| { | |
| "best_metric": 0.40675491094589233, | |
| "best_model_checkpoint": "./ko-xlsr5/checkpoint-18000", | |
| "epoch": 9.995938811425477, | |
| "eval_steps": 1800, | |
| "global_step": 18460, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 0.00029839643652561246, | |
| "loss": 11.754, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 0.00028839086859688194, | |
| "loss": 2.455, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 0.00027836859688195985, | |
| "loss": 1.4452, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eval_cer": 0.23372905998439186, | |
| "eval_loss": 0.9257286190986633, | |
| "eval_runtime": 2065.8069, | |
| "eval_samples_per_second": 7.149, | |
| "eval_steps_per_second": 0.447, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 0.0002683630289532294, | |
| "loss": 1.2319, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 0.00025834075723830735, | |
| "loss": 1.1229, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 0.0002483184855233853, | |
| "loss": 1.0684, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "eval_cer": 0.18344095102682245, | |
| "eval_loss": 0.6873096227645874, | |
| "eval_runtime": 2137.5713, | |
| "eval_samples_per_second": 6.909, | |
| "eval_steps_per_second": 0.432, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 0.0002383129175946548, | |
| "loss": 0.9894, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 0.00022829064587973272, | |
| "loss": 0.9499, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "learning_rate": 0.00021826837416481068, | |
| "loss": 0.9269, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "eval_cer": 0.15986839995296181, | |
| "eval_loss": 0.5857065916061401, | |
| "eval_runtime": 2062.0942, | |
| "eval_samples_per_second": 7.162, | |
| "eval_steps_per_second": 0.448, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 0.00020826280623608017, | |
| "loss": 0.8687, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 0.0001982405345211581, | |
| "loss": 0.856, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 0.00018821826280623606, | |
| "loss": 0.8264, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "eval_cer": 0.14419078265145766, | |
| "eval_loss": 0.5356672406196594, | |
| "eval_runtime": 2072.5322, | |
| "eval_samples_per_second": 7.126, | |
| "eval_steps_per_second": 0.445, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 4.22, | |
| "learning_rate": 0.00017819599109131402, | |
| "loss": 0.7946, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "learning_rate": 0.00016819042316258353, | |
| "loss": 0.781, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 0.00015816815144766144, | |
| "loss": 0.7637, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "eval_cer": 0.1364936231171358, | |
| "eval_loss": 0.5068702697753906, | |
| "eval_runtime": 2140.8687, | |
| "eval_samples_per_second": 6.898, | |
| "eval_steps_per_second": 0.431, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 5.2, | |
| "learning_rate": 0.00014814587973273943, | |
| "loss": 0.736, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 5.52, | |
| "learning_rate": 0.00013812360801781736, | |
| "loss": 0.7097, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 5.85, | |
| "learning_rate": 0.00012811804008908684, | |
| "loss": 0.7033, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 5.85, | |
| "eval_cer": 0.12771672314813826, | |
| "eval_loss": 0.47439366579055786, | |
| "eval_runtime": 2140.9721, | |
| "eval_samples_per_second": 6.898, | |
| "eval_steps_per_second": 0.431, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 6.17, | |
| "learning_rate": 0.0001180957683741648, | |
| "loss": 0.6863, | |
| "step": 11400 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 0.00010807349665924275, | |
| "loss": 0.6564, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 6.82, | |
| "learning_rate": 9.806792873051224e-05, | |
| "loss": 0.652, | |
| "step": 12600 | |
| }, | |
| { | |
| "epoch": 6.82, | |
| "eval_cer": 0.12095765493206187, | |
| "eval_loss": 0.4477333426475525, | |
| "eval_runtime": 2142.7063, | |
| "eval_samples_per_second": 6.892, | |
| "eval_steps_per_second": 0.431, | |
| "step": 12600 | |
| }, | |
| { | |
| "epoch": 7.15, | |
| "learning_rate": 8.804565701559018e-05, | |
| "loss": 0.6258, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 7.47, | |
| "learning_rate": 7.802338530066816e-05, | |
| "loss": 0.6085, | |
| "step": 13800 | |
| }, | |
| { | |
| "epoch": 7.8, | |
| "learning_rate": 6.801781737193764e-05, | |
| "loss": 0.5999, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 7.8, | |
| "eval_cer": 0.11436428945596049, | |
| "eval_loss": 0.4309966564178467, | |
| "eval_runtime": 2141.1385, | |
| "eval_samples_per_second": 6.897, | |
| "eval_steps_per_second": 0.431, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 8.12, | |
| "learning_rate": 5.799554565701558e-05, | |
| "loss": 0.5889, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 8.45, | |
| "learning_rate": 4.798997772828508e-05, | |
| "loss": 0.5754, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 8.77, | |
| "learning_rate": 3.7967706013363024e-05, | |
| "loss": 0.5606, | |
| "step": 16200 | |
| }, | |
| { | |
| "epoch": 8.77, | |
| "eval_cer": 0.11144845575736842, | |
| "eval_loss": 0.41504111886024475, | |
| "eval_runtime": 2070.5583, | |
| "eval_samples_per_second": 7.132, | |
| "eval_steps_per_second": 0.446, | |
| "step": 16200 | |
| }, | |
| { | |
| "epoch": 9.1, | |
| "learning_rate": 2.7945434298440978e-05, | |
| "loss": 0.5528, | |
| "step": 16800 | |
| }, | |
| { | |
| "epoch": 9.42, | |
| "learning_rate": 1.792316258351893e-05, | |
| "loss": 0.5452, | |
| "step": 17400 | |
| }, | |
| { | |
| "epoch": 9.75, | |
| "learning_rate": 7.917594654788419e-06, | |
| "loss": 0.5371, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 9.75, | |
| "eval_cer": 0.10851124106006992, | |
| "eval_loss": 0.40675491094589233, | |
| "eval_runtime": 2143.676, | |
| "eval_samples_per_second": 6.889, | |
| "eval_steps_per_second": 0.431, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 18460, | |
| "total_flos": 1.2921277443498574e+20, | |
| "train_loss": 1.1888314855705748, | |
| "train_runtime": 43297.9494, | |
| "train_samples_per_second": 27.297, | |
| "train_steps_per_second": 0.426 | |
| } | |
| ], | |
| "logging_steps": 600, | |
| "max_steps": 18460, | |
| "num_train_epochs": 10, | |
| "save_steps": 1800, | |
| "total_flos": 1.2921277443498574e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |