812, | |
| "loss": 0.1909, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 0.24357128143310547, | |
| "learning_rate": 0.00020854700854700857, | |
| "loss": 0.19, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 0.2084117978811264, | |
| "learning_rate": 0.00020512820512820512, | |
| "loss": 0.1857, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "grad_norm": 0.23245294392108917, | |
| "learning_rate": 0.00020170940170940172, | |
| "loss": 0.1858, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 0.23836293816566467, | |
| "learning_rate": 0.0001982905982905983, | |
| "loss": 0.183, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.19184565544128418, | |
| "learning_rate": 0.00019487179487179487, | |
| "loss": 0.183, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 0.20401564240455627, | |
| "learning_rate": 0.00019145299145299148, | |
| "loss": 0.1829, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "grad_norm": 0.21579188108444214, | |
| "learning_rate": 0.00018803418803418803, | |
| "loss": 0.1812, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 0.23108145594596863, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.1781, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 0.2311713844537735, | |
| "learning_rate": 0.0001811965811965812, | |
| "loss": 0.1755, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "grad_norm": 0.19794794917106628, | |
| "learning_rate": 0.00017777777777777779, | |
| "loss": 0.1768, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 0.2516119182109833, | |
| "learning_rate": 0.00017435897435897436, | |
| "loss": 0.1737, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 0.20975567400455475, | |
| "learning_rate": 0.00017094017094017094, | |
| "loss": 0.1712, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "grad_norm": 0.22168505191802979, | |
| "learning_rate": 0.00016752136752136754, | |
| "loss": 0.1693, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 0.22844062745571136, | |
| "learning_rate": 0.0001641025641025641, | |
| "loss": 0.168, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 0.22804197669029236, | |
| "learning_rate": 0.0001606837606837607, | |
| "loss": 0.1721, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 0.22620578110218048, | |
| "learning_rate": 0.00015726495726495727, | |
| "loss": 0.1703, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 0.21445313096046448, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.1673, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 0.207479327917099, | |
| "learning_rate": 0.00015042735042735043, | |
| "loss": 0.1648, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 0.22134087979793549, | |
| "learning_rate": 0.00014700854700854703, | |
| "loss": 0.1629, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 0.20121484994888306, | |
| "learning_rate": 0.0001435897435897436, | |
| "loss": 0.1638, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 0.2002618908882141, | |
| "learning_rate": 0.00014017094017094016, | |
| "loss": 0.1621, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 0.19750453531742096, | |
| "learning_rate": 0.00013675213675213676, | |
| "loss": 0.1667, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.22286508977413177, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.1642, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 0.21668635308742523, | |
| "learning_rate": 0.00012991452991452992, | |
| "loss": 0.1482, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 0.233961820602417, | |
| "learning_rate": 0.0001264957264957265, | |
| "loss": 0.1453, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "grad_norm": 0.1865084022283554, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.1455, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.1853141337633133, | |
| "learning_rate": 0.00011965811965811966, | |
| "loss": 0.1442, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "grad_norm": 0.17371739447116852, | |
| "learning_rate": 0.00011623931623931625, | |
| "loss": 0.1382, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 0.19631154835224152, | |
| "learning_rate": 0.00011282051282051283, | |
| "loss": 0.1415, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 0.194850891828537, | |
| "learning_rate": 0.00010940170940170942, | |
| "loss": 0.141, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 0.18121449649333954, | |
| "learning_rate": 0.000105982905982906, | |
| "loss": 0.1388, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "grad_norm": 0.2176773101091385, | |
| "learning_rate": 0.00010256410256410256, | |
| "loss": 0.1399, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "grad_norm": 0.19013133645057678, | |
| "learning_rate": 9.914529914529915e-05, | |
| "loss": 0.137, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "grad_norm": 0.22148679196834564, | |
| "learning_rate": 9.572649572649574e-05, | |
| "loss": 0.139, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "grad_norm": 0.20861493051052094, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.139, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "grad_norm": 0.17541790008544922, | |
| "learning_rate": 8.888888888888889e-05, | |
| "loss": 0.1362, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "grad_norm": 0.1971459984779358, | |
| "learning_rate": 8.547008547008547e-05, | |
| "loss": 0.1346, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "grad_norm": 0.20883004367351532, | |
| "learning_rate": 8.205128205128205e-05, | |
| "loss": 0.1351, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "grad_norm": 0.18058577179908752, | |
| "learning_rate": 7.863247863247864e-05, | |
| "loss": 0.1363, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "grad_norm": 0.19193512201309204, | |
| "learning_rate": 7.521367521367521e-05, | |
| "loss": 0.1359, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "grad_norm": 0.17777132987976074, | |
| "learning_rate": 7.17948717948718e-05, | |
| "loss": 0.1363, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 0.1730206310749054, | |
| "learning_rate": 6.837606837606838e-05, | |
| "loss": 0.1339, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 0.172698512673378, | |
| "learning_rate": 6.495726495726496e-05, | |
| "loss": 0.1317, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 0.1746242642402649, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.132, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 0.1631608009338379, | |
| "learning_rate": 5.8119658119658126e-05, | |
| "loss": 0.1292, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "eval_cer": 0.9320702386692806, | |
| "eval_loss": 0.11450555920600891, | |
| "eval_runtime": 136.638, | |
| "eval_samples_per_second": 14.637, | |
| "eval_steps_per_second": 0.461, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1170, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 6.737241075941376e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |