| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.9568302779420463, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 4.9546619357382224e-05, | |
| "loss": 7.0936, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 4.905381431105855e-05, | |
| "loss": 4.7571, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.856100926473487e-05, | |
| "loss": 4.3059, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.80682042184112e-05, | |
| "loss": 3.9402, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.7575399172087526e-05, | |
| "loss": 4.5039, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.708259412576385e-05, | |
| "loss": 5.0274, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.658978907944018e-05, | |
| "loss": 4.4826, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.60969840331165e-05, | |
| "loss": 4.2794, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.560417898679283e-05, | |
| "loss": 4.1005, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.5111373940469154e-05, | |
| "loss": 3.6633, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_cer": 1.0385071558472365, | |
| "eval_loss": 3.3380627632141113, | |
| "eval_runtime": 108.6612, | |
| "eval_samples_per_second": 13.832, | |
| "eval_steps_per_second": 3.46, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.461856889414548e-05, | |
| "loss": 3.2821, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.41257638478218e-05, | |
| "loss": 3.1914, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 4.363295880149813e-05, | |
| "loss": 3.0884, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 4.3140153755174456e-05, | |
| "loss": 2.9599, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.264734870885078e-05, | |
| "loss": 2.8756, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 4.215454366252711e-05, | |
| "loss": 2.7076, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 4.166173861620343e-05, | |
| "loss": 2.6674, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 4.116893356987976e-05, | |
| "loss": 2.5175, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 4.0676128523556084e-05, | |
| "loss": 2.5046, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 4.018332347723241e-05, | |
| "loss": 2.3915, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_cer": 0.6388866924962442, | |
| "eval_loss": 2.358058214187622, | |
| "eval_runtime": 91.2055, | |
| "eval_samples_per_second": 16.479, | |
| "eval_steps_per_second": 4.123, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 3.969051843090874e-05, | |
| "loss": 2.2984, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 3.919771338458506e-05, | |
| "loss": 2.2398, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 3.8704908338261386e-05, | |
| "loss": 2.1845, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 3.821210329193771e-05, | |
| "loss": 2.0596, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 3.771929824561404e-05, | |
| "loss": 1.9458, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 3.722649319929036e-05, | |
| "loss": 1.8901, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 3.673368815296669e-05, | |
| "loss": 1.8351, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 3.6240883106643014e-05, | |
| "loss": 1.6854, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.574807806031934e-05, | |
| "loss": 1.6954, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 3.525527301399567e-05, | |
| "loss": 1.5061, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "eval_cer": 0.47789989720882425, | |
| "eval_loss": 1.4634768962860107, | |
| "eval_runtime": 93.897, | |
| "eval_samples_per_second": 16.007, | |
| "eval_steps_per_second": 4.004, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.476246796767199e-05, | |
| "loss": 1.5024, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.4269662921348316e-05, | |
| "loss": 1.5311, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.377685787502464e-05, | |
| "loss": 1.334, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 3.328405282870097e-05, | |
| "loss": 1.2053, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 3.27912477823773e-05, | |
| "loss": 1.1355, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 3.229844273605362e-05, | |
| "loss": 1.0514, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 3.1805637689729944e-05, | |
| "loss": 0.9352, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 3.131283264340627e-05, | |
| "loss": 0.8822, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 3.08200275970826e-05, | |
| "loss": 0.9451, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 3.0327222550758922e-05, | |
| "loss": 0.8043, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "eval_cer": 0.34268996599984186, | |
| "eval_loss": 0.826034426689148, | |
| "eval_runtime": 95.3049, | |
| "eval_samples_per_second": 15.77, | |
| "eval_steps_per_second": 3.945, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 2.9834417504435245e-05, | |
| "loss": 0.8479, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 2.9341612458111573e-05, | |
| "loss": 0.7237, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 2.88488074117879e-05, | |
| "loss": 0.6919, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 2.8356002365464223e-05, | |
| "loss": 0.6698, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 2.786319731914055e-05, | |
| "loss": 0.5657, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 2.7370392272816874e-05, | |
| "loss": 0.6165, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 2.68775872264932e-05, | |
| "loss": 0.6387, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 2.6384782180169525e-05, | |
| "loss": 0.5372, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 2.5891977133845852e-05, | |
| "loss": 0.5254, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 2.539917208752218e-05, | |
| "loss": 0.46, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "eval_cer": 0.22416383332015496, | |
| "eval_loss": 0.5074146389961243, | |
| "eval_runtime": 96.0964, | |
| "eval_samples_per_second": 15.641, | |
| "eval_steps_per_second": 3.913, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 2.4906367041198502e-05, | |
| "loss": 0.4557, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 2.441356199487483e-05, | |
| "loss": 0.445, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 2.3920756948551153e-05, | |
| "loss": 0.4523, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 2.342795190222748e-05, | |
| "loss": 0.3584, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 2.2935146855903804e-05, | |
| "loss": 0.4356, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 2.244234180958013e-05, | |
| "loss": 0.3989, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 2.1949536763256458e-05, | |
| "loss": 0.333, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 2.145673171693278e-05, | |
| "loss": 0.3701, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.096392667060911e-05, | |
| "loss": 0.4125, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 2.0471121624285432e-05, | |
| "loss": 0.3393, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "eval_cer": 0.12192614849371393, | |
| "eval_loss": 0.2698688209056854, | |
| "eval_runtime": 95.6778, | |
| "eval_samples_per_second": 15.709, | |
| "eval_steps_per_second": 3.93, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 1.997831657796176e-05, | |
| "loss": 0.2519, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 1.9485511531638083e-05, | |
| "loss": 0.2175, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 1.899270648531441e-05, | |
| "loss": 0.2418, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 1.8499901438990734e-05, | |
| "loss": 0.2157, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 1.800709639266706e-05, | |
| "loss": 0.2653, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 1.7514291346343388e-05, | |
| "loss": 0.2321, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 1.702148630001971e-05, | |
| "loss": 0.1881, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 1.652868125369604e-05, | |
| "loss": 0.1818, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 1.6035876207372362e-05, | |
| "loss": 0.1204, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 1.554307116104869e-05, | |
| "loss": 0.1077, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "eval_cer": 0.09330275954771883, | |
| "eval_loss": 0.17944632470607758, | |
| "eval_runtime": 96.1807, | |
| "eval_samples_per_second": 15.627, | |
| "eval_steps_per_second": 3.909, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 1.5050266114725015e-05, | |
| "loss": 0.174, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1.455746106840134e-05, | |
| "loss": 0.1222, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 1.4064656022077665e-05, | |
| "loss": 0.0985, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "learning_rate": 1.357185097575399e-05, | |
| "loss": 0.0946, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 1.3079045929430316e-05, | |
| "loss": 0.0832, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1.2586240883106643e-05, | |
| "loss": 0.074, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 1.209343583678297e-05, | |
| "loss": 0.0768, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 1.1600630790459296e-05, | |
| "loss": 0.1018, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 1.1107825744135621e-05, | |
| "loss": 0.0668, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "learning_rate": 1.0615020697811946e-05, | |
| "loss": 0.063, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_cer": 0.06167470546374634, | |
| "eval_loss": 0.13426831364631653, | |
| "eval_runtime": 95.5453, | |
| "eval_samples_per_second": 15.731, | |
| "eval_steps_per_second": 3.935, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 1.0122215651488272e-05, | |
| "loss": 0.0742, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 9.629410605164597e-06, | |
| "loss": 0.0679, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 9.136605558840922e-06, | |
| "loss": 0.0818, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 8.643800512517248e-06, | |
| "loss": 0.0704, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "learning_rate": 8.150995466193575e-06, | |
| "loss": 0.0537, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.54, | |
| "learning_rate": 7.6581904198699e-06, | |
| "loss": 0.0831, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 7.1653853735462255e-06, | |
| "loss": 0.0547, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 6.672580327222551e-06, | |
| "loss": 0.0765, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "learning_rate": 6.179775280898876e-06, | |
| "loss": 0.0408, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 5.6869702345752024e-06, | |
| "loss": 0.0356, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "eval_cer": 0.06918636830868981, | |
| "eval_loss": 0.0790172666311264, | |
| "eval_runtime": 97.3434, | |
| "eval_samples_per_second": 15.44, | |
| "eval_steps_per_second": 3.863, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "learning_rate": 5.194165188251528e-06, | |
| "loss": 0.059, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 4.70628819239109e-06, | |
| "loss": 0.0775, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 4.213483146067416e-06, | |
| "loss": 0.058, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 3.7206780997437416e-06, | |
| "loss": 0.0327, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 3.227873053420067e-06, | |
| "loss": 0.0378, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "learning_rate": 2.7350680070963927e-06, | |
| "loss": 0.0752, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "learning_rate": 2.2422629607727185e-06, | |
| "loss": 0.0269, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 1.749457914449044e-06, | |
| "loss": 0.0359, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "learning_rate": 1.2566528681253697e-06, | |
| "loss": 0.0345, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 7.638478218016952e-07, | |
| "loss": 0.0292, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "eval_cer": 0.044595556258401205, | |
| "eval_loss": 0.06197139248251915, | |
| "eval_runtime": 96.6477, | |
| "eval_samples_per_second": 15.551, | |
| "eval_steps_per_second": 3.89, | |
| "step": 10000 | |
| } | |
| ], | |
| "max_steps": 10146, | |
| "num_train_epochs": 3, | |
| "total_flos": 7.242528907589714e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |