| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 29.75128998968008, | |
| "global_step": 14400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 24.132, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "eval_cer": 1.0, | |
| "eval_loss": 4.829305171966553, | |
| "eval_runtime": 666.1596, | |
| "eval_samples_per_second": 3.82, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0002935805991440799, | |
| "loss": 4.7165, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_cer": 1.0, | |
| "eval_loss": 4.607272148132324, | |
| "eval_runtime": 670.8713, | |
| "eval_samples_per_second": 3.794, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 0.00028502139800285305, | |
| "loss": 2.9637, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "eval_cer": 0.34719755926062756, | |
| "eval_loss": 1.3734972476959229, | |
| "eval_runtime": 671.6914, | |
| "eval_samples_per_second": 3.789, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "learning_rate": 0.0002764621968616262, | |
| "loss": 1.3373, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "eval_cer": 0.26771859991180924, | |
| "eval_loss": 0.9831737875938416, | |
| "eval_runtime": 669.2245, | |
| "eval_samples_per_second": 3.803, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.13, | |
| "learning_rate": 0.0002679029957203994, | |
| "loss": 1.0378, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.13, | |
| "eval_cer": 0.24304901738788448, | |
| "eval_loss": 0.8199232220649719, | |
| "eval_runtime": 670.0042, | |
| "eval_samples_per_second": 3.798, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "learning_rate": 0.00025934379457917263, | |
| "loss": 0.8747, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "eval_cer": 0.21944010773576136, | |
| "eval_loss": 0.7384957075119019, | |
| "eval_runtime": 673.0546, | |
| "eval_samples_per_second": 3.781, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 5.78, | |
| "learning_rate": 0.0002507845934379458, | |
| "loss": 0.7648, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.78, | |
| "eval_cer": 0.21221799806933703, | |
| "eval_loss": 0.7131306529045105, | |
| "eval_runtime": 671.747, | |
| "eval_samples_per_second": 3.789, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 6.61, | |
| "learning_rate": 0.00024222539229671895, | |
| "loss": 0.6863, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 6.61, | |
| "eval_cer": 0.19983553611650717, | |
| "eval_loss": 0.6777553558349609, | |
| "eval_runtime": 679.4426, | |
| "eval_samples_per_second": 3.746, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 7.44, | |
| "learning_rate": 0.00023366619115549212, | |
| "loss": 0.6269, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 7.44, | |
| "eval_cer": 0.18854950005362953, | |
| "eval_loss": 0.6489213705062866, | |
| "eval_runtime": 682.8113, | |
| "eval_samples_per_second": 3.727, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 8.26, | |
| "learning_rate": 0.00022510699001426532, | |
| "loss": 0.5893, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 8.26, | |
| "eval_cer": 0.19089728157885327, | |
| "eval_loss": 0.6407034993171692, | |
| "eval_runtime": 687.6153, | |
| "eval_samples_per_second": 3.701, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "learning_rate": 0.0002165477888730385, | |
| "loss": 0.5301, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "eval_cer": 0.1817445089322957, | |
| "eval_loss": 0.6261754035949707, | |
| "eval_runtime": 681.4189, | |
| "eval_samples_per_second": 3.735, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 9.92, | |
| "learning_rate": 0.0002079885877318117, | |
| "loss": 0.4909, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 9.92, | |
| "eval_cer": 0.17955165715239127, | |
| "eval_loss": 0.623394250869751, | |
| "eval_runtime": 672.8551, | |
| "eval_samples_per_second": 3.782, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 10.74, | |
| "learning_rate": 0.00019942938659058487, | |
| "loss": 0.4602, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 10.74, | |
| "eval_cer": 0.17870550238949337, | |
| "eval_loss": 0.6308777928352356, | |
| "eval_runtime": 673.8601, | |
| "eval_samples_per_second": 3.777, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 11.57, | |
| "learning_rate": 0.00019087018544935804, | |
| "loss": 0.4244, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 11.57, | |
| "eval_cer": 0.17706086355456507, | |
| "eval_loss": 0.6224568486213684, | |
| "eval_runtime": 677.416, | |
| "eval_samples_per_second": 3.757, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 12.4, | |
| "learning_rate": 0.0001823109843081312, | |
| "loss": 0.3936, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 12.4, | |
| "eval_cer": 0.17660799199132393, | |
| "eval_loss": 0.6143100261688232, | |
| "eval_runtime": 683.1583, | |
| "eval_samples_per_second": 3.725, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 13.22, | |
| "learning_rate": 0.0001737517831669044, | |
| "loss": 0.3792, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 13.22, | |
| "eval_cer": 0.17075641468733985, | |
| "eval_loss": 0.627468466758728, | |
| "eval_runtime": 676.725, | |
| "eval_samples_per_second": 3.761, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 14.05, | |
| "learning_rate": 0.00016519258202567759, | |
| "loss": 0.3555, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 14.05, | |
| "eval_cer": 0.16945738836120083, | |
| "eval_loss": 0.6296263933181763, | |
| "eval_runtime": 678.0284, | |
| "eval_samples_per_second": 3.754, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 14.88, | |
| "learning_rate": 0.00015663338088445076, | |
| "loss": 0.3282, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 14.88, | |
| "eval_cer": 0.16861123359830293, | |
| "eval_loss": 0.6353692412376404, | |
| "eval_runtime": 677.9049, | |
| "eval_samples_per_second": 3.754, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 15.7, | |
| "learning_rate": 0.00014807417974322396, | |
| "loss": 0.3105, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 15.7, | |
| "eval_cer": 0.16284308000333694, | |
| "eval_loss": 0.6332650780677795, | |
| "eval_runtime": 678.3344, | |
| "eval_samples_per_second": 3.752, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 16.53, | |
| "learning_rate": 0.00013951497860199713, | |
| "loss": 0.2899, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 16.53, | |
| "eval_cer": 0.16626345207307916, | |
| "eval_loss": 0.6411857604980469, | |
| "eval_runtime": 680.0199, | |
| "eval_samples_per_second": 3.743, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 17.36, | |
| "learning_rate": 0.0001309557774607703, | |
| "loss": 0.2769, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 17.36, | |
| "eval_cer": 0.1635343050209155, | |
| "eval_loss": 0.629165768623352, | |
| "eval_runtime": 679.3365, | |
| "eval_samples_per_second": 3.746, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 18.18, | |
| "learning_rate": 0.0001223965763195435, | |
| "loss": 0.2685, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 18.18, | |
| "eval_cer": 0.16531003825572943, | |
| "eval_loss": 0.6566583514213562, | |
| "eval_runtime": 677.0871, | |
| "eval_samples_per_second": 3.759, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 19.01, | |
| "learning_rate": 0.00011383737517831669, | |
| "loss": 0.2481, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 19.01, | |
| "eval_cer": 0.16478566065618705, | |
| "eval_loss": 0.652852475643158, | |
| "eval_runtime": 678.2412, | |
| "eval_samples_per_second": 3.752, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 19.83, | |
| "learning_rate": 0.00010527817403708987, | |
| "loss": 0.2355, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 19.83, | |
| "eval_cer": 0.1621041842948909, | |
| "eval_loss": 0.6581872701644897, | |
| "eval_runtime": 679.7908, | |
| "eval_samples_per_second": 3.744, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 20.66, | |
| "learning_rate": 9.671897289586304e-05, | |
| "loss": 0.2266, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 20.66, | |
| "eval_cer": 0.1623544554219452, | |
| "eval_loss": 0.6539607644081116, | |
| "eval_runtime": 679.8539, | |
| "eval_samples_per_second": 3.743, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 21.49, | |
| "learning_rate": 8.815977175463624e-05, | |
| "loss": 0.2183, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 21.49, | |
| "eval_cer": 0.1604952984781132, | |
| "eval_loss": 0.6687005758285522, | |
| "eval_runtime": 678.9594, | |
| "eval_samples_per_second": 3.748, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 22.31, | |
| "learning_rate": 7.960057061340941e-05, | |
| "loss": 0.2089, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 22.31, | |
| "eval_cer": 0.16017352131475765, | |
| "eval_loss": 0.6894858479499817, | |
| "eval_runtime": 681.5459, | |
| "eval_samples_per_second": 3.734, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 23.14, | |
| "learning_rate": 7.10413694721826e-05, | |
| "loss": 0.2016, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 23.14, | |
| "eval_cer": 0.1619850075677222, | |
| "eval_loss": 0.6835413575172424, | |
| "eval_runtime": 682.6349, | |
| "eval_samples_per_second": 3.728, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 23.97, | |
| "learning_rate": 6.248216833095577e-05, | |
| "loss": 0.1904, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 23.97, | |
| "eval_cer": 0.16073365193245062, | |
| "eval_loss": 0.6923198103904724, | |
| "eval_runtime": 681.4566, | |
| "eval_samples_per_second": 3.735, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 24.79, | |
| "learning_rate": 5.392296718972896e-05, | |
| "loss": 0.1865, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 24.79, | |
| "eval_cer": 0.15916051913382354, | |
| "eval_loss": 0.6866386532783508, | |
| "eval_runtime": 680.6783, | |
| "eval_samples_per_second": 3.739, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 25.62, | |
| "learning_rate": 4.536376604850214e-05, | |
| "loss": 0.1754, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 25.62, | |
| "eval_cer": 0.15712259709923845, | |
| "eval_loss": 0.693188488483429, | |
| "eval_runtime": 682.8906, | |
| "eval_samples_per_second": 3.727, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 26.45, | |
| "learning_rate": 3.680456490727532e-05, | |
| "loss": 0.1739, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 26.45, | |
| "eval_cer": 0.15602617120928625, | |
| "eval_loss": 0.6990784406661987, | |
| "eval_runtime": 682.7736, | |
| "eval_samples_per_second": 3.727, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 27.27, | |
| "learning_rate": 2.82453637660485e-05, | |
| "loss": 0.1657, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 27.27, | |
| "eval_cer": 0.1563002776817743, | |
| "eval_loss": 0.7062773108482361, | |
| "eval_runtime": 681.4615, | |
| "eval_samples_per_second": 3.735, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 28.1, | |
| "learning_rate": 1.968616262482168e-05, | |
| "loss": 0.1584, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 28.1, | |
| "eval_cer": 0.15688424364490103, | |
| "eval_loss": 0.7029954791069031, | |
| "eval_runtime": 682.9133, | |
| "eval_samples_per_second": 3.727, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 28.92, | |
| "learning_rate": 1.1126961483594864e-05, | |
| "loss": 0.1625, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 28.92, | |
| "eval_cer": 0.1568604082994673, | |
| "eval_loss": 0.6991938352584839, | |
| "eval_runtime": 683.4936, | |
| "eval_samples_per_second": 3.724, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 29.75, | |
| "learning_rate": 2.5677603423680457e-06, | |
| "loss": 0.1582, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 29.75, | |
| "eval_cer": 0.156514795790678, | |
| "eval_loss": 0.7013677358627319, | |
| "eval_runtime": 682.2827, | |
| "eval_samples_per_second": 3.73, | |
| "step": 14400 | |
| } | |
| ], | |
| "max_steps": 14520, | |
| "num_train_epochs": 30, | |
| "total_flos": 1.668287739327801e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |