| { | |
| "best_metric": 0.004195804195804196, | |
| "best_model_checkpoint": "custom_model_stage1_base_concat_synth/checkpoint-900", | |
| "epoch": 2.9605263157894735, | |
| "global_step": 900, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 4.9945175438596495e-05, | |
| "loss": 9.5118, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 4.9451754385964915e-05, | |
| "loss": 3.2727, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 4.890350877192983e-05, | |
| "loss": 2.6133, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.8355263157894734e-05, | |
| "loss": 3.1323, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.780701754385965e-05, | |
| "loss": 2.6306, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.7258771929824566e-05, | |
| "loss": 2.9819, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_cer": 0.3053536021150033, | |
| "eval_loss": 1.6067169904708862, | |
| "eval_runtime": 22.608, | |
| "eval_samples_per_second": 6.281, | |
| "eval_steps_per_second": 0.796, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.671052631578948e-05, | |
| "loss": 2.3285, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.616228070175439e-05, | |
| "loss": 1.9334, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.56140350877193e-05, | |
| "loss": 1.9042, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.506578947368421e-05, | |
| "loss": 2.0802, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.451754385964912e-05, | |
| "loss": 2.144, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "eval_cer": 0.08821502412129566, | |
| "eval_loss": 0.9317633509635925, | |
| "eval_runtime": 27.231, | |
| "eval_samples_per_second": 5.215, | |
| "eval_steps_per_second": 0.661, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.3969298245614036e-05, | |
| "loss": 1.7753, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 4.342105263157895e-05, | |
| "loss": 1.9885, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 4.287280701754386e-05, | |
| "loss": 1.8163, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.2324561403508774e-05, | |
| "loss": 1.8354, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 4.177631578947369e-05, | |
| "loss": 2.1775, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "eval_cer": 0.05823293172690763, | |
| "eval_loss": 0.8989996314048767, | |
| "eval_runtime": 24.7772, | |
| "eval_samples_per_second": 5.731, | |
| "eval_steps_per_second": 0.726, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 4.12280701754386e-05, | |
| "loss": 2.0764, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 4.067982456140351e-05, | |
| "loss": 2.1435, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 4.0131578947368425e-05, | |
| "loss": 1.9279, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 3.958333333333333e-05, | |
| "loss": 1.6143, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 3.9035087719298244e-05, | |
| "loss": 1.6676, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "eval_cer": 0.05952380952380952, | |
| "eval_loss": 0.6604741215705872, | |
| "eval_runtime": 25.9554, | |
| "eval_samples_per_second": 5.471, | |
| "eval_steps_per_second": 0.693, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 3.848684210526316e-05, | |
| "loss": 1.5091, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 3.7938596491228076e-05, | |
| "loss": 1.5469, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 3.739035087719299e-05, | |
| "loss": 1.7522, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 3.6842105263157895e-05, | |
| "loss": 1.6574, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 3.629385964912281e-05, | |
| "loss": 1.2664, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "eval_cer": 0.019313304721030045, | |
| "eval_loss": 0.6368018388748169, | |
| "eval_runtime": 26.7506, | |
| "eval_samples_per_second": 5.308, | |
| "eval_steps_per_second": 0.673, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.574561403508772e-05, | |
| "loss": 1.481, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 3.519736842105263e-05, | |
| "loss": 1.404, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.4649122807017546e-05, | |
| "loss": 1.2388, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.410087719298246e-05, | |
| "loss": 1.6538, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 3.355263157894737e-05, | |
| "loss": 1.2714, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "eval_cer": 0.04748982360922659, | |
| "eval_loss": 0.6383101344108582, | |
| "eval_runtime": 27.9443, | |
| "eval_samples_per_second": 5.082, | |
| "eval_steps_per_second": 0.644, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 3.3004385964912284e-05, | |
| "loss": 1.2309, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 3.24561403508772e-05, | |
| "loss": 1.2347, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 3.190789473684211e-05, | |
| "loss": 1.1184, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 3.1359649122807015e-05, | |
| "loss": 1.2385, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 3.081140350877193e-05, | |
| "loss": 0.9966, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "eval_cer": 0.025748086290883786, | |
| "eval_loss": 0.5416922569274902, | |
| "eval_runtime": 27.2235, | |
| "eval_samples_per_second": 5.216, | |
| "eval_steps_per_second": 0.661, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 3.0263157894736844e-05, | |
| "loss": 0.9694, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 2.9714912280701757e-05, | |
| "loss": 1.0465, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 2.916666666666667e-05, | |
| "loss": 0.9811, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 2.861842105263158e-05, | |
| "loss": 1.0744, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 2.8070175438596492e-05, | |
| "loss": 1.0068, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "eval_cer": 0.02813127930341594, | |
| "eval_loss": 0.4663618803024292, | |
| "eval_runtime": 28.8768, | |
| "eval_samples_per_second": 4.917, | |
| "eval_steps_per_second": 0.623, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 2.7521929824561405e-05, | |
| "loss": 1.1744, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 2.6973684210526317e-05, | |
| "loss": 1.0358, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "learning_rate": 2.642543859649123e-05, | |
| "loss": 0.8434, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 2.5877192982456143e-05, | |
| "loss": 0.901, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 2.5328947368421052e-05, | |
| "loss": 1.1405, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "eval_cer": 0.020134228187919462, | |
| "eval_loss": 0.5748289823532104, | |
| "eval_runtime": 28.573, | |
| "eval_samples_per_second": 4.97, | |
| "eval_steps_per_second": 0.63, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 2.4780701754385965e-05, | |
| "loss": 0.9414, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 2.4232456140350878e-05, | |
| "loss": 1.1748, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 2.368421052631579e-05, | |
| "loss": 1.0134, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 2.3135964912280703e-05, | |
| "loss": 1.0102, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 2.2587719298245616e-05, | |
| "loss": 0.9627, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "eval_cer": 0.02203856749311295, | |
| "eval_loss": 0.4435074031352997, | |
| "eval_runtime": 27.4558, | |
| "eval_samples_per_second": 5.172, | |
| "eval_steps_per_second": 0.656, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 2.2039473684210525e-05, | |
| "loss": 1.0858, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 2.149122807017544e-05, | |
| "loss": 0.8016, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.0942982456140354e-05, | |
| "loss": 0.813, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 2.0394736842105264e-05, | |
| "loss": 0.792, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "learning_rate": 1.9846491228070176e-05, | |
| "loss": 0.8551, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_cer": 0.0291583830351226, | |
| "eval_loss": 0.5917633771896362, | |
| "eval_runtime": 30.3337, | |
| "eval_samples_per_second": 4.681, | |
| "eval_steps_per_second": 0.593, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 1.929824561403509e-05, | |
| "loss": 0.8459, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 1.8750000000000002e-05, | |
| "loss": 0.8129, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 1.8201754385964914e-05, | |
| "loss": 0.8238, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "learning_rate": 1.7653508771929824e-05, | |
| "loss": 0.9552, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 1.7105263157894737e-05, | |
| "loss": 0.6585, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "eval_cer": 0.01694915254237288, | |
| "eval_loss": 0.4142173230648041, | |
| "eval_runtime": 28.0204, | |
| "eval_samples_per_second": 5.068, | |
| "eval_steps_per_second": 0.642, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 1.655701754385965e-05, | |
| "loss": 0.6229, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 1.6008771929824562e-05, | |
| "loss": 0.8082, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 1.5460526315789475e-05, | |
| "loss": 0.751, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 1.4912280701754386e-05, | |
| "loss": 0.6503, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 1.4364035087719299e-05, | |
| "loss": 0.6518, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "eval_cer": 0.008419689119170985, | |
| "eval_loss": 0.3558412492275238, | |
| "eval_runtime": 28.5056, | |
| "eval_samples_per_second": 4.981, | |
| "eval_steps_per_second": 0.631, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1.3815789473684213e-05, | |
| "loss": 0.6165, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 1.3267543859649122e-05, | |
| "loss": 0.7267, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 1.2719298245614037e-05, | |
| "loss": 0.8219, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 1.2171052631578948e-05, | |
| "loss": 0.7055, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "learning_rate": 1.162280701754386e-05, | |
| "loss": 0.5255, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "eval_cer": 0.0047879616963064295, | |
| "eval_loss": 0.3521033823490143, | |
| "eval_runtime": 27.0564, | |
| "eval_samples_per_second": 5.248, | |
| "eval_steps_per_second": 0.665, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 1.1074561403508772e-05, | |
| "loss": 0.7562, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "learning_rate": 1.0526315789473684e-05, | |
| "loss": 0.6348, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 9.978070175438597e-06, | |
| "loss": 0.7211, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 9.42982456140351e-06, | |
| "loss": 0.488, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 8.881578947368421e-06, | |
| "loss": 0.469, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "eval_cer": 0.004864489228630994, | |
| "eval_loss": 0.32997873425483704, | |
| "eval_runtime": 26.4415, | |
| "eval_samples_per_second": 5.37, | |
| "eval_steps_per_second": 0.681, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 8.333333333333334e-06, | |
| "loss": 0.4705, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "learning_rate": 7.785087719298246e-06, | |
| "loss": 0.5315, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 7.236842105263158e-06, | |
| "loss": 0.4967, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 6.68859649122807e-06, | |
| "loss": 0.6951, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "learning_rate": 6.140350877192982e-06, | |
| "loss": 0.434, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "eval_cer": 0.005936675461741424, | |
| "eval_loss": 0.3225802481174469, | |
| "eval_runtime": 27.8587, | |
| "eval_samples_per_second": 5.097, | |
| "eval_steps_per_second": 0.646, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 5.592105263157895e-06, | |
| "loss": 0.4898, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "learning_rate": 5.043859649122807e-06, | |
| "loss": 0.8127, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 4.4956140350877196e-06, | |
| "loss": 0.5916, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "learning_rate": 3.9473684210526315e-06, | |
| "loss": 0.3977, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 3.399122807017544e-06, | |
| "loss": 0.5842, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_cer": 0.006997900629811057, | |
| "eval_loss": 0.3334667980670929, | |
| "eval_runtime": 26.2691, | |
| "eval_samples_per_second": 5.406, | |
| "eval_steps_per_second": 0.685, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 2.850877192982456e-06, | |
| "loss": 0.4293, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 2.3026315789473684e-06, | |
| "loss": 0.4953, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "learning_rate": 1.7543859649122807e-06, | |
| "loss": 0.6281, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "learning_rate": 1.206140350877193e-06, | |
| "loss": 0.5186, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 6.578947368421053e-07, | |
| "loss": 0.5946, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "eval_cer": 0.004195804195804196, | |
| "eval_loss": 0.3096241056919098, | |
| "eval_runtime": 27.0145, | |
| "eval_samples_per_second": 5.256, | |
| "eval_steps_per_second": 0.666, | |
| "step": 900 | |
| } | |
| ], | |
| "max_steps": 912, | |
| "num_train_epochs": 3, | |
| "total_flos": 6.369477453856899e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |